Australas Phys Eng Sci Med DOI 10.1007/s13246-016-0494-2

EPSM 2016 ABSTRACTS

EPSM 2016, Engineering and Physical Sciences in Medicine 6–10 November 2016, Sydney, Australia

 Australasian College of Physical Scientists and Engineers in Medicine 2016

Contents Monday 7th November 1000–1045 Keynote Speaker Session Medical Physics: Quo Vadis Robert Jeraj 1115–1230 Concurrent Session Innovation in treatment techniques Keynote Speaker Engineering protein nonocages for biomedical applications Sierin Lim 1330–1355 Invited Speaker Session Real-time adaptive radiotherapy: development and translation to clinical practice Jeremy Booth 1355–1415 Invited Speaker Session Report on Richard Bates Travel Scholarship Robin Hill 1330–1500 Concurrent Session Radiotherapy – Tracking and Imaging Keynote Speaker Innovations in anatomical modelling to advance image guided therapy Kristy Brock

123

Australas Phys Eng Sci Med

1330–1500 Concurrent session Diagnostic – Quality control programs and ultrasound

Radiotherapy Imaging – EPID and CBCT 1330–1500 Concurrent Session

Keynote Speaker

Radiotherapy – Deformable Image Registration

Individualised CT Dosimetry

Keynote Speaker

Hilde Bosmans

Implementing Dose Accumulation in Standard Practice: Is it clinically feasible and evidence driven?

Invited Speaker Australia’s role in the development of diagnostic ultrasound technology Robert Gill Invited Speaker Current Australian Medical Ultrasound Research Christian Langton Invited Speaker

Kristy Brock 1415–1500 Concurrent Session Diagnostic – Performance Testing of Diagnostic Equipment 1415–1500 Concurrent Session Radiotherapy – Stereotactic

Clinical Impact of technology advances in ultrasound

1530–1600 Concurrent Session

Lyndal Macpherson

Medical Physics Education 1530–1610 Concurrent Session

1430–1500 Concurrent Session Management issues 1530–1600 Poster Session – Mini Orals 1630–1715 Concurrent Session Diagnostic – MRI and PET Tuesday 8th November 0800–0900 Invited Speaker Session

Radiotherapy – Treatment Planning 1530–1700 Concurrent Session Radiotherapy – Dosimetry Standards Wednesday 9th November 0900–1030 Concurrent Session Radiotherapy – Adaptive Radiotherapy

The role of MR physics in the public health service

Keynote Speaker

Donald McRobbie

Adaptive Radiotherapy @ NKI

0900–1030 Concurrent Session Radiotherapy – Patient specific QA Keynote Speaker IMRT dose delivery in practice: what are we actually delivering?

Jan-Jakob Sonke 0900–1030 Concurrent Session Radiotherapy – Brachytherapy 0900–1030 Concurrent Session Diagnostic – Mammography

Stephen Kry

Keynote Speaker

Invited Speaker Pre-clinical evaluation of the Integral Quality Monitor for real-time beam verification

Breast Imaging: virtual clinical trials Hilde Bosmans

Jurgen Oellig 0900–1030 Concurrent Session Radiotherapy – Treatment Planning and Modelling 0900–1030 Concurrent Session Radiation Protection Keynote Speaker Accreditation of Radiation Protection Experts and Medical Physics Experts in the UK Medical Sector David Sutton 1100–1230 Concurrent Session Radiotherapy – MRI for Radiotherapy and MRI Linac Keynote Speaker MRI Guided Radiation Therapy Jan-Jakob Sonke 1100–1230 Concurrent Session Radiotherapy – New Treatment Techniques and Equipment

123

1100–1230 Concurrent Session

1100–1230 Concurrent Session Radiotherapy – Small Field Dosimetry 1100–1230 Concurrent Session Radiotherapy – Nanoparticles Keynote Speaker Protein nanocages as a display platform with high special control Sierin Lim 1100–1230 Concurrent Session Diagnostic – Reference Levels and Audits Keynote Speaker Benchmarking and Optimising CT in Scotland – an exercise in patience David Sutton 1330–1500 Concurrent Session Radiotherapy – Particle Therapy 1330–1500 Concurrent Session Radiotherapy – Quality Management and Control

Australas Phys Eng Sci Med

1530–1630 Poster Session – Mini Orals 1630–1730 Concurrent Session Radiotherapy – Gating and Image Guidance Thursday 10th November 0900–1015 Concurrent Session Science and Engineering in Medicine: The Needs, The Vision, The Future Invited Speaker Science and engineering in medicine: the vision (from the perspective of radiation oncology) Verity Ahern Invited Speaker Medical Physics innovations today – Medical Physics jobs in the future Robert Jeraj 0900–1015 Concurrent Session Radiotherapy – Innovations in In Vitro Dosimetry 0900–1015 Concurrent Session Diagnostic – Patient Dose Measurement Invited Speaker Optimisation of image quality and patient dose in radiography of paediatric extremities using direct digital radiography Adam Jones 1045–1230 Concurrent Session Science and Engineering in Medicine: Reality and Potential Invited Speaker The importance of high quality radiotherapy Stephen Kry 1045–1145 Concurrent Session Radiotherapy – The Impact of Radiation on the Patient 1330–1430 Science and Engineering in Medicine: Our responsibilities Invited Speaker The 4 ‘R’s: Responsibilities, Roles and Remaining Relevant Sandra Turner

KS01 Medical physics: Quo vadis? Robert Jeraj University of Wisconsin, Madison, USA. ([email protected]) Medical physics is intimately connected with medicine, and is progressing along a similar path. General trend of medicine, particularly oncology, towards personalized treatment gave rise to precision

medicine, which addresses the highly complex nature of disease. In the past, little could be done to tackle this complexity, but the emergence of targeted therapies is bringing personalized therapies within reach. However, there are severe obstacles to overcome. For example, cancers evolve in time to become harder targets to treat. Understanding treatment resistance, and its development, often connected with the highly heterogeneous nature of the disease, is the key obstacle. Use of multi-modality imaging techniques such as molecular imaging is one of the solutions that medical physics can offer. Radiomics, where large amounts of useful data are extracted from a single medical scan, is turning into a promising emerging field. However, much more data are available from genomic and other tests, which need to be integrated into the picture as well. To analyze such large amounts of data, we must learn from ‘‘big-science’’ physics approaches. For example, when analyzing data to detect the Higgs boson, CERN physicists relied on a deep understanding of the underlying fundamental physics. Likewise, when we move to medicine, we have to better understand the underlying biological principles that drive observations. However, medical physicists cannot do this alone. To achieve this effectively, it’s essential that physicists much more effectively partner with biologists and other scientists beyond current typical collaborative frameworks. In summary, in order for medical physics to remain at the forefront of scientific research, it will have to move beyond the current boundaries towards better addressing advanced disease, better integration of ‘‘big-science’’ physics approaches, better understanding of the biological basis of the disease, and better collaborations beyond current collaborative frameworks.

O001 A small-animal model of radiotherapy R. A. Begley1, A. Kennedy2, M. House1, M. A. Ebert2 1

School of Physics, University of Western Australia, Australia. ([email protected] [Presenting author]), ([email protected]). 2Medical Physics Research Group, Sir Charles Gairdiner Hospital. ([email protected]), ([email protected]) Introduction Small-animal radiotherapy devices (SARDs) allow precise irradiation for pre-clinical studies of radiotherapy. Recently developed devices mimic human treatments with low energy X-ray beams. The brain is a region where dose accuracy is particularly important due to the proximity of small, sensitive structures. This study assesses the accuracy of a SARD and compares this with two human clinical systems; volumetric arc therapy and Cyberknife. Method Treatment plans were produced for two situations in the rat; irradiation of a brain tumour target and irradiation of the unilateral hippocampus. Generalised organ-at risk (OAR) structures representative of typical clinical scenarios were outlined. The tumour target was planned with a prescribed dose of 3 Gy to 90% of the target. Past research has found that total doses of 1.8–2 Gy induce toxicities in the hippocampus and optic nerve. These values were used as the threshold for the small OAR. This study also compared the SARD and clinical systems in selectively targeting the unilateral hippocampus. Results Cyberknife achieved the highest dose conformity followed by the SARD and then VMAT. For targets adjacent sensitive structures such as the hippocampus or optic nerve, only the SARD platform could keep the dose below threshold (Fig. 1). Conclusion The SARD was theoretically the only platform successful in selectively inducing toxicities in the unilateral hippocampus. The differences in precision arise primarily from beam energy and beam delivery geometry. The steeper gradients better represent those

123

Australas Phys Eng Sci Med

a

Volume (%)

b 100 SARD

VMAT

Cyberknife

50

0 0

2

4 6 Dose (Gy)

8

10

Fig. 1 (a) Colour map dose distributions for unilateral irradiation of the hippocampus with the SARD. (b) Dose-volume histogram for dose to the spared right hippocampus achievable in human clinical treatments, scaled relative to the anatomy of humans and small animals, allowing for better translational research. References 1. Emami, B., et al., Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys, 1991. 21(1): p. 109–22. 2 Mayo, C., et al., Radiation dose-volume effects of optic nerves and chiasm. Int J Radiat Oncol Biol Phys, 2010. 76(3 Suppl): p. S28–35. Acknowledgements We acknowledge the Department of Health WA, Cancer Council WA and BHP Billiton for their funding support.

O002 Indirect radio-chemo-beta therapy: A targeted approach to increase radiosensitisation of resistant cancer cells to radiotherapy S. Oktaria1, S. Corde2, M. L. F. Lerch1, A. B. Rosenfeld1, M. Tehei1 1

Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected] [Presenting author]). 2 Radiation Oncology Department, Prince of Wales Hospital, Randwick, NSW, Australia. ([email protected]) Introduction Despite the use of multimodal treatments incorporating surgery, chemotherapy and radiotherapy, local control of gliomas remains a major challenge. Our innovative approach uses the synergy created by combining two pharmacological drugs with optimised x-ray energies. The two drugs are a radiosensitiser (bromodeoxyuridine – BrUdR), which makes the tumour more sensitive to radiation, and a chemotherapy drug (methotrexate – MTX), which impairs the growth of tumour cells and DNA damage repair after irradiation. In combination with X-rays tuned at optimum energy, a synergy was revealed, inducing more biologically efficient radiation doses. Methods BrUdR incorporation and radiosensitisation in 9L rat gliosarcoma cells were analysed. Incorporation of BrUdR in the DNA

123

Fig. 1 Incorporation of BrUdR in 9L cells. Cell nuclei labeled with propidium iodide (PI) (left); MoBU-1 (middle); and overlay of PI and MoBU-1 (right) was detected by the fluorescence of labelled anti-BrUdR antibodies (MoBU-1). Cells pre-treated with 0.01 lM MTX and/or 10 lM BrUdR were irradiated in vitro with 50, 125, 250 kVp, 6 MV and 10 MV X-rays. The cytotoxicity was assessed using clonogenic survival as the radiobiological endpoint. The photon energy with maximum effect was determined using radiation sensitisation enhancement factors at 10% clonogenic survival (SER10). Results Fig. 1 supports the hypothesis that BrUdR is incorporated into DNA. The SER10 of irradiated BrUdR and MTX-treated cells was found to be greatest at 125 kVp (SER10 = 2.3) [1]. Conclusion We assessed a new indirect radio-chemo-beta therapy treatment approach. We established in vitro that the combination of drugs with optimised X-ray energies more than doubles the number of glioma tumour cells killed compared to irradiation alone, or either drug and irradiation. This could in principle maximise the killing of tumour cells whilst minimising the effect on healthy tissue. We showed that this is a highly effective form of chemo-radiation therapy. References 1. Oktaria S, Corde S, Lerch ML, Konstantinov K, Rosenfeld AB, Tehei M (2015) Indirect radio-chemo-beta therapy: A targeted approach to increase biological efficiency of x-rays based on energy. Phys Med Biol 60:7847–7859. doi: 10.1088/0031-9155/ 60/20/7847.

IS01 Engineering protein nanocages for biomedical applications Sierin Lim School of Chemical and Biomedical Engineering, NTU-Northwestern Institute for Nanomedicine, Nanyang Technological University, Singapore. ([email protected] [Presenting author]) Introduction Protein nanocages can be engineered to tailor their function as carriers for therapeutic and diagnostic agents. They are formed by the self-assembly of multiple subunits forming hollow cage-like structures of nanometer size. Due to their proteinaceous nature, the protein nanocages allow facile modifications on their internal and external surfaces, as well as the subunit interfaces. In this presentation, we will focus on three protein cage representatives that are ferritin from Archaeoglobus fulgidus, E2 protein from Geobacillus stearothermophilus, and vault from rat as carriers of MRI contrast agent and drug that are responsive to external stimuli. Method Modifications on the internal and the external surfaces of the nanocages allow conjugation of small molecule drugs or contrast agent and targeting ligands, respectively. The subunit interaction is of special interest in engineering a controlled release property onto the

Australas Phys Eng Sci Med protein cage. Upon achieving the modifications by site-directed mutagenesis, the purified modified protein nanocages were characterized by electron microscopy, dynamic light scattering and circular dichroism techniques, vibrating sample magnetometry, and surface plasmon resonance technique. Results Manganese-loaded ferritin shows enhancement of relaxivity values that are about one order of magnitude over conventional Mn-based MRI contrast agent. The loading of metal ions in ferritin can be translated to other protein nanocages such as the E2 protein that naturally are incapable of carrying metal ions. Engineering of E2 subunit interactions results in pH-responsive disassembly of the cage-like structure while resurfacing of the vault lumen, made the modulation of release rate of the packaged cargo molecules possible. Conclusion Engineering of protein nanocages imparts non-natural functions for biomedical applications. In the presented works, we have shown that MRI contrast agents and drugs can be loaded to the lumen of the protein nanocages without jeopardizing their efficacies. Controlling the release rate of molecules from the lumen is possible by modifying the interactions between subunits or other interacting partners. References 1. 1. Sana B, Johnson E, Lim S* (2015) The Unique Self-assembly/ disassembly Property of Archaeoglobus fulgidus Ferritin and Its Implications on Molecular Release from the Protein Cage, Biochimica et Biophysica Acta (BBA) – General Subjects, 1805(12):2544–2551. 2. 2. Bu¨cheler J, Howard C**, de Bakker CJ, Goodall S, Jones ML, Win T, Peng T, Tan CH, Chopra A, Mahler S, Lim S* (2015) Development of a Protein Nanoparticle Platform for Targeting EGFR Expressing Cancer Cells, Journal of Chemical Technology & Biotechnology, 80(7):1230–1236. 3. 3. Peng T, Lee H, Lim S* (2015) Design of a Reversible Inversed pH-responsive Caged Protein, Biomaterials Science, 3:627–635. 4. 4. Walsh EG*, Mills DR, Lim S, Sana B, Brilliant KE, Park WKC (2013) MRI Contrast Demonstration of Antigen-specific Targeting with an Iron-based Ferritin Construct, Journal of Nanoparticle Research 15:1409–1418. 5. 5. Sana B, Poh CL, Lim S* (2012) A Manganese-ferritin Nanocomposite as an Ultrasensitive T2 Contrast Agent, Chemical Communications 48(6):862–864. 6. 6. Sana B, Johnson E, Sheah K, Poh CL*, Lim S* (2010) IronBased Ferritin Nanocore as a Contrast Agent, Biointerphases 5(3):FA48–FA52.

5005 Australia. 5Radiation Physics Laboratory, Sydney Medical School, University of Sydney, NSW Australia. Introduction Recurrent metastatic prostate cancer is treated with androgen deprivation therapy (ADT), providing early but transient control of progression. The benefits of ADT are lost as the cancer progresses to the ‘‘late-stage’’ castrate-resistant form which has low median survival. The minimal residual disease state derived from ADT provides a window of opportunity for effective systemic adjuvant treatment such as targeted alpha therapy (TAT), where antibodies specific for cancer biomarkers are labelled with alpharadionuclides to increase the therapeutic ratio. Method Both preclinical and clinical efficacy of TAT for residual prostate cancer management is examined. Monte Carlo calculations show the preferred radioisotopes for efficacy and therapeutic gain. This work reviews the current status of TAT for prostate cancer, looking at options and directions that the modality should take in near future. Discussion Three vector/receptor approaches for targeted radiotherapy of prostate cancer are available: viz c595/muc1, J591/PSMA and PSMA617/PSMA. They are labelled with beta or alpha emitters for therapy or PET imaging. The most advanced is Lu177 or Ac225PSMA617. Promising results are obtained with Ac225 with 2-year survivals. However, Ga68-PSMA shows high uptake in the salivary glands for PET. Consequently, Ac225-PSMA617 causes xerostomia in most patients. Lu177-PSMA617 spares the salivary glands but has reduced efficacy. A Lu177-J591 trial was completed with some efficacy at maximum tolerance dose. Bi213-C595 is a generic anti-cancer therapy tested in vitro and in vivo, but has not yet reached the clinic. The only FDA-approved TAT is Ra223Cl2 for palliative therapy of bone metastasis. However, it does not target prostate cancer cells. Conclusion Based on clinical status, the current pecking order of radio-conjugates for curative therapy is Ac225-PSMA617 (with blocking of salivary gland receptors), followed by Lu177-J591 and Bi213-C595. Note that a palliative therapy was approved before potentially curative therapies, thus throwing a challenge to the Nuclear Medicine profession.

O004 Personalised therapies for breast cancer subtypes: What is the role for radiotherapy? L. J. Rogers1,2, N. Suchowerska1,2, J. Toohey1, S. Carroll1, J. G. Lyons3,4,5, J. Beith6, D. R. McKenzie1,2 1

O003 Quo vadis targeted radio-immunotherapy for metastatic prostate cancer? Barry J. Allen1, Loredana G. Marcu2, Eva Bezak3,4, Chen-Yu Huang5 1

Faculty of Medicine, University of Western Sydney, Liverpool, NSW 2170 Australia. ([email protected] [Presenting author]). 2 Faculty of Science, University of Oradea, Oradea 410087, Romania School of Physical Sciences, The University of Adelaide, North Terrace, Adelaide, South Australia 5005, Australia. 3International Centre for Allied Health Evidence and Sansom Institute for Health Research, Division of Health Sciences, University of South Australia, Adelaide, South Australia 5001. 4School of Physical Sciences, University of Adelaide, North Terrace, Adelaide, South Australia

VectorLAB, Radiation Oncology, Chris O’Brien Lifehouse, Sydney, Australia. 2School of Physics, The University of Sydney, Sydney, Australia. ([email protected]), ([email protected] [Presenting Author]). 3Immune Imaging, Centenary Institute, Sydney, Australia; Dermatology. 4Cancer Services, Royal Prince Alfred Hospital, Sydney, Australia. 5Department of Dermatology, The University of Sydney, Sydney, Australia. 6Medical Oncology, Chris O’Brien Lifehouse, Sydney, Australia. Introduction HER2-positive breast cancer patients are now routinely given Trastuzumab (Herceptin), a monoclonal antibody against HER2, which has led to a significant improvement in outcomes in this previously high risk breast cancer subtype. Ionising radiation and HER2-targeting agents are sometimes used concurrently, but it is not known if there is an interaction between them. There are preliminary indications that Trastuzumab may act as a radiation sensitizer both in vitro1, and in vivo2, but no information is available for other commonly used HER2-targeting agents. The aim of this study is to develop a protocol for identifying any synergies and using them in a mathematical model to maximise therapeutic benefits. If the

123

Australas Phys Eng Sci Med targeting agents make the cancer cells more sensitive to treatment, can we customise the treatment to reduce the prescribed radiation dose? Method Two HER2+ breast cancer cell lines (HCC-1954 and BT474) were treated to a range of doses of HER2-targeting agents (Trastuzumab and TDM-1) and ionizing radiation (0–4 Gy). Synergy (S) is defined as the fractional difference between the observed (S0) and the predicted survival for each treatment given alone (S1 x S2): S¼

S0  S1  S2 S0

The observed response was determined using two assays: the clonogenic assay and the confluence assay. Results 1.0 0.9

*

0.8 0.7 0.6

Trastuzumab Radiation

B

C

1.4

1.4

1.2

**

1.0 0.8 0.6

-

+ -

+

+ +

Trastuzumab Radiation

Survival Fraction

Predicted Survival

1.1

Survival Fraction

Confluence Fraction

A

1.2 1.0

**

0.8 0.6

-

+ -

+

+ +

TDM-1 Radiation

-

+ -

+

+ +

Figure 1 shows the confluence fraction (A) and the survival fraction (B,C) for HCC-1954 cells with and without radiation and HER2targeting agents. A significant synergy for all radiation doses was observed for the confluence assay, indicating that an improvement of 10% is achievable. (*p \ 0.05, **p \ 0.01) Conclusion The innovation of this study is the potential to customize radiation treatment, exploiting the sensitizing action of targeting agents to give the best outcome with a minimum dose and potentially modifying any increased toxicity of the combined treatment modalities. Acknowledgements We acknowledge funding from the Sydney Breast Cancer Foundation

the development and translation to clinical practice of Kilovoltage Intrafraction Monitoring (KIM) to localise the target in real-time and MLC tracking to continuously adapt treatment delivery in real-time. Method Four key ‘First in human’ clinical trials have clinically realised the improved ability to track and adapt radiotherapy delivery in real-time. The KIMGating clinical trial evaluates the ability of fluoroscopy during treatment (KIM) to provide real time 3D/6DoF localisation of fiducial markers in prostate cancer with gated radiotherapy. The CALYPSO trial evaluated electromagnetic guided MLC tracking for prostate. The LIGHT SABR trial evaluates electromagnetic guided MLC tracking to deliver exquisite compact dose for lung cancer SABR. The SPARK trial brings kV localisation and real time adaptive RT together to enable KIM-guided MLC tracking to the prostate on a standard linac. Results The KIM technology is shown to deliver sub-mm accuracy and precision in localising the prostate. KIM gating enabled safe dose escalation, SABR, focal boosts, and lower normal tissue dose. Electromagnetic-guided MLC tracking delivered high fidelity prostate radiotherapy for over 850 fractions, including plans with dose painting. MLC tracking with electromagnetic guidance has been delivered successfully to 7 lung SABR patients ensuring compact dose distributions with motion exceeding 2 cm peak to peak. The first patient has now received KIM-guided MLC tracking for prostate SABR, which was delivered with high fidelity. Conclusion Real time adaptive radiotherapy has been translated into clinical practice on a standard linac. Development continues towards deformation tracking and use in the liver, pancreas and kidney SABR environment. Acknowledgements This research is funded by Varian Medical Systems. A large team of Physicists, Radiation Therapists, Radiation Oncologists and Academics have contributed to these implementations at Northern Sydney Cancer Centre with Radiation Physics Lab, University of Sydney.

References 1. Liang K, Lu Y, Jin W, Ang KK, Milas L, Fan Z. Sensitization of breast cancer cells to radiation by trastuzumab. Molecular cancer therapeutics. 2003;2(11):1113–20. 2. Horton JK, Halle J, Ferraro M, Carey L, Moore DT, Ollila D, et al. Radiosensitization of Chemotherapy-Refractory, Locally Advanced or Locally Recurrent Breast Cancer With Trastuzumab: A Phase II Trial. International Journal of Radiation Oncology Biology Physics. 2010;76(4):998–1004.

IS01 Boyce Worthley Young Achiever Award Winner 2015: Real-time adaptive radiotherapy: Development and translation to clinical practice J. Booth1,2 1

Northern Sydney Cancer Centre, Royal North Shore Hospital, NSW, Australia. 2School of Physics, University of Sydney, Sydney, Australia. ([email protected]) Introduction Image-guided radiotherapy has led to demonstrable gains in local tumour control across a number of cancer types, enabling accurate targeting and modulated field delivery. Real-time adaptive radiotherapy, where the tumour is localised and targeted continuously during radiotherapy delivery, will deliver further improvements in patient outcomes by directly mitigating motion. Real-time adaptive radiotherapy has not been applied clinically before now on standard radiotherapy devices. This presentation demonstrates

123

IS02 Report on Richard Bates travel scholarship R. F. Hill 1

Department of Radiation Oncology, Chris O’Brien Lifehouse, Sydney, Australia. ([email protected] [Presenting author]) Introduction This presentation will summarise the visits to several organisations as part of the Richard Bates Travel scholarship. Method The three specific topics investigated during the scholarship were: 1. Small field dosimetry, 2. Radiation dosimetry audits and 3. Kilovoltage x-ray beam dosimetry These three topics were selected due to being of interest (clinical and research), the need to develop more advanced dosimetry tools and the need for increased accuracy and auditing. The first centre visited was to the Mt Vernon Cancer Centre, London, to meet with the staff of the National Trials QA Group. This provided opportunity to gain knowledge from their experience in setting up dosimetry audits for clinical trials. The next centre was Medical Physics group at the Queen Elizabeth Hospital, Birmingham due to their extensive experience in small field dosimetry with their Cyberknife unit. The final radiotherapy centre I had the opportunity to visit was the Cantonal Hospital of Lucerne. This gave me the opportunity to see a newly released kilovoltage x-ray unit as well as talk to physicists who are involved in ESTRO teaching course. The final visit was to the Section of Dosimetry and Medical Radiation

Australas Phys Eng Sci Med Physics at the IAEA in Vienna. I had the opportunity to meet with a number of Medical Physicists and Radiation Oncologists. Results The visits to all centres provided opportunity to develop both clinical and research collaborations with a number of projects started from these visits. It was also an opportunity to share ideas and protocols on developing dosimetry audits from centres that have extensive experience in this. Conclusion The Richard Bates travel scholarship has provided an excellent opportunity for developing collaborative links with a number of overseas centres. Acknowledgements The support of the ACPSEM Council through this scholarship is greatly appreciated to allow this travel to take place.

O005 First integration of online 4D ultrasound guidance with MLC tracking for real-time motion compensation in radiation therapy S. Ipsen1, R. Bruder1, R. O’Brien2, P. Keall2, A. Schweikard1, P. Poulsen3

Fig. 1 3%/3 mm c-failure rates for high-modulation VMAT plans and nine trajectories

Results The overall system latencies were 95.5, 170 and 186 ms for the small, medium and large volumes respectively. When using a medium volume size, ultrasound tracking reduced the mean c-failure rate by 76% for prostate (from 13.9 to 4.6%) and by almost 99% for lung (from 21.8 to 0.6%) with high-modulation VMAT plans and from 5% (prostate) and 18% (lung) to 0% with low modulation (Fig. 1). Conclusion For the first time, real-time ultrasound tracking was successfully integrated with MLC tracking. The combined system showed similar accuracy and latency as other methods (e.g. [4]) while holding the potential to measure target motion non-invasively.

1

Institute for Robotics and Cognitive Systems, University of Luebeck, Germany. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2 Radiation Physics Laboratory, Sydney Medical School, University of Sydney, NSW Australia. ([email protected]), ([email protected]). 3Department of Clinical Medicine, Aarhus University and Department of Oncology, Aarhus University Hospital, Denmark. ([email protected]) Introduction Motion compensation of moving targets has now become possible with MLC tracking. However, rather than direct target visualization, real-time localization methods still rely on correlation models with x-ray imaging or implanted electromagnetic transponders . Modern ultrasound imaging yields volumetric data in real-time (4D) without ionizing radiation. These are the first results of online 4D ultrasound-guided MLC tracking in a phantom. Method We used a 4D ultrasound station (Vivid7 dimension, GE) modified for real-time tracking to detect a 2 mm spherical lead marker inside a water tank. The marker was rigidly attached to a motion stage programmed to reproduce nine tumor trajectories (five prostate, four lung). The ultrasound frame rate depended on volume size and lay between 4 and 56 ms (Table 1). The 3D marker position from ultrasound was used for real-time MLC aperture adaption. The overall system latency was measured and compensated by prediction for lung trajectories. 358 VMAT fields were delivered to a biplanar diode array dosimeter using the same trajectories and a medium volume size for dosimetric evaluation. Dose measurements with and without MLC tracking were compared to a static reference dose using a 3%/3 mm c-test. Table 1 Parameters of different ultrasound volume sizes Volume type

Small Medium (high-resolution zoom)

Large (maximum volume size)

Acquisition rate

250 Hz (4 ms)

17.8 Hz (56 ms)

21.3 Hz (47 ms)

Depth

105–150 mm

170 mm

170 mm

Angle 1

8

30

30

Angle 2

6

22

46

Motion range in 130 mm depth 17 mm 9 14 mm

67 mm 9 50 mm 67 mm 9 100 mm

References 1. P. J. Keall, V. R. Kini, S. S. Vedam, and R. Mohan, ‘‘Motion adaptive x-ray therapy: a feasibility study.,’’ Phys. Med. Biol., vol. 46, no. 1, pp. 1–10, 2001. 2. A. Schweikard, G. Glosser, M. Bodduluri, M. J. Murphy, and J. R. Adler, ‘‘Robotic motion compensation for respiratory movement during radiosurgery.,’’ Comput. aided Surg., vol. 5, no. 4, pp. 263–77, Jan. 2000. 3. P. J. Keall, E. Colvill, R. O’Brien, J. A. Ng, P. R. Poulsen, T. Eade, A. Kneebone, and J. T. Booth, ‘‘The first clinical implementation of electromagnetic transponder-guided MLC tracking.,’’ Med. Phys., vol. 41, no. 2, p. 020702, Feb. 2014. 4. R. Hansen, T. Ravkilde, E. S. Worm, J. Toftegaard, C. Grau, K. Macek, and P. R. Poulsen, ‘‘Electromagnetic guided couch and multileaf collimator tracking on a TrueBeam accelerator,’’ Med. Phys., vol. 43, no. 5, pp. 2387–2398, May 2016.

O006 First measurements of lung tumour motion using Kilovoltage Intrafraction Monitoring (KIM) C.-Y. Huang1, F. Hegi-Johnson1, D. T. Nguyen1, R. O’Brien1, K. Makhija1, C.-C. Shieh1, E. Hau2, R. Yeghiaian-Alvandi2, S. White2, J. Barber2, J. Luo2, S. Cross2, B. Ng3, K. Small2, P. Keall1 1

Faculty of Medicine, The University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Nepean Cancer Care Centre, NSW, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]),

123

Australas Phys Eng Sci Med ([email protected]). 3Respiratory Medicine, Nepean Hospital, NSW, Australia. ([email protected]) Introduction Kilovoltage Intrafraction Monitoring (KIM) has been successfully applied in sites with large cylindrical markers. However, for lung cancer, large markers will increase the risk of implant toxicity and cylindrical markers can migrate. In this study, we are acquiring data with goals of investigate KIM using small, easy to implant markers that anchor to the lungs in vivo, and develop KIM to enable real-time IGRT for lung cancer. Method A locally advanced stage III lung cancer patient undergoing conventionally fractionated VMAT was enrolled in an ethics-approved study of KIM. A Gold Anchor fiducial marker (0.4 mm diameter 9 20 mm length) was implanted in the tumour near the right hilum. kV images were acquired at 5.5 Hz during treatment. Posttreatment, markers were segmented and reconstructed to obtain 3D tumour trajectories. A Microsoft Kinect audio and depth sensing device was also mounted on the couch to get the external respiratory signal (Fig 1, left). Results KIM was successfully applied in the two imaging fractions delivered up to date. The fiducial marker was visible on 65.4% of the kV images (Fig 1, right). The average lung tumour motion (mean ± SD) in superior-inferior (SI), anterior-posterior (AP) leftright (LR), directions were 0.0 ± 2.1, 0.4 ± 1.5 and -0.1 ± 1.1 mm respectively. Four arcs of lung tumour 3D motion and two arcs of Kinect external signal were acquired, with the results illustrated in Fig 2.

Fig. 1 Treatment setup (left) and kV image of the Gold Anchor marker (right)

Fig. 2 Lung tumour motion during two fractions of radiotherapy treatment. The Kinect signal was only acquired for fraction 1 arc 1 and fraction 2 arc 2

123

Conclusion This is the first time that KIM has been used for intrafractional tumour motion monitoring during lung cancer radiotherapy, and also the first implementation of KIM on an Elekta imaging platform. This clinical translational research milestone paves the way for the broad implementation of image guidance to facilitate the detection and correction of geometric error for lung radiotherapy, and resultant improved clinical outcomes. References 1. Huang, C.-Y., et al., Six degrees-of-freedom prostate and lung tumor motion measurements using kilovoltage intrafraction monitoring. Int J Radiat Oncol Biol Phys, 2015. 91(2): p. 368–375. 2. Keall, P.J., et al., The first clinical treatment with kilovoltage intrafraction monitoring (KIM): A real-time image guidance method. Med Phys, 2015. 42(1): p. 354–358.

O007 MLC tracking enables dose reduction to OARs during lung cancer SABR V. Caillet1,2, N. Hardcastle1, K. Szymurah1, C. Haddad1, P.J. Keall2, J.T. Booth1,2 1

Northern Sydney Cancer Centre, Royal North Shore Hospital, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2School of Medicine, University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected])

Introduction Six lung Stereotactic Ablative Body Radiotherapy (SABR) patients have been successfully treated as part of the world’s first MLC tracking clinical trial (LIGHT SABR, NCT02514512). For each patient, three electromagnetic transponders are bronchoscopically implanted around the tumour. The delivered dose for real-time adaptive treatments will depend on the tumour motion so methods are required to estimate the delivered dose using measured tumour trajectory and dynalog files. This study presents reconstructed dose for the first lung SABR patients treated with MLC tracking. Method Six patients have been treated with MLC tracking on a Varian Trilogy linac. The MLC tracking plans are prescribed to a single phase GTV + 5 mm, and comparison ‘standard’ plans established prescribed to ITV + 5 mm. An isocentre shift method is utilised for reconstructing the delivered dose. Doses delivered to PTV, mean lung dose, lung D2%, spine and heart doses are compared between MLC tracking and standard plans. The dose that covers 100% of the target volume (GTV D100%) was taken as reference. Statistical differences were assessed using the Wilcoxon signed-rank test. Results Tumour motion was \4 mm for 4/6 patients. MLC-tracking resulted in reduced PTV volume and DVH metrics for all patients (Fig. 1). PTV volume reduction was 29% (9.6 cc, p \ 0.1). For the OARs and GTV D100% no statistical differences were found between regimes. MLD and D2% (spine and heart) reduction were 8.4% (0.22 Gy), 11.5% (0.75 Gy) and 3.1% (0.16 Gy). Tracking and standard D100 GTV were 99.0 and 97.9% of the prescribed dose (Fig. 2, p [ 0.1) suggesting superior dose coverage during tracking. Conclusion This study provides initial promising results in favour for MLC tracking. The current clinical trial aims for 20 patients so differences between regimes are expected to achieve more significant results.

Australas Phys Eng Sci Med 1

Northern Sydney Cancer Centre, Royal North Shore Hospital, NSW, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Radiation Physics Laboratory, University of Sydney, NSW, Australia. ([email protected]), ([email protected]). 3 Department of Respiratory Medicine, Royal North Shore Hospital, NSW, Australia. ([email protected])

Fig. 1 Dosimetric comparison between tracking and standard SABR for the (a) mean lung dose, (b) spine (D2%) and (c) heart (D2%). All patients benefited from MLC tracking. Higher than standard dose were observed in some cases. For the spine, patient 4 and 5 had in average both 0.6 Gy higher than standard (For patient 5, planned tracking dose equated delivered tracking dose). For the heart, patient 1 had 0.9 Gy higher than standard. Differences arise when the tumour hovers toward the OAR and increases the dose to them. More importantly, in no instances the dose delivered to the OAR exceeded protocol

Fig. 2 Box plot of the percentage dose coverage to the GTV D99, D95, D100 and D2 using the prescribed dose as the reference. The GTV dose coverage for MLC tracking shows dose coverage with better agreement than no tracking. The D2% for standard and tracking is higher than planned potentially due to interplay effect due tumour motion relative to the MLC

Introduction The LIGHT SABR clinical trial involves implantation of electromagnetic beacons surrounding thoracic targets treated with stereotactic ablative body radiotherapy. This facilitates realtime tracking of the tumour as it moves with respiration. Motion data is provided to the treatment beam by the beacons, which may not be moving in synchrony with the tumour. This study investigates the surrogacy error in tracking implanted beacons for lung tumours. Method Seven patients each had three beacons implanted endobronchially in the ipsilateral lung using fluoroscopic guidance. The patients received a simulation 4DCT in which the tumour motion was compared to that of the centre-of-mass (COM) of the beacons over the 10 phases. The surrogacy error was calculated as the difference in each phase between the tumour and beacon COM relative to the reference (exhale) phase. Prior to treatment at the first fraction, lateral and anterior-posterior fluoroscopy of the beacons and tumour was obtained for three breaths. The tumour and beacon COM position was determined in each frame from which probability distribution functions (PDFs) of the surrogacy error were created. From the these PDFs, the geometric margin to account for surrogacy error required to cover the tumour 95% of the time for each patient was computed.

Fig. 1 Example data (Patient 6) showing superior-inferior motion of the tumour and beacon COM and resultant surrogacy error probability distribution function

References 1. P.R. Poulsen, M.L. Schmidt, P. Keall, E.S. Worm, W. Fledelius, L. Hoffmann, A method of dose reconstruction for moving targets compatible with dynamic treatments, Medical physics, 39 (2012) 6237–6246.

O008 Surrogacy error in tumour tracking based on implanted fiducial markers in lung stereotactic ablative body radiotherapy N. Hardcastle1, V. Caillet1,2, B. Harris3, C. Haddad1, P. Keall2, J. Booth1

Fig. 2 Component of geometric uncertainty margin required to account for surrogacy for each patient and each direction

123

Australas Phys Eng Sci Med Results The maximum surrogacy error measured on the 4DCT ranged from 0.7–3.0 mm, 0.7–2.5 mm and 1.2–2.2 mm; on fluoroscopy 0.9–4.1 mm, 1.0–1.9 mm and 1.6–3.0 mm for left-right, anteriorposterior and superior-inferior respectively (Fig. 1). The component of the geometric uncertainty margin due to surrogacy error was up to 3 mm (Fig. 2). Conclusion Surrogacy error in real time tracking of tumours moving with respiration based on fiducial markers in the lung presents a significant geometric uncertainty. The magnitude of the surrogacy error varies between patients. Treatment margins should take into account surrogacy error on a patient specific basis. Acknowledgements This research is funded by Varian Medical Systems.

O009 Localization accuracy of PET/CT defined target volumes under respiratory motion for radiation therapy T. Osman1, S. Downes2, B. McBride3, A. Malaroda1, T. Hennessy3, P. Metcalfe1, A. Rozenfeld1 1

Centre of Medical Radiation Physics, Faculty of Engineering, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]). 2Radiation Oncology Department, Prince of Wales Hospital, Australia. ([email protected]). 3Nuclear Medicine Department, Prince of Wales Hospital, Australia. ([email protected]), ([email protected]) Introduction PET in radiation therapy allows the metabolic function of a tumour to be observed. Furthermore, the use of 4D-PET in conjunction with 4D-CT can improve tumour localization. However, the accuracy of localizing the tumour and defining a planning target volume needs to be assessed with each hospital’s imaging systems. The aim of this study was to assess the accuracy and hence determine the viability of 4D-PET/CT for radiation therapy treatment planning at Prince of Wales Hospital. Method Six 18F-FDG filled hollow spheres with radii ranging from 5.5 to 18 mm, were individually placed in the lung cavity of an anthropomorphic phantom. The phantom was placed on a moving platform designed to simulate patient respiration with respiration amplitude of 2.6 cm and frequency of 14 bpm. Following clinical workflow, 4D-PET and 4D-CT images were acquired on a Philips Ingenuity TF 128 PET/CT scanner using the Philips Bellows respiratory motion tracking system in the Nuclear Medicine department. A 4D-CT was then acquired on the Toshiba Aquilion Large Bore 90 cm using the Varian RPM 4D respiratory tracking system in the Radiation Oncology department. All acquisitions were gated into 10 bins per respiratory cycle. Lesion maximum intensity projections (MIPs) were generated from PET and CT images from the 2 scanners. The internal target volume (ITV) was defined via manual 3D region of interest placement around each lesion MIP. Concordance of ITVs were assessed by comparison between the 3 images as well as the predicted ITV based on sphere movement. Results/Conclusions Preliminary results showed that partial volume effects in PET will have a significant influence on the final ITV for smaller lesions. This has a direct consequence on the margin to be applied to the ITV when the planning target volume is defined for radiotherapy treatment. Full results for ITV concordance will be presented.

123

KS03 Innovations in anatomical modelling to advance image guided therapy Kristy K. Brock University of Michigan. ([email protected] [Presenting author]) The increasing focus on personalized and adaptive radiotherapy using information from both anatomical and functional imaging has placed significant demands on the ability to understand how tissues are changing during and following a course of high precision radiation therapy. This demand has expanded the role of traditional deformable image registration into the development of novel anatomical modeling techniques. Here the focus is shifted from making two images visually look the same to understanding the complex anatomical and physiological changes that are occurring at multiple levels of the tissue between the two states represented by the images. Two anatomical sites will be highlighted in the presentation, lung and liver. Anatomical modeling of the lung is a complex application with many exciting applications. Modeling the breathing motion of the lung is important for accurate assessment of the delivered dose due to both the motion as well as the heterogeneity of the tissue and the impact that this may have on the radiation beam, especially when the beam is highly modulated. However, expanding these models to describe the ventilation of the lung is of increasing interest as clinical trials investigate the ability to spare areas of functioning lung to reduce toxicities when delivering escalated doses. The use of biomechanical models and stress maps that result from these models will be demonstrated. In addition, during standard fractionation, data has shown that both the normal lung and the lung tumor often respond. Anatomical models have been developed to describe these changes and enable the safe adaptation of dose in these instances. The role of radiation in the treatment of both primary and metastatic cancer in the liver has increased. Functional imaging, such as dynamic contrast enhanced MRI, has been used to investigate the impact of radiation on normal liver. This can be critical for determining safe dose escalation strategies as well as in the retreatment setting. However, the complex response of the liver due to its regenerative capabilities places significant demands on anatomical modeling to enable accurate correlation of functional imaging with the delivered dose as well as mapping the delivered dose onto a subsequent planning scan. Novel biomechanical models will be presented to model these complex changes. Acknowledgements The author would like to acknowledge the work of Guillaume Cazoulat, Daniel Polan, Molly McCulloch, Martha Matuszak, Mary Feng, and Dawn Owen that contributed to the work highlighted in this abstract and presentation.

KS04 Individualised CT dosimetry Hilde Bosmans1, Xochitl Lopez Rendon1, Andreas Stratis1, Guozhi Zhang1 1

KU Leuven, Belgium. ([email protected] [Presenting author]) CT examinations contribute the largest part to the radiation exposure of medical imaging. Its quantification starts usually from measures such as CTDIvol or Dose-Length-Product (DLP). Radiation induced risk studies would typically require a first transition towards organ dose and subsequent application of life attributable risk factors. These

Australas Phys Eng Sci Med steps often use population and age averaged conversion factors. These days, large scale surveys are typically organized with an aim to investigate the (national) doses and start optimisation. Individualized dosimetry in radiological applications has another aim: quantify the potential burden to a specific patient, or optimize procedures for a specific patient group. In this presentation, we will explain how to obtain individualized doses using Monte Carlo techniques with a customized Monte Carlo framework and appropriate voxel models. This approach serves as gold standard. New concepts such as size specific dose estimate are a good approach towards these individualized doses. We will focus our study to dental cone beam CT applications and abdominal MultiSlice CT.

O010 Implementing a general X-ray quality control program: Early experience at a multi-campus public hospital A. Perdomo, S. Campaigne, S. Ravizza, H. Scoullar, G. Houston, Z. Brady Radiology Department, Alfred Health, Melbourne, Victoria. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction Alfred Health operates 20 general X-ray units of various manufacturers and models across two campuses and four subdepartments in Victoria for diagnostic imaging. A general X-ray quality control (QC) program was implemented to comply with the Royal Australian and New Zealand College of Radiologists (RANZCR) general X-ray quality assurance (QA) and QC guidelines as well as the manufacturers’ recommendations. Method The general X-ray units include four different manufacturers, a range of models and both fixed and mobile units. Both digital radiography (DR) and computed radiography (CR) detectors are available for every unit. Each of the recommended RANZCR and manufacturers’ items were initially assessed by a multi-disciplinary working group to see how useful the activity was, how long the activity took to complete, the recommended frequency and the financial impact. A general X-ray QC program was then established by medical physicists which includes daily, weekly and quarterly QC activities undertaken by radiographers. An electronic method for recording the QC results for each X-ray unit and each individual CR and DR detector was created. Results After a number of iterations, the general X-ray QC program was implemented across all sub-departments. The electronic records permit real time input and analysis by the radiographer and easy ongoing tracking by the physicist. A number of occupational health and safety (OHS), image quality and radiation dose issues have been identified, particularly during the first round of testing on the units. These included workstation monitor replacement, correcting filtration during detector calibration and correction of light beam alignment. Conclusion The implementation of the general X-ray QC program has allowed radiographers and physicists to identify issues enabling timely rectification and reducing potential clinical impact. This program has also created a collaborative working environment between professional groups. References 1. RANZCR General X-ray QA and QC Guideline, Version 1, November 2013

O011 An imaging quality control program for Spectral micro-CT N. Schleich1, R. Aamir2, S. M. Midgley3, M. Chen4, N. Anderson2, A. Butler2,5,6,7 1

Department of Radiation Therapy, University of Otago Wellington, New Zealand. ([email protected] [Presenting author]). 2 Department of Radiology, University of Otago, Christchurch, New Zealand. ([email protected]), ([email protected]). 3 School of Physics, Monash University, Melbourne, Australia. ([email protected]). 4University of Otago, Auckland, New Zealand. ([email protected]). 5Department of Electrical and Computer Engineering, University of Canterbury, Christchurch, New Zealand. 6MARS Bioimaging, Christchurch, New Zealand. 7European Organisation for Nuclear Research, Geneva, Switzerland. ([email protected]) Introduction Spectral computed tomography (CT) is a new imaging modality. In comparison to the X-ray detectors in conventional CT scanners, Spectral CT uses photon counting detectors that measure multiple energies simultaneously. Our NZ based research team is developing the MARS scanner, a small-bore Spectral CT system for pre-clinical use (Anderson et al. 2010; Butler et al 2011; Aamir et al. 2014; Anderson et al. 2014). Quality control (QC) testing is a vital link in the imaging chain, monitoring and quantifying the performance of each subsystem. Commercially available phantoms and QC tests used with medical CT scanners cannot be transferred to micro-CT scanners by simply downscaling their size. Instead, micro-CT requires specifically designed phantoms and protocols appropriately adapted to the task. In particular, the energy resolving capability of Spectral CT requires new tests that are not performed on conventional CT scanners. Method Imaging quality control methodologies for medical CT were adapted and test objects appropriate to micro-CT were designed and custom-built. Scanning protocols and analysis algorithms were developed for measurement of image noise and uniformity, high contrast spatial resolution, low contrast detail threshold and other parameters. Results MARS Spectral CT scans of the phantoms underwent 3D algebraic reconstruction delivering volumetric data in four to five energy windows 15–120 keV. QC results were used to establish suitable scan settings, data reconstruction parameters and analysis approaches, to acquire baselines and to perform constancy checks. Results obtained with the customized phantoms have assisted in rapid diagnosis of non-ideal behaviour, leading to improved image quality. Pre-clinical research scans performed during the development of the QC program also demonstrate image quality improvement. Conclusion Imaging QC for Spectral CT requires new procedures focussing on its energy resolving capabilities and the modification of established imaging tests. Phantoms and procedures suitable for Spectral micro-CT were developed and tested for this purpose. References 1. Anderson NG, Butler AP. Clinical applications of spectral molecular imaging: potential and challenges. Contrast Media Mol Imaging. 2014;9(1):3–12. 2. Aamir R, Chernoglazov A, Bateman CJ, et al. MARS spectral molecular imaging of lamb tissue: data collection and image analysis. JINST. 2014;9:P02005, 1–10. 3. Butler, APH, Butzer, J, Schleich, N, et al. Processing of spectral X-ray data with principal components analysis. Nucl Inst & Meth Phys Res A. 2011;633(Suppl.1):S140–2.

123

Australas Phys Eng Sci Med 4. Anderson NG, Butler AP, Scott N, et al. Spectroscopic (multienergy) CT distinguishes iodine and barium contrast material in mice. Eur Radiol. 2010;20:2126–34.

IS03 Australia’s role in the development of diagnostic ultrasound technology R. W. Gill Faculty of Medicine, UNSW. ([email protected] [Presenting author]) In 2004 Australia Post issued a postage stamp featuring a pregnant woman walking along the beach holding an ultrasound image. What was this all about? This was one of a series of five stamps honouring Australian innovations; other stamps featured well-known Australian inventions such as the Hills Hoist and the Black Box Flight Recorder. Many people do not realise that Australia made enormous contributions to the development of diagnostic ultrasound. Starting in 1959, just one year after Dr Ian Donald published his ground-breaking paper in The Lancet, the Ultrasonics Institute (UI) in Sydney became one of the world’s major research groups in the race to develop instrumentation suitable for clinical use. In its 30 years of operation the Institute pioneered many aspects of the technology, most importantly introducing grey-scale imaging in 1969. This was one of two essential developments (the other being real-time scanning) which enabled widespread adoption of ultrasound. UI’s collaborators and other pioneering Australian clinicians contributed to clinical applications ranging from eyes, hearts and breasts to obstetrics, paediatrics and the abdomen. In addition to research, UI and its colleagues encouraged and supported the development of a strong ultrasound profession. UI staff helped to found the Australian Society for Ultrasound in Medicine (ASUM) and the Australasian College of Physical Scientists & Engineers in Medicine (ACPSEM). The Institute also provided education and training for doctors and sonographers and organised a major world ultrasound congress in Sydney (WFUMB ’85) and the first World Congress of Sonographers. In this talk I will review some of Australia’s contributions to the development of ultrasound.

IS04 Current Australian Medical Ultrasound Research C. M. Langton Quantitative Ultrasound Imaging and Characterisation (QUIC) Research Group, Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, QLD. ([email protected]) Today, we use ultrasound for a broad range of routine clinical investigations; examples include monitoring the developing foetus in obstetrics, observing the movement of the heart and blood flow in cardiology, guiding a biopsy needle in oncology, and assessing musculoskeletal tissues following injury; but there could be other ‘left-field’ applications. Aimed at determining the dosimetric consequences of tumour movement during radiotherapy, we have created a robotic-ultrasound ‘proof of concept’ system. It is envisaged that by recording tumour movement, the actual dose delivered may be accurately estimated through Monte Carlo modelling, that would inform subsequent

123

fraction prescriptions. Complementary to this is the haptic robotic arm developed at Deakin University. Quantitative 3D imaging is being studied using two platforms. A ‘flat-bed’ scanner is being developed to characterise tissue breakdown in the diabetic foot, and to assess mammographic density. An ultrasound computed tomography scanner is being developed to characterise tissue breakdown associated with the post-amputation residuum, and to assess osteoporotic fracture risk. Noting that the initial medical utility of ultrasound was aimed at therapeutic interventions of the brain in the 1940’s; with a shift to diagnostic imaging from the 1960’s, it is interesting to note the renewed therapeutic interest. Inconsistencies in the composition and thickness of the skull however lead to variations in transit-time and significant wave degradation at a tissue target. The current ‘active’ solution is to vary transmission-delay, necessitating electronic control of each individual transducer element. A ‘passive’ 3D-printed twinlayer ‘ultrasound phase-interference compensator’ (UPIC) has recently been developed, aimed at providing constant transit-time. Another therapeutic ultrasound research project is aimed at local drug delivery through bursting of micro-capsules within a tissue target. Recognising the pioneers at the Ultrasonics Institute, I advocate that we continue to extend the utility boundaries of this ‘unsung hero’ of the medical diagnosis and treatment world.

IS05 Clinical impact of technology advances in ultrasound Lyndal Macpherson CEO of ASUM. ([email protected] [Presenting author]) Ultrasound continues to have a clinical impact on lives. It has many advantages compared to other imaging technologies. These include real time dynamic studies and bedside examinations and the fact that ultrasound remains relatively cost effective. Within the ever changing world of imaging, ultrasound continues to reinvent itself while remaining an essential part of the clinical care of our patients. In this talk I will look at some of the new developments and their clinical impact. How can we utilise the acoustic force radiation impulse to diagnose cancer and liver disease? Could we prevent disease utilising stiffness measures for the vessels, not just the heart? Can we model and measure the valves of the heart for greater accuracy in diagnosis and surgical intervention? Are there ways to create greater consistency across exams?

O012 Weekend work for radiotherapy medical physicists: A pilot project at Peter MacCallum Cancer Centre Tomas Kron, Kosta Gerontzos, Ajit Mullen, Li Zhu, Brad Schilling, Simon Castles, Peter Buchanan, Nilgun Touma, Felicity Topp Peter MacCallum Cancer Centre, Melbourne, Australia. ([email protected] [Presenting author]) Background Radiotherapy is one of the main treatment modalities for cancer patients and timely access to treatment is one of the main

Australas Phys Eng Sci Med

O013 The estimation of missing values in Lung Cancer radiotherapy medical records M. S. Barakat1, M. Field1, D. Thwaites2, A. Ghose1, D. Stirling1, A. Dekker3, M. Carolan4, S. Vinod5,6, L. Holloway7 1

Faculty of Engineering and Information Sciences, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected]). 3MAASTRO Clinic, Maastricht, The Netherland. ([email protected]). 4Illawarra and Shoalhaven Cancer Care Centres and Centre of Medical Radiation Physics, University of Wollongong. ([email protected]). 5Liverpool Cancer Therapy Centre, Liverpool Hospital, Australia. 6SWSCS, University of New South Wales Australia. ([email protected]). 7 Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute for Applied Medical Research, Institute of Medical Physics, University of Sydney, Centre of Medical Radiation Physics, University of Wollongong and SWSCS, University of New South Wales Australia. ([email protected]) Introduction Many medical records are incomplete or data is stored in a format which is difficult to access. Such a scenario was observed in our previous work . The estimation of such missing values from other data stored for the same patient can be used for epidemiology and decision support system work and could potentially be used where a given test can not be carried out for a given patient. The aim of this work was to investigate statistical methods for imputing missing values and evaluate their error.

0.25 0.2 Error

predictors for treatment outcome and patient satisfaction. Our institution embarked on a trial of performing Physics Quality Assurance for linear accelerators on Saturdays. Methods Monthly quality control (QC) activities at our institution follow the recommendations of the AAPM TG 142 Report. At our main site in East Melbourne QC for 6 linacs was performed from 7 a.m. to 9 a.m. one day a week each rotating through mechanical, dosimetric, beam parameters and imaging component checks. Commencing in June 2015 these activities were combined into one single Saturday shift as typical patient treatment hours are until 6:30 p.m. or longer which makes late shifts less attractive. Results Moving physics QC to weekends freed five hours for patient radiotherapy every week and allowed for better patient booking. Initially three physicists (at least one senior) were rostered on for weekend QC of half of the linacs alternating every fortnight. As weekend work results in time penalties and increased annual leave entitlements the change resulted in an effective loss of 0.3EFT. The impact was reduced by paying the senior staff member over-time. The trial also highlighted that combining four nominally two hour sessions did not result in time savings as documentation and training that previously had occurred after the two hour session in effect prolonged the time required for completing all tasks. This additional time and the need to interact with engineering staff for faults and QC after maintenance resulted in rostering an additional physicist for the weekend shift. Conclusion Moving physics QC activities out of normal working hours is possible. However, the workflow tested in our institution only works for relatively large departments and additional resources must be made available to implement the changes.

0.15

mean hot-deck

0.1

knn EM

0.05 0 5%

10%

15% 20% Missing Percentage

25%

30%

Fig. 1 Imputation error Method Data was extracted from Liverpool Cancer Therapy Centre oncology information systems. Clinical variables included FEV1, ECOG, age, gender and GTV. The data was filtered to remove records with a missing value for the clinical variables listed earlier. New datasets were generated by randomly deleting values from the ECOG and FEV1 variables. The new datasets have missing data percentages between 5 and 30%. The common Mean, hot-cold-deck, Knn and Expectation Maximization statistical imputation algorithms were used separately to impute the missing values. They were then evaluated by calculating the error as the average normalized difference between the imputed values and the actual values over the original complete dataset population sand. The whole process was repeated to reach statistical significance. Results The original complete dataset consisted of 108 records. Figure 1 shows the average errors for different datasets. The error for the hot-cold-deck method was significantly the highest. With 5% missing data, estimation resulted in a best-case scenario of 0.11 errors. With 15% or more missing data the error exceeds 0.15. Conclusion With small size data sets, statistical imputation of 5–10% of missing values can have a reasonable error margin of 0.11–0.13 and the impact the imputed values is being assessed for its impact on the modeling. References 1. A. Dekker, S. Vinod, L. Holloway, C. Oberije, A. George, G. Goozee, G. P. Delaney, P. Lambin, and D. Thwaites, ‘‘Rapid learning in practice: A lung cancer survival decision support system in routine patient care data,’’ Radiotherapy and Oncology, vol. 113, no. 1, pp. 47–53, Oct. 2014. 2. P. J. Garcı´a-Laencina, J.-L. Sancho-Go´mez, and A. R. FigueirasVidal, ‘‘Pattern classification with missing data: a review,’’ Neural Computing and Applications, vol. 19, no. 2, pp. 263–282, Sep. 2009

O014 OzCAT: The Australian computer aided theragnostics network L. C. Holloway1,2,3,4, M. Field5, M. S. Barakat5, S. Vinod3,6,7, G. Delaney3,6, M. Carolan2,8,9, M. Bailey8,9,10, A. Miller8,11, J. Sykes12,13, R. Alvandi12, E. Hau12, V. Ahern12, J. Lehman13,14, J. Ludbrook14, A. Ghose5, D. Stirling5, T. Lustberg15, J. Van Soest15, S. Walsh15, A. Dekker15, D. Thwaites16 1

Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute, Australia. 2Centre for Medical Radiation Physics, University

123

Australas Phys Eng Sci Med of Wollongong, Australia. 3South West Clinical School, University of New South Wales, Australia. ([email protected]), ([email protected]). 4School of Physics, University of Sydney, Australia. 5Faculty of Engineering and Information Sciences, University of Wollongong, Australia. 6Cancer Therapy Centre, Liverpool Hospital, Australia. 7Western Sydney University, NSW, Australia. 8Illawarra and Shoalhaven Cancer Care Centres. 9 Illawarra Health and Medical Research Institute, Wollongong, Australia. 10Centre for Oncology Informatics and Centre for Medical Radiation Physics, University of Wollongong, Wollongong, Australia. 11Centre for Oncology Informatics, University of Wollongong, Wollongong, Australia. 12Sydney West Cancer Network, NSW Australia. 13Institute of Medical Physics, University of Sydney, NSW, Australia. 14Calvary Mater Hospital Newcastle, NSW Australia. 15Department of Radiation Oncology (MAASTRO), GROW School for Oncology and Developmental Biology, Maastricht University Medical Centre+, Maastricht, The Netherlands. 16Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected] [Presenting author]) Introduction Radiotherapy treatment guidelines are based on randomised clinical trial (RCT) evidence. However, many patients are not eligible for RCTs (Fig. 1). The concept of a distributed data network where radiotherapy data remains at local centres but can be learnt from jointly has been presented by the MAASTRO clinic, with the ability to provide additional clinical evidence. In this work an assessment of the number of lung patients who do not meet trial criteria was determined and an initial distributed radiotherapy data network established across four NSW centres with the goal of improving available evidence for these patients. Method A non-small cell lung radiotherapy dataset of 298 patients was extracted from Liverpool and Macarthur Cancer Therapy Centres. The dataset was reviewed to determine the number of patients who met criteria for RTOG 9410. A local distributed learning network with an additional link to the MAASTRO clinic was established across Liverpool and Macarthur Cancer Therapy Centres, Illawarra and Shoalhaven Cancer Care Centres, Westmead Cancer Care Centre and Calvary Mater Newcastle radiotherapy department and lung radiotherapy data extracted at each centre. Results A maximum of only 23% patients within the non-small cell lung cancer cohort met RTOG 9410 criteria. Lung cancer radiotherapy data has been successfully extracted at the four radiotherapy centres and transfer of model learning parameters between centres has been demonstrated between centres and with MAASTRO. Inclusion of further centres in the learning network including both Australian and additional international centres is planned. An international community for computer aided theragnostics (CAT) is also in development.

Conclusion Many patients do not meet randomised clinical trial criteria on which radiotherapy treatment guidelines are based. An Australian distributed radiotherapy data network has been established with links internationally to enable generation of additional clinical evidence to support treatment decisions for such patients. Acknowledgements NSW Health bioinformatics ‘proof of concept’ grant.

MO01 Independent external validation of predictive models for urinary dysfunction following external beam radiotherapy of the prostate N. Yahya1, M. A. Ebert2,3, M. Bulsara4, A. Kennedy3, David J. Joseph3,5, James W. Denham6 1

School of Health Sciences, National University of Malaysia, Kuala Lumpur, Malaysia. ([email protected]). 2School of Physics, University of Western Australia, Western Australia, Australia. 3 Department of Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia, Australia. ([email protected] [Presenting author]). 4Institute for Health Research, University of Notre Dame, Fremantle, Western Australia, Australia. 5School of Surgery, University of Western Australia, Western Australia, Australia. 6School of Medicine and Public Health, University of Newcastle, New South Wales, Australia. Introduction Most predictive models are not sufficiently validated for prospective use. We performed independent external validation of published predictive models for urinary dysfunctions following radiotherapy of the prostate. Method Multivariable models developed to predict atomised and generalised urinary symptoms, both acute and late, were considered for validation using a dataset representing 754 participants from the TROG 03.04-RADAR trial. Harmonisation of endpoints and features were performed to match the predictive models. The overall performance, calibration and discrimination were assessed. Results 14 models from four publications were validated. The discrimination of the predictive models in an independent external validation cohort, measured using the area under the receiver operating characteristic (ROC) curve, ranged from 0.473 to 0.695, generally lower than in internal validation. 4 models had ROC [ 0.6. Shrinkage was required for all predictive models’ coefficients ranging from -0.309 (prediction probability was inverse to observed proportion) to 0.823. Predictive models which include baseline symptoms as a feature produced the highest discrimination. Two models produced a predicted probability of 0 and 1 for all patients. Conclusion Predictive models vary in performance and transferability illustrating the need for improvements in model development and reporting. Several models showed reasonable potential but efforts should be increased to improve performance. Acknowledgements This study was supported by NHMRC Grants 1006447 and 1077788.

MO02 Evaluation of multileaf collimator performance in high magnetic fringe fields for MRI-Linac development

Fig. 1 Many patients don’t meet RCT eligibility criteria

123

Bin Dong1,2, Kevin K. Zhang1,3, Lois Holloway1,2,3,4,5, Peter E. Metcalfe1,2, Paul J. Keall1,5, Gary P. Liney1,2,3,4

Australas Phys Eng Sci Med 1

Ingham Institute for Applied Medical Research, Australia. 2Centre for Medical Radiation Physics (CMPR), University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]). 3South Western Sydney Clinical School, University of New South Wales, Australia. ([email protected]), ([email protected]). 4Department of Medical Physics, Cancer Therapy Centre, Liverpool Hospital, Australia. ([email protected]). 5Radiation Physics Laboratory, Sydney Medical School, The University of Sydney, Australia. ([email protected])

Introduction A multi-leaf collimator (MLC) has permanent magnet motors/encoders which will be affected in the fringe field of MRI. This study evaluated the MLC performance over a range magnetic fields and investigated possible solutions to reduce the impact for MRI-Linac. Methods A Single Leaf Test Platform (SLTP) was built to simulate the MLC working conditions. This consisted of a motor/encoder assembly from a clinical MLC (Varian Millennium-120) driven by a single-board computer (Raspberry Pi2) and placed in the fringe field of an MRI scanner (Siemens Skyra, Fig. 1). Speed measurements were taken with the SLTP in two orientations with/without shielding of a steel box of approximately 0.5 mm thickness; 90 to B0 (in-line beam alignment) and 0 to B0 (perpendicular beam alignment). The motor was driven by 14 V 10 kHz Pulse Width Modulation (PWM) signal with high and low duty cycles and in clockwise and counterclockwise directions. The motor speed was calibrated from the encoder output and verified using an optical tachometer (Digitech QM1448). Results For 0 arrangement (Fig. 2), in slow mode, the motor speed reduced up to 40% as fringe field increased with the encoder was able to feedback true rotational speed, and there was a threshold (300G without shielding or 400G with shielding) when the motor experiences sufficient Lorentz force to fail. In fast mode, the motor speed increased up to 6.5% and the encoder failed above 400G. For 90 setup, no motor failure was observed up to 800G. However, an encoder discrepancy in fast mode was observed beyond 400G with little shielding improvement. Conclusion This study has shown some important variations and discrepancies in both the encoder and motor output of a clinical MLC when placed in a magnetic field. Results will be used to help designing the MLC for our subsequent prototype MR-Linac.

Fig. 2 Motor speed (counter clockwise) versus magnetic field for 0 setup with and without shielding References 1. J. Yun, J. St. Aubin, S. Rathee, and B. G. Fallone (2010) Brushed permanent magnet DC MLC motor operation in an external magnetic field. Med. Phys. Vol 37 No. 5, 2131–2134 2. P. J. Keall, M. Barton, and S. Crozier (2014). The Australian magnetic resonance imaging–linac program. Sem. Radiat. Oncol. Vol. 24, No. 3, 203–206

MO03 Workflow and software for MRI real-time guided radiotherapy K. Zhang1,2, X. Jia3, P. Keall4, G. Liney1,2,5,6 1

Ingham Institute for Applied Medical Research, Liverpool, NSW, Australia. 2South Western Sydney Clinical School, The University of New South Wales, Liverpool, NSW, Australia. ([email protected] [Presenting author]). 3 School of Engineer and Information Technology, The University of New South Wales, Canberra, ACT, Australia. ([email protected]). 4School of Physics, The University of Sydney, Sydney, Australia. ([email protected]). 5Department of Medical Physics, Cancer Therapy Centre, Liverpool Hospital, Liverpool, NSW, Australia. 6Centre for Medical Radiation Physics (CMPR), The University of Wollongong, Wollongong, NSW, Australia. ([email protected]) Fig. 1 Single leaf test platform in the magnetic field of a 3T scanner. The motor/encoder (90 set-up) is mounted on a moving bracket to accurately position in the fringe field

Introduction Concepts of image guided radiotherapy [1], especially MRI guided radiotherapy [2–4], are broadly studied in recent years.

123

Australas Phys Eng Sci Med Conclusion This study develops a workflow with software for MRI real-time guided radiotherapy. The under development software can be used in the MRI-Linac program. References

Fig. 1 Workflow of MRI real-time guided radiotherapy

Fig. 2 Software for MRI real-time guided radiotherapy MRI images are gratefully used for radiotherapy planning [5] supplementing CT [6] owing to their superior soft tissue contrast and non-ionizing radiation utilization compared with CT. However, geometric distortion [7] due to inhomogeneity of magnetic field and nonlinearity of gradient coil has to be considered. Studies [8–10] also show feasibility of real-time anatomy/cancer tracking. All the previous studies have proved the possibility of adaptive radiotherapy by real-time MRI assistance. This study develops a workflow with software for MRI real-time guided radiotherapy. Method The workflow (Fig. 1) starts from acquiring 3D volumetric images through pre-treatment MRI scan. A 3D volume sequence HASTE (TE/TR = 92/2000 ms, resolution = 1.25 9 1.25 9 4.0 mm) is first acquired during free breathing with navigation. The acquired images are used for cancer segmentation, template extraction and treatment plan generation. The segmentation can be automatic or manual with interactive User Interface (Fig. 2) support. A template is a texture map of cancer in a plane, which can be either orthogonal or oblique. A treatment plan is paired with each template. All plans are downloaded to treatment delivery unit for motion adaption. A real-time motion sequence is subsequently acquired using a rapid single-sagittal-plane 2D TrueFISP acquisition (TE/TR = 1.44/ 419 ms; in-plane resolution 0.70 9 0.70 mm and 4 mm slice thickness) at 2 Hz frame rate. By matching template[11] with real-time images, the motion is detected, hence the plan is adapted. Results The 3D and 2D MRI pulse sequences performed good image quality. The tracking tests with a lung cancer patient dataset produced accurate results. The workflow and software design were compact and applicable.

123

1. D.A. Jaffray, Image-guided radiotherapy: from current concept to future perspectives, Nature Reviews Clinical Oncology, 2012 Dec, Vol. 9(12), pp. 688–699. 2. B. W. Raaymakers, J. J. W. Lagendijk, J. Overweg, J. G. M. Kok, A. J. E. Raaijmakers, E. M. Kerkhof, R. W. van der Put, I. Meijsing, S. P. M. Crijns, F. Benedosso, M. van Vulpen, C. H. W. de Graaff, J. Allen and K. J. Brown, Integrating a 1.5T MRI scanner with a 6 MV accelerator: proof of concept, Physics in Medicine and Biology, 54 (2009), 229–237. 3. B. G. Fallone, B. Murray, S. Rathee, T. Stanescu, S. Steciw, S. Vidakovic, E. Blosser, and D. Tymofichuk, First MR images obtained during megavoltage photon irradiation from a prototype integrated linac-MR system, Medical Physics, 36(6) (2009), 2084–2088. 4. P. Keall, M. Barton, and S. Crozier, The Australian Magnetic Resonance Imaging-Linac Program. Seminars in Radiation Oncology, 24(3) (2014), 203–206. 5. M. A Schmidt, and G. S Payne, Radiotherapy planning using MRI, Physics in Medicine and Biology 60 (2015) R323–R361. 6. G. C. Pereira, M. Traughber, and R. F. Muzic Jr., The Role of Imaging in Radiation Therapy Planning: Past, Present, and Future, BieMed Research International, vol. 2014, Article ID 231090, 9 pages. 7. L. N. Baldwin, K. Wachowicz, S. D. Thomas, R. Rivest, and B. G. Fallone, Characterization, prediction, and correction of geometric distortion in 3T MR images, Medical Physics, 34(2), February 2007, 388–399. 8. L. Brix, S. Ringgaard, T. S. Sorensen, and P. R. Poulsen, Threedimensional liver motion tracking using real-time two-dimensional MRI, Medical Physics, 41(4) (2014), 042302 (10 pp). 9. L. I. Cervino, J. Du, and S. B. Jiang, MRI-guided tumor tracking in lung cancer radiotherapy, Physics in Medicine and Biology, 56 (2011), 3773–3785. 10. K. Zhang, S. Kumar, R. Rai, A. George, B. Dong, G.P. Liney, Three-dimensional lung tumour motion tracking using an advanced template matching technique: Texture Reformatted Angle Correlation (TRAC), ISMRM, 24 (2016). 11. L. Ding, A. Goshtasby, and M. Satter, Volume image registration by template matching, Image and Vision Computing, 19 (2001), 821–832.

MO04 An ideal detector for the quality assurance of stereotactic brain metastases E. K. Inness1, L. K. Webb1, P. H. Charles1,2 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected]), ([email protected] [Presenting author]). 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected]) Introduction Point dose measurements are commonly used as a way of verifying patient specific doses prior to treatment delivery. With an increasing interest in stereotactic radiotherapy (SRT), the choice of

Australas Phys Eng Sci Med detector in reducing inherent problems associated with small field dosimetry has become a topic of consideration. This study aims to assess the dose differences between three detectors, namely the Electron diode, Exradin W1 scintillator (PSD), and a Pinpoint ionisation chamber, in 11 SRT Brain metastases patients. Method A total of 43, 6 MV, non-conformal arcs were delivered to three detectors in a Lucy 3D QA Phantom, on an Elekta Axesse with a Beam modulator collimator. Each detector was setup to the isocentre using Exactrac, and oriented parallel to the radiation beam. All of the detectors were independently calibrated in plastic water under reference conditions, and the PSD measurements were corrected for Cerenkov radiation using the method defined in the user manual. All of the measured doses were compared to those calculated in BrainLAB iPlan version 3.0.0. Results For the range of field sizes measured, 8–24 mm, the average agreement with the TPS was found to be 1.5% for the PSD, with the diode measuring consistently high, and the pinpoint chamber, consistently low. When considering field sizes \15 mm, the pinpoint and diode gave an average agreement to within 3.5, and 4%, respectively; the worst result being -14.9% for a 10 mm field size using the pinpoint, and 6.3%, for an 11 mm field size, using the diode. For small fields, \15 mm, the PSD results continued to be within 1.5% of the average for all field sizes. Conclusion When measuring patient specific point doses for small fields, it is recommended to use a small, water equivalent detector, such as the Exradin W1 scintillator, in order to minimise volume averaging and mass-density chamber effects.

MO05 A Simple and fast method to measure attenuation caused by treatment couch and immobilization devices Omemh Bawazeer1,2, Sisira Herath3, Siva Sarasanandarajah1,3,4, Tomas Kron3,4, Leon Dunn6, Pradip Deb1 1

Discipline of Medical Radiations, RMIT University, Australia. Discipline of Sciences, Umm Al-Qura University, Saudi Arabia. ([email protected] [Presenting author]). 3Department of Physical Sciences, Peter MacCallum Cancer Centre, Australia, Discipline of Medical Radiations, RMIT University, Australia. 4 Department of Medical Imaging and Radiation Sciences, Monash University, Australia. 5Sir Peter MacCallum Cancer Institute, University of Melbourne, Australia. 6Epworth Radiation Oncology, Melbourne, Australia 2

Introduction With the increasing requirement of including attenuation arising from treatment couch and Immobilization devices in treatment planning system or monitor unit calculations, the objective of this study is to propose a simple and fast method to measure attenuation using electronic portal imaging devices (EPID) without a complex procedure to convert the signal to dose. Method A aS500 EPID attached to a Varian linear accelerator (21iX) was irradiated with delivering 100 MU using 6 MV and 600 MU/min at source to EPID distance 150 cm. Exact and IGRT Varian couches were positioned at 110 cm. Integrated images without and with treatment couch, belly board and head and neck baseboard were acquired for jaw and MLC defined fields at gantry angles of 0, 30, 60. Some measurements were repeated with the presence of 20 cm

Fig. 1 Data points are averages of mean values at the centre of attenuated images for four jaw defined field sizes, all measurement at source to couch distance 110 cm, error bars represent the deviation between the mean values for four field sizes

Fig. 2 Attenuation through the exact couch as a function of field size and presence of phantom, with phantom (red) and without (black), measurements for jaw defined field with gantry angle zero and source to couch distance 110 cm

solid water phantom and with changing couch position to 120 cm. Attenuated 2D images were assessed as the percentage differences between images with and without attenuation. Results Attenuation data are shown in Fig. 1. The highest attenuation was observed with the combination of the couch and head and neck baseboard 9.8 ± 0.41%. Interestingly, attenuation showing a reduction in the presence of the phantom approximately from 3.5 to 2.2% for large field size, see Fig. 2. The changing of couch position to 120 cm also reduced the attenuation. Conclusion A simplified method of measuring attenuation using aS500 EPID is proposed. Unlike the conventional approach, this approach is not time consuming and provides attenuation data in the center as well as the entire field. These results could also be useful to quantify the effect of attenuation arising from treatment couch and immobilization devices on EPID images when EPID is used as transit dosimetry. References 1. Olch, A.J., L. Gerig, H. Li, I. Mihaylov, and A. Morgan, Dosimetric effects caused by couch tops and immobilization devices: report of AAPM Task Group 176. Medical physics, 2014. 41(6): p. 061501. 2. Li, H., A.K. Lee, J.L. Johnson, R.X. Zhu, and R.J. Kudchadker, Characterization of dose impact on IMRT and VMAT from couch attenuation for two Varian couches. Journal of Applied Clinical Medical Physics, 2011. 12(3).

123

Australas Phys Eng Sci Med

MO06 Characterization of a novel Dual detection system designed for Simultaneous dose verification and imaging assessment in Intensity-Modulated Radiation Therapy

MO07 Ion chamber dosimetry within lung equivalent media

S. Alhujaili1, S. Deshpande2, O. Brace1, M. Petasecca1, P. Vial2,3, L. Holloway1,2,3, A. Rosenfeld1, P. Metcalfe1

1

1

Centre for Medical Radiation Physics, University of Wollongong, Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centres. ([email protected]). 3Institute of Medical Physics, University of Sydney, Sydney, Australia. ([email protected]), ([email protected])

Purpose To investigate a novel dual detection system that consists of a transparent 2D silicon Array detector (called Magic Plate, MP) coupled to an EPID to work simultaneously as a dose verification device (MP) and imaging co-registration (EPID). Method In this novel system, MP is placed directly above the EPID, positioned at 150 cm from LINAC source. Response of MP was measured in term of Output Factor and Beam Profile using 6 MV photon beam with SW phantom of 30 m3 dimensions placed at couch and SW build-up overlayer of 5, 10 and 15 cm. The MP measurements were compared to 2D ionisation chamber array (ICA) measurements and the PinnacleTM treatment planning system (TPS). A dosimetric performance of several clinical IMRT beams have been evaluated using !-evaluation method. The effect of the MP on the quality of the EPID Imaging has been evaluated by measuring the contrast-to-noise ratio (CNR) and spatial resolution. Images of a Rando phantom were used for qualitative assessment. Results The beam profile and output factor measurement of the MP with 15 cm build-up exhibits good agreement with ICA and TPS with a difference not exceeding 2%. Clinical IMRT beams had gamma pass rates of C95% at 3%/3 mm criteria. When the dual detection system was used with build-up 5 and 15 cm the CNR and spatial resolution (f50) were 264.96, 210.6, and 0.41, 0.40 respectively but when the EPID was used only the CNR and spatial resolution (f50) were 643.9 and 0.41 (Fig. 1). Conclusion The dual detector approach allows using EPID for MV X-ray imaging while Magic Plate for 2D dosimetry only that will optimize 3D dose reconstruction during delivery. The presence of MP above the EPID doesn’t perturb the resolution of the images acquired.

T. Pham1, P. H. Charles1,2 Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected]). 2Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]) Introduction Disequilibrium dosimetry is known to be complicated, particularly when the dosimeter differs in density from the medium in which the dose is being measured. Electronic disequilibrium occurs more readily in low density materials such as lung. The aim of this study is to quantify the perturbation effects of using a small ionisation chamber in lung equivalent media as a function of field size. Method The perturbation factors were determined for the PTW31014 PinPoint chamber using the egs_chamber Monte Carlo code. This detector was simulated within a slab geometry phantom (2 cm of water, 13 cm of lung, 3 cm of water [beam direction]), at a physical depth of 8.5 cm (i.e. within lung); 100 cm from the source. Two sources (approximate 6MV and 10MV beams); and 5 square field sizes (side lengths from 2 to 10 cm) were simulated. The total perturbation factor within lung was found by comparing the dose simulated within the active volume of the chamber to a detectorless simulation (dose scored to 1 mm3 cube of lung). Individual perturbation factors, quantifying the effects of the detector cavity, wall, stem, and central electrode were simulated by progressively adding these components. Results The total perturbation factor was close to unity for the 10 cm field, however decreased to 0.87 and 0.85 respectively for the 6 and 10 MV beams (field size of 2 cm). The plastic wall had the largest effect (e.g. 8% over-response for the 3 cm field; 10MV beam), followed by the stem (3.5% over-response for the 3 cm field; 10MV beam). Conclusion The PTW31014 PinPoint has a strong over-response in conditions of electronic disequilibrium. This should be corrected for when performing lung dosimetry with this ion chamber. The lower density of lung exacerbates the over-response of high density components.

MO08 Assessment of dose variation for accelerated partial breast irradiation using rigid and deformable image registrations V Batumalai1,2,3, L Holloway1,2,3,4,5, A Walker1,2,4, M Jameson1,2,4, G P Delaney1,2,3 1

Liverpool and Macarthur Cancer Therapy Centres, NSW Australia. Ingham Institute of Applied Medical Research, NSW Australia. 3 South Western Clinical School, University of New South Wales, NSW Australia. ([email protected] [Presenting author]), ([email protected]). 4Centre for Medical Radiation Physics, University of Wollongong, NSW Australia. ([email protected]), ([email protected]). 5Institute of Medical Physics, School of Physics, University of Sydney, NSW Australia. ([email protected]) 2

Fig. 1 Beam profile and output factor of MP compared with ICA and TPS

123

Introduction Setup uncertainties are particularly relevant when highly conformal plans such as external beam accelerated partial

Australas Phys Eng Sci Med

MO09 Impact of small changes in linear accelerator photon energy on beam quality: A Monte Carlo investigation D. M. Hossain, J. Mathew, L. Marsh, T. Kron

Fig. 1 Average dose-volume histogram for all patients comparing the original plan with error sources from bone, soft-tissue and deformable registration. *Statistically significant deviation from the original plan

Department of Physical Sciences, Peter MacCallum Cancer Centre, Melbourne, Victoria. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected])

Fig. 2 Dose volume histogram plots of one of the patient showing target volumes and organs at risk doses

Introduction The final energy of photons and electrons produced by modern medical linacs is defined by the magnetic field of the bending magnet. The bending magnet current for a specific energy could alter for different reasons which can result in changes to beam quality. In this study we have done Monte Carlo simulations to investigate small changes in energies for a Varian linac and their impact on beam qualities. Method In our previous study [1], a photon beam of nominal energy 6MV was modelled and validated for a Varian linac using BEAMnrc and DOSXYZnrc Monte Carlo code. The input file of the verified model was replicated for 5 different electron incident kinetic energies ranging from 5.9 to 6.1 MeV with intervals of 0.05 MeV. Simulations were performed for each energy and their phase space files were stored at 100 cm SSD. Those phase space files were used to investigate changes in spectral distribution, mean energy and PDDs in a water phantom. DOSXYZnrc software was used to calculate the PDDs. For each of the energy changes the D20,10 and the corresponding TPR20,10 was calculated from the PDDs using the empirical equation of TRS398 [2].

breast irradiation (APBI) is being considered. The study aimed to estimate the delivered dose to the target and organs at risk (OAR) for APBI accounting for setup uncertainties, using rigid and deformable image registration (DIR). Method One planning computed tomography (CT) scan and five cone-beam CT (CBCT) scans for each of 25 patients were used. All CBCT scans were registered to the planning-CT scan using three techniques; (i) rigid registration based on bony-anatomy only, (ii) rigid registration based on soft-tissue only, and (iii) deformable image registration. For each patient, four dose distributions were calculated for APBI. The first was the original plan, while the other three were ‘‘dose-of-the-day’’ for each of the registration approaches. The effects of image registrations on estimating delivered dose to targets and organs at risk were determined. Results The average reductions in V95 (percentage of the PTV that received 95% of the prescribed dose) were 6, 7, and 5% for bone, softtissue and DIR, respectively (Fig. 1). The average increase in mean dose to the heart were 9, 9 and 18% for bone, soft-tissue and DIR, respectively, while the average increase in maximum dose to the contralateral breast were 19, 20, and 28%, respectively. Figure 2 shows the dose-volume histogram for one of the patients showing the differences between the four plans. Conclusion The results of this study have shown that there are differences between the planned and estimated delivered dose for APBI due to setup uncertainties which may need to be accounted for. Estimated dosimetric impact of setup variation and breast deformation assessed using DIR was greater for OARs and smaller for target volumes compared to rigid registration. Acknowledgements The authors acknowledge funding assistance from the Ingham Institute Breast Cancer Grant (2014).

Results A linear trend for spectral distribution and for mean energy was observed with changing energy. The TPR20,10 for the 5 different KEs were calculated from their corresponding D20,10 and those TPR20,10 values are within 0.663 ± 0.005. The TPR20,10 for 6 MeV was relatively higher than for the other electron KEs. Conclusion There are some small changes in D10 and D20 with respect to the small changes of the beam energy. From the TPR20,10 results it can be concluded that changes in TPR20,10 from changes of ±1% in energy for a nominally 6 MV photon beam are not significant.

123

Australas Phys Eng Sci Med References

References

1. HM Deloar et al. IFMBE Proceedings Volume 14, 2007, pp 1956–1957. 2. TRS 398, IAEA Vienna, June 2006, pp 62.

1. J. Chojnowski, ‘A novel automated method of linac source alignment QA’, Combined Scientific Meeting, Sep. 2014, Melbourne, Australia.

MO10 Assessment of radiation source positions for a linac with multiple photon energies using EPID

MO11 A Comparison of a TrueBeam LINAC output using the machine performance check and the daily QA3

Jacek Chonjnowski1,2, Dr. Jonathan Skyes2,3, Shan Yau3, Prof David Thwaites2

A. Espinoza1, J. Sykes1, J. Barber1, E. Sullivan1, F. Q. Chen2, S Yau1,2,3

1

1

Mid North Coast Cancer Care Institute, Coffs Harbour, Australia. Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected] [Presenting author]). 3Sydney West Cancer Network, Australia.

2

Introduction Assessment of radiation source positions for a linac with multiple photon energies is important for advanced treatment techniques such as SRS/SBRT. High dose rate FFF beams have advantages for such treatments. However, image guided quality assurance tests are generally only performed with one regularly used photon energy i.e. 6X. Therefore misalignments of radiation source positions between different energies might contribute to the overall geometrical treatment accuracy. Method The radiation source position errors for each available energy on a Varian Truebeam machine at the Crown Princess Mary Cancer Care Centre (Westmead, Australia) are measured using a novel EPID-based methodology, where images acquired for 4 predefined fields are analysed using an in-house developed software (as presented at CSM20141). Results The results of measured radiation source positions are illustrated in Fig. 1. It was observed that the radiation source for 10XFFF energy is slightly misaligned by 0.3 mm with the reference energy 6X, normally used for image guided quality assurance. However, the radiation source for 10XFFF energy shows the minimal total (radial) misalignment with the reference mechanical isocentre (0.15 mm), when compared to radiation source offsets for other energies (70.3 mm). Conclusion The presented method of assessment of radiation source positions for a linac with multiple photon energies is user-independent, reproducible, easy and quick to execute. It should be part of a regular image guided quality assurance program.

Varian Truebeam®

Introduction Quality assurance (QA) of a modern linac is critical and time consuming. The Varian TrueBeam linac system features the Machine Performance Check (MPC) software, a series of self-checks performed using the kilovoltage and megavoltage imagers. MPC performs an automated assessment of beam properties and geometry within a run time of less than 5 min. There are few studies assessing the accuracy and long term performance of the MPC and none investigating the use for electron beams. Blacktown Hospital recently commissioned two linacs and the MPC measured beam output was compared to the Sun Nuclear DailyQA3 (DQA3) device over a period of 4 months. Method Daily machine outputs for one TrueBeam were measured for photon energies 6, 10 and 18 MV, flattening filter free (FFF) 6 and 10 MV, and for electron energies 6, 9, 12 and 15 MeV using both the MPC and the DQA3. MPC and DQA3 data were assessed in terms of day to day variation and for evidence of a correlation between the MPC and DQA3 results. Results The overall sample variance of the change in daily output for MPC was \0.1% for all beam energies except 6 MeV which was prone to random large fluctuations; and for DQA3 \0.2% for all beam energies. The Pearson’s correlation coefficient (r) for all energies, Table 1, ranged 0.55–0.86 over first two months and following a beam adjustment, ranged 0.54–81.

Table 1 Pearson’s correlation coefficient between MPC and DQA3 (p \ 0.01 for all energies)

0.4

Source Posion Offset [mm]

Blacktown Cancer & Haematology Centre, Blacktown Hospital. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Crown Princess Mary Cancer Centre, Westmead Hospital. ([email protected]). 3 Nepean Cancer Care Centre, Nepean Hospital.

0.3

Beam energy

0.2 0.1

3rd of Feb–8th of Apr 18th of Apr–6th of Jun Pearson’s r Pearson’s r

0 Photons

-0.1 -0.2 -0.3 -0.4

6 MV

0.82 (n = 30)

0.54 (n = 33)

6 FFF

0.76 (n = 30)

0.60 (n = 34)

10 MV

0.86 (n = 31)

0.68 (n = 34)

10 FFF

0.82 (n = 30)

0.77 (n = 34)

6X

6XFFF

10X

10XFFF

18X

Crossplane

-0.31

-0.36

-0.32

-0.07

-0.33

18 MV

0.83 (n = 31)

0.75 (n = 34)

Inplane

0.01

0.04

0.02

-0.14

0.01

Electrons 6 MeV

0.55 (n = 30)

0.59 (n = 33)

Total

0.31

0.36

0.32

0.16

0.33

Fig. 1 An example of measured energy dependent radiation source offset in crossplane and inplane directions for one linac

123

9 MeV

0.89 (n = 31)

0.72 (n = 33)

12 MeV

0.79 (n = 31)

0.79 (n = 33)

15 MeV

0.85 (n = 31)

0.81 (n = 33)

Australas Phys Eng Sci Med Conclusion The MPC measured output difference was found to have an overall strong correlation to the DQA3 across all measured energies with a lower day to day variation. Both methods require careful setting of baselines. While the error detection ability is yet to be assessed, MPC allows for a quick and easy beam assessment.

MO12 Investigating the factors contributing to dose variation in kilovoltage treatments of large nasal tumours E. C. Cosgriff, J. Foo, S. White Nepean Cancer Care Centre, Nepean Hospital, Kingswood, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Kilovoltage radiotherapy is a common treatment option for basal and squamous cell carcinomas located in the nasal region. Dosimetry in this region can prove particularly challenging because of significant variations in patient contour and local tissue heterogeneities. There is an absence of side-scatter material on either side of the nose, and SSD changes mean the dose distribution across the entire nose is nonuniform. A wax compensator can be used to improve uniformity, but other separate issues such as the percentage depth dose and delivered surface dose still need to be considered. Method Film measurements were performed on a Gulmay D3225 kilovoltage X-ray therapy unit using a solid water representation of a nose with and without side-scatter material. Patient-specific measurements were also performed using a plaster mould of the patient’s face and the filter and applicator specified by the treatment plan (Fig. 1). Radiochromic film was irradiated for a prescribed number of monitor units with or without the use of a wax compensator. A set of calibration films were collected for each case using the reference cone (10 cm diameter, 20 cm SSD). Films were scanned using an Epson flatbed scanner at a resolution of 50 dpi and analysed with ImageJ. Results Measured dose profiles revealed an underdose of 10–20% mainly due to a lack of side scatter. A wax compensator improved uniformity by around 10%. The lack of side scatter and wax compensator also affect the percentage depth dose.

Fig. 1 Photograph of the setup for the patient-specific measurements

Conclusion The lack of side scatter and nonuniformity across the field need to be taken into account in the monitor unit calculations for kilovoltage nose treatments; otherwise, there could be a detrimental effect on tumour control.

MO13 A non-destructive filmless approach to gynaecological applicator commissioning and reproducibility analysis M. Hanlon1, R. L. Smith1,2, C. Dempsey3,4, J. L. Millar2,5, R. D. Franich1 1

School of Science, RMIT University, Melbourne, VIC, Australia. ([email protected] [Presenting author]), ([email protected]). 2Alfred Health Radiation Oncology, The Alfred Hospital, Melbourne, VIC, Australia. ([email protected]). 3Department of Radiation Oncology, Calvary Mater Newcastle Hospital, NSW, Australia. 4Department of Radiation Oncology, University of Washington, Seattle, WA, USA. ([email protected]). 5School of Applied Sciences, RMIT University, Melbourne, VIC, Australia. ([email protected]) Introduction Commissioning and quality assurance of HDR Brachytherapy applicators requires accurate confirmation of true source dwell positions relative to applicator libraries. Existing approaches are either imprecise (using radiochromic films) or destructive (cutting applicator in half for visual observation.1 We present a method allowing the source to be tracked inside the applicator without damage to the applicator. Method A 30 mm diameter intrauterine ring applicator (Nucletron, The Netherlands) was disassembled and placed on the face of our flat panel detector (FPD) source tracking system. The source was delivered to the 34 distal dwell positions. At each dwell positron, an autoradiograph was taken, and an external x-ray tube was used to capture a dual-exposure image (with both the x-ray tube and HDR source

Fig. 1 Comparison between the measured source positions via two methods and the TPS applicator library

123

Australas Phys Eng Sci Med radiation contributing to the image). The difference between these allows the active source to be visualised. Reproducibility was assessed by repeated deliveries, and the measured source positions were compared to the TPS applicator library to confirm coincidence. Results The source was delivered three times. For each dwell position, the mean deviation from the centroid of the three measurements was 0.2 mm (r = 0.1 mm), showing that the source position was reproducible. Imaging also allowed the source cable movements to be seen, and showed that consequential irregularly spaced dwell positions were also reproducible. The dwell positions measured via the subtraction-imaging method were shown to be coincident with the TPS applicator library, with a mean deviation of 0.4 mm (r = 0.2 mm). The auto-radiograph source-tracking method agreed within a mean of 0.5 mm (r = 0.3 mm) (Fig. 1). Conclusion Our filmless approach allows for the commissioning and accurate characterisation of every dwell position in each applicator without damage. The reproducibility of the HDR source delivery can be confirmed, and dwell positions can be compared to the applicator library to confirm coincidence. References 1. Humer I, Kirisits C, Berger D, Trnkova´ P, Po¨tter R, Nesvacil N (2015) Improved source path localisation in ring applicators and the clinical impact for gynecological brachytherapy. J Contemp Brachytherapy. 7(2):

MO14 Detecting and managing vulnerabilities in medical software: Pinnacle3Ò scripting as a case study

Fig. 1 Software QA process described for commercial (dashed) and in-house software (solid) based on IEEE reference (blue), FMEA (yellow), and TRL (green)

Table 1 Mankins Technical Readiness Level [6] as adopted for inhouse medical software and customised for radiotherapy Technical readiness levels (Mankins) adopted for medical technology

In-house software release

1

Basic principles observed and reported

Pre-Alpha

2

Technology concept and application formulated

Pre-Alpha

3

Theoretical or experimental validation of critical functions

Pre-Alpha

4

Non-clinical testing of subsystems with phantom-equivalent data

Alpha

5

Non-clinical testing of subsystems with patient-equivalent data

Alpha

6

Non-clinical end-to-end system testing

Alpha

7

Clinical system testing with patients

Beta

8

Clinical system use with after commissioning and QA

Released

9

Clinical system use with proven results in operation and research

Released

J. Yuen, A. Ralston St George Hospital Cancer Care Centre, Sydney Australia. ([email protected] [Presenting author]), ([email protected]) Introduction Determining an appropriate amount of testing is difficult as it is impossible to test all possible conditions and situations [1, 2]. Guidelines exist that relate to the QA of commercial medical software or medical hardware reliant on software however there may be residual software vulnerabilities. The aim is to minimise risks of medical software using a system based on failure mode effects analysis (FMEA) recommended by TG-100 [3]. This system also applies to in-house software, with variations in systematic testing and design and specific testing unique to the medical physics community. Method Figure 1 illustrates the guideline which is based on FMEA and the IEEE Software Engineering Body of Knowledge [4]. Particular notes include: • • • •

•

Commercial and in-house software share similar QA components Testing of commercial software subsystems may be required if inhouse software has such dependencies In-house software involves unique steps such as software requirement validation Risk profile numbers (RPN) are based on multiplication of severity, occurrence, and detectability which scale tables available in the literature Software testing involves RPN which modulates testing required [5] by testers and test-time allocated, with increasing cases/conditions tested increasing trustworthiness.

123

•

Consideration of technology readiness level (TRL) [6] to utilise the software as either alpha, beta, or full clinical release (Table 1).

Results This methodology was demonstrated with a Pinnacle3 treatment planning system couch script, including: 1. Insufficient volume of planning CT: addresses invalid contours outside CT dataset. 2. Abutting couch structures: addresses couch override density order vulnerability.

Australas Phys Eng Sci Med 3. Couch contour discontinuities: prevents homogeneous calculation treating couch interior with incorrect density. 4. Prone and supine scans: prevents couch flip with scan orientation variation. 5. Quality assurance based on RPNs. Conclusion The aim of developing a versatile FMEA-based system to manage the risks of medical software vulnerabilities was achieved for both commercial and in-house software. References 1. 1. Steinke, G., V. Kurniawati, and J. Nindel-Edwards (2010), Integrating Failure Mode Effect Analysis into the Medical Device Approval Process: Communications of the IIMA, v. 10. 2. 2. Parnas, D. L., A. J. van Schouwen, and S. P. Kwan (1990), Evaluation of safety-critical software: Communications of the ACM, v. 33, p. 636–648. 3. 3. Huq, M. S., B. A. Fraass, P. B. Dunscombe, G. J. J. P., G. S. Ibbott, A. J. Mundt, S. Mutic, J. R. Palta, F. Rath, B. R. Thomadsen, J. F. Williamson, and E. D. Yorke (2016), The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management: Medical Physics. 4. 4. Bourque, P., and R. E. Fairley (2014), Guide to the Software Engineering Body of Knowledge (SWEBOK (R)): Version 3.0, IEEE Computer Society Press. 5. 5. Felderer, M., C. Haisjackl, V. Pekar, and R. Breu (2014), A risk assessment framework for software testing, Leveraging Applications of Formal Methods, Verification and Validation. Specialized Techniques and Applications, Springer, p. 292–308. 6. 6. Mankins, J. C. (2009), Technology readiness and risk assessments: A new approach: Acta Astronautica, v. 65, p. 1208–1215.

MO15 Free breathing lung MRI protocol at 3T Shivani Kumar1,2,3, Robba Rai2, Daniel Moses4,5, Callie Choong B2, Lois Holloway1,2,3,6,7,8, Shalini Kavita Vinod1,2,8, Gary Liney1,2,3,7 1

South Western Clinical School, School of Medicine, University of New South Wales, NSW, Australia. 2Liverpool and Macarthur Cancer Therapy Centres, Liverpool Hospital, NSW, Australia. 3 Ingham Institute of Applied Medical Research, NSW, Australia. ([email protected] [Presenting author]). 4Prince of Wales Clinical School, University of New South Wales, NSW, Australia. 5Department of Medical Imaging, Prince of Wales Hospital, NSW, Australia. 6Institute of Medical Physics, School of Physics, University of Sydney, NSW, Australia. 7Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. 8Western Sydney University, NSW, Australia. Introduction Magnetic resonance imaging (MRI) can provide morphological and functional information and is increasingly being utilised in imaging for radiotherapy. However for lung cancer most protocols are based on breath-hold imaging and patient’s non-compliance to breath-hold manoeuvers can lead to significant artefacts. For patients presenting for lung cancer radiotherapy, maintaining a breath-hold can be difficult. The purpose of this study was to develop a completely freebreathing lung MRI protocol for use in radiotherapy for lung cancer. Method The image protocol was developed using published evidence, and an in-house pilot study. Using sequences available on the

Fig. 1 (a) Planning CT, (b) Planning PET, and (c) T2 weighted HASTE. T2 weighted HASTE image clearly demonstrates the margin between tumour volume and atelectasis scanner, parameters were modified to accommodate lung imaging to allow free-breathing. All imaging was performed on the departmental radiotherapy dedicated 3-Tesla (3T) wide bore MRI (Siemens Magnetom Skyra, Erlangen, Germany) on a flat-bed insert with 18 channel surface coil and 32 channel spine coil. Optimum image quality was determined using in-house image quality score table based on tumour delineation and image artefacts. Images were evaluated by an experienced thoracic radiologist and lung specialist radiation oncologist. Results Sequences with short echo time were selected to maximise signal from lung parenchyma. To manage respiratory and cardiac motion, sequences with the ability to compensate for motion were utilised. For T2-weighted anatomical imaging, a respiratory navigated HASTE sequence showed best image quality for tumour delineation. For T1-weighted imaging, the StarVIBE (radial k-space sampling acquisition) sequence was the best for anatomical and dynamic contrast enhanced imaging. Conclusion This protocol demonstrates the feasibility of a freebreathing protocol for lung cancer radiotherapy using 3T MRI. The image quality was adequate for both tumour delineation and motion evaluation (Fig. 1).

MO16 Not all 3D-printed PLAs are created equal: An initial assessment of physical and radiological variability of 3D-printed bolus samples sourced from different manufacturers in Melbourne Eka Moseshvili, Tom Kupfer, Richard Khor Radiation Oncology Centre, Olivia Newton-John Cancer Wellness and Research Centre, Austin Health, Heidelberg, Australia. ([email protected]), ([email protected] [Presenting author]), ([email protected]) Introduction PLA (polylactic acid) is a promising material for customized bolus printing (1–3), however, it is manufactured using an assortment of chemical formulations (4). To assess the potential impact on dosimetry, we investigated the physical and radiological properties of 3D-printed PLA sourced from different local manufacturers. Method 5 9 5 9 1 cm3 and 100% infill 3D-printed samples were acquired from two commercial services (X and Y). X printed 2 samples using PLA brand A, while Y printed 3 and 2 samples from brands B and C, respectively. We measured weight and size as well as average Hounsfield units (HU) (GE LightspeedCT, 1.25 mm slice thickness, 120 kVp). The coefficient of equivalent thickness (CET) in 6 MV and 10 MV photon fields was derived by comparing against CIRS Plastic Water the TPR measured with an ionization chamber. CET for 6 MeV electron fields was evaluated using radiochromic film by determining the shift in depth dose profile behind the sample compared to a 1 cm thick slab of Plastic Water (5). Results Physical and radiological properties varied with service and brand (Table 1). Warping was visually evident for service Y. The

123

Australas Phys Eng Sci Med Table 1 Physical and radiological properties of PLA samples Service, brand, sample

Weight (g)

Thickness (mm)

Density (g cm-3)

HU (1 SD)

CET (no unit) 6MV U = 0.02

10 MV U = 0.02

6 MeV U = 0.05 1.16

X, A, 1

29.6

10.05

1.19

140.6 (11.5)

1.11

1.14

X, A, 2

30.0

10.01

1.22

137.8 (14.3)

1.12

1.13

1.19

Y, B, 1

27.0

9.95

1.09

60.9 (25.3)

0.96

1.02

1.06

Y, B, 2

28.1

9.95

1.13

118.4 (41.3)

1.07

1.05

1.10

Y, B, 3

29.7

9.90

1.19

221.0 (12.5)

1.12

1.17

1.16

Y, C, 1

26.5

10.0

1.06

-5.1 (58.9)

1.04

1.09

1.02

Y, C, 2

26.7

10.0

1.08

-4.0 (52.0)

1.05

1.09

1.02

U: Uncertainty (k = 2)

maximum weight difference between samples was 13% with intrabrand variability of brand B being 9%. Physical density ranged from 1.06 to 1.22 g cm-3 and this was reflected in HU. However, some samples with similar HU did not yield similar CET. 6 MeV CET was generally higher than MV CET, except for the least dense brand. Conclusion 3D-printed PLAs show considerable variability in density, HU and CET. The observed 15% variability of CET values may be clinically significant, depending on application and bolus thickness. The absence of a clear relationship between HU and CET means that HU cannot be used to predict CET. This may be due to variations in chemical formulations and requires further investigation. References 1. Kim SW, Shin HJ, Kay CS, Son SH. A customized bolus produced using a 3-dimensional printer for radiotherapy. PLoS ONE. 2014;9(10). 2. Zou W, Fisher T, Zhang M, Kim L, Chen T, Narra V, et al. Potential of 3D printing technologies for fabrication of electron bolus and proton compensators. Journal of applied clinical medical physics. 2015;16(3):4959. 3. Burleson S, Baker J, Hsia AT, Xu Z. Use of 3D printers to create a patient-specific 3D bolus for external beam therapy. Journal of applied clinical medical physics. 2015;16(3):5247. 4. Sin L, Rahmat A, Rahman W. Polylactic Acid: PLA Biopolymer Technology and Applications. Elsevier. Amsterdam. 2012. pp 23–35. 5. Su S, Moran K, Robar JL. Design and production of 3D printed bolus for electron radiation therapy. Journal of applied clinical medical physics. 2014;15(4):194.

([email protected]). 5Ingham Institute of Applied Medical Research, Liverpool Hospital. ([email protected]) Introduction SBRT is now a common treatment for early stage NSCLC however toxicities have led to the avoidance of treatment within 2 cm of the proximal bronchial tree [1]. This distance can be impacted by inter-observer contouring variation [2]. With the Australian MRI-linac program comes real-time imaging and the possibility to reduce planning margins around tumours and increase eligibility for SBRT. This project aimed to demonstrate that margin-less treatment plans can meet standard SBRT protocol requirements [3] and assess the accuracy of auto-contouring for treatment planning purposes. Method Two plans were created for five T1/T2, N0 NSCLC patients with GTVexhale \5 cm that were previously ineligible for SBRT due to proximity to various OARs. Plan 1 treated the ITV plus a 5 mm PTV margin and plan 2 treated the GTVexhale with no planning margin. Target doses and normal tissue tolerances for the two plans were analysed. Auto-segmentation of the bronchus was performed and compared to existing manual contours. The distance between the ITV and the proximal bronchial tree was measured for both contours. Results All plans met the requirements in the lung SBRT protocol. The no-margin plans were able to meet the target dose for the GTVexhale whilst resulting in lower doses to the OARs than the PTVmargin plans. The distance between the ITV and the bronchus when using the manual contours ranged from 18.9–67.8 mm compared to a range of 17.9–66.9 mm using the auto-contours. Conclusion The no-margin plans may be considered superior as removing margins means a smaller volume is treated. Similar plans may be suitable for future use with the MR-linac and increase eligibility for SBRT. Clinically a CTV margin may be required but the results for the no-margin situation should translate. Although bronchus auto-segmentation resulted in different contours, there could still be value in auto-contouring for removing inter-observer variation. References 1. Timmerman, R., McGarry, R., Yiannoutsos, C., Papiez, L., Tudor, K., DeLuca, J., Ewing, M., Abdulrahman, R., DesRosiers, C., Williams, M. and Fletcher, J. (2006). Excessive Toxicity When Treating Central Tumors in a Phase II Study of Stereotactic Body Radiation Therapy for Medically Inoperable Early-Stage Lung Cancer. Journal of Clinical Oncology, 24(30), pp. 4833–4839. 2. Jameson, M., Holloway, L., Vial, P., Vinod, S. and Metcalfe, P. (2010). A review of methods of analysis in contouring studies for radiation oncology. Journal of Medical Imaging and Radiation Oncology, 54(5), pp. 401–410. 3. RT 4.188.1 Lung SABR VMAT Planning Protocol. Liverpool & Macarthur Cancer Therapy Centres, SWSLHD Cancer Services. Issue Date: 27/5/15 using Zhang et al. (2011) Radiation Oncology, 6:152. http://www.ro-journal.com/content/6/1/152

MO17 Reduction of planning margins for lung stereotactic body radiation therapy (SBRT) R. Newstead1, R. Short2, J. Dowling3, M. Boxer4, M. Yap4, K. Neville4, J. Veneran4, L. Holloway5 1

University of Sydney, Australia. ([email protected] [Presenting author]). 2Macarthur Cancer Therapy Centre, Campbelltown Hospital. ([email protected]). 3 CSIRO. ([email protected]). 4Liverpool and Macarthur Cancer Therapy Centres. ([email protected]), ([email protected]), ([email protected]),

123

MO18 A practical proposal for a collaborative networked national approach to the provision of particle therapy in Australia; comparison to other countries D. I. Thwaites1, J. Sykes2, V. Ahern2 1

Institute of Medical Physics, School of Physics, The University of Sydney, Australia. ([email protected] [Presenting author]). 2Sydney West Radiation Oncology Network, Westmead,

Australas Phys Eng Sci Med Blacktown and Nepean Hospitals, Australia. ([email protected]), ([email protected]) Introduction Interest in particle therapy is growing internationally, with [100 centres in operation or being commissioned or constructed. Most are protons-only, but *10–15% have carbon ion beams. Australia has no such facility yet, although there are various announced proposals at different stages of development. Some of the evidence can be compared to inform the optimum approach for Australia. Method Various national approaches to modelling the patients likely to benefit from particle therapy (mainly proton and carbon ion beam therapy) are compared for the Australian situation, examining assumptions and levels of evidence bases involved. There are some unique features of the Australian healthcare environment, such as the distances between population centres, which should be accounted for. Different national approaches to the provision of particle beam services are also compared, ranging from a nationally co-ordinated selection process, based on clear clinical, research and infrastructure criteria, to a freely competitive approach. The current Australian proposals are compared to these. Results Depending on the level of evidence base selected, the number of Australian patients who might benefit from particle therapy ranges from 600 per year (on the tightest evidence base criteria) to 10–15 times that. Considering various national frameworks for particle service provision and potential associated research and innovation spin-offs, the optimum for Australia might be a growing collaborative, research-led and networked approach, with a national research centre (ideally with protons and C-ions) and a number of disseminated proton-only systems. This could provide a consistent approach to quality, research, education, patient and plan selection, technology evaluation, clinical and technical innovation, evidence base development and dissemination of outcomes. Conclusion A consistent collaborative national approach to particle therapy provision in Australia could consider a single larger proton and C-ion treatment facility, with smaller compact proton-only facilities where clear unmet service needs exist.

measures the tumour size, with the bottom surface representing the microscopic or Pre-clinical Phase, whilst the upper surface represents the Clinical Phase. Conclusion Several trajectories are demonstrated in the figure below: 1. A classic logistic growth route. 2. A meandering route to emphasise the long latency of many tumours. When the trajectory intercepts the cusp edge there is a catastrophic jump to the upper Clinical plane. 3. A classic treatment route with a concentration on external treatment factors, driving the tumour primarily along the K-axis. If the tumour is not driven beyond the bifurcation region then recurrence is likely. 4. A novel treatment route, driving the tumour along the R-axis until it hits the cusp and catastrophically falls to the Pre-clinical plane. This behaviour is also observed in cases of ‘‘spontaneous remission’’ 4.

References

MO19 A Logistic catastrophe model of tumour behaviour P. J. Riley Faculty of Medicine, Deakin University, Geelong, Australia. ([email protected]) Introduction A simple cusp-catastrophe model captures various aspects of tumour growth, such as long latency, response to therapies and the occurrence of spontaneous remission. Method The model is based upon a Logistic Equation (LE)1 modified to incorporate a sigmoid killing function2. In the classic LE there are two controlling parameters, r = the growth rate and k = the carrying capacity. In the Cusp LE (CLE) these parameters become generalised: so K captures the overall response of factors external to the tumour which will either support or suppress the tumour (eg. vascularisation & immune response); whilst R captures the overall response of factors internal to the tumour (eg. metabolic rate & apoptosis)3. Results The state-space of the tumour is represented by the CLE control surface, shown below. The tumour state at any moment is a point on the surface, whilst a trajectory across the surface shows a possible development history of the tumour. The vertical scale N

1. McAneney, H., O’Rourke, S.F.C. Investigation of various growth mechanisms of solid tumour growth within the linear-quadratic model for radiotherapy. Phys. Med. Biol. 52 (2007) p. 1043 2. Strogatz, S. Nonlinear Dynamics and Chaos. Perseus Books Publishing, 1994, p. 69–79 3. Riley, P. Tumour Growth as a Logistic Catastrophe. World Congress on Med. Phys. & Biom. Eng., Sep. 1997, J. Intl. Fed. Med. Biol. Eng., 35, Suppl.I, p. 497: (F70-PS1.02). 4. Dische, S. et al. The Concentration of Desmethylmisonidazole in Human Tumours and in Cerebrospinal Fluid. Br. J. Cancer (1981) 43, 344ff.

MO20 2-Dimensional noise power spectrum assessment of varied parameters on computed tomographic image quality – A phantom study C. Boyd, K. Hickson, D. McRobbie Medical Physics and Radiation Safety, SA Medical Imaging, Adelaide, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected])

123

Australas Phys Eng Sci Med Introduction Given the clinical value of diagnostic images can be significantly affected by noise properties, an accurate calculation method of the NPS and proper interpretation of results allows noise properties of scanners to be quantified. Evaluation of dosimetric and noise properties between different scanners and protocols provides a more complete understanding through which scanning settings can be optimised. Method Using the approach described in ICRU 87 [1], a script for evaluation of the NPS was written. Based off a matrix and corresponding NPS supplied in AAPM TG16 report [2], the function was validated before being used on phantom images. The approach requests an input DICOM image, uses thresholding to select a uniform region of interest within the centre of a water phantom and a Fast Fourier Transform (FFT) to allow computation of the Normalised Noise Power Spectrum (NNPS). Using a Catphan 650, a series of images were acquired using a variety of machine settings such as kVp, mAs, Pitch, Slice thickness and reconstruction kernel. Each image in the series was anticipated to exhibit different noise properties. Results After generation of the noise power spectrum of each image, a correlation was observed between images acquired with settings associated with poor quality and their corresponding NPS. An example of the effect of mAs on the NPS can be seen in Fig. 1 below.

equipment for the tube voltage range of 40 to 150 kV, establishing the RQR 2 to 10 beam qualities [1–2], cross-comparison of the medium energy free-air chamber (MEFAC) with existing low-energy primary standard, calibrating the monitor chamber of the X-ray equipment with respect to the MEFAC, perform uncertainty budget analysis [3] to establish the uncertainty of this monitor chamber calibration, establish the temporal stability of the monitor chamber and finally validate the monitor chamber against a calibrated dosimeter using the RQR beam qualities using the substitution method [1]. Results Item

Value

Requirement

Monitor chamber leakage current

B±10fA

\2% of the maximum indication in the most sensitive range

Shutter

25 mm tungsten

2 mm lead

Shutter transit time 85.70 ms

N/A

Tube port material 3 mm \2.5 mmAl quality equivalent beryllium filtration Target material

Tungsten

Tungsten

Anode angle

20

B27

Cooling

Oil-to-water Liquid cooled cooler

Percentage Ripple 0.09% over \10% 5 m cable Conclusion From the result shown, the Hopewell X80-320-A X-Ray Irradiator System and the monitor chamber acquired by ARPANSA are adequate for the purpose of dosimetry tasks in the diagnostic beam energy range. Further work can then be carried out as mentioned in the Method section. The availability of such service can enhance patient dosimetry in diagnostic imaging clinics across Australia. References

Conclusion The 2D NPS represents noise properties of digital computed tomography scans with sufficient reliability to warrant further work on possible application as a protocol optimisation tool.

1. International Atomic Energy Agency (2007) Dosimetry in Diagnostic Radiology: An International Code of Practice. IAEA Technical Report Series 457, Vienna 2. International Electrotechnical Commission (2005) Medical Diagnostic X-ray Equipment – Radiation Conditions for Use in the Determination of Characteristics. IEC 61267. IEC, Geneva 3. Joint Committee for Guides in Metrology (2010) Evaluation of Measurement Data – Guide to the Expression of Uncertainty in Measurement. JCGM 100:2008.

MO21 Calibration at diagnostic beam energies Kam L. Lee ARPANSA. ([email protected]) Introduction Radiation dosimetry has been well established in radiotherapy and ARPANSA has also established a clinical dosimetry service called the Australian Clinical Dosimetry Services to calibrate dosimeters and audit clinical dosimetry. However, such services are lacking in diagnostic imaging area. ARPANSA has recently acquired a dosimetry grade X-ray system (the Hopewell X80-320-A X-Ray Irradiator System) and is preparing to provide calibration service in the diagnostic beam energy space. Method The method adopted will essentially follow the IAEA TR457 [1] document. The stages involve measuring the shutter transit time of the X-ray equipment, establishing the inherent filtration of the X-ray

123

MO22 Dosimetric definition of a normal coronary angiogram from a retrospective review of 13 months of clinical data L. Wilkinson, D. Kim Department of Medical Engineering and Physics, St. Vincent’s Hospital, Melbourne. ([email protected] [Presenting author]), ([email protected]) Introduction A coronary angiogram is a complex procedure that incorporates high dose acquisition runs and lower dose fluoroscopic imaging both of which may be affected by a range of factors. Understanding what is a normal coronary angiogram is not intuitive

Australas Phys Eng Sci Med and there is a need for a description of what might be expected in terms of the common radiation dosimetric terms. Method A dataset of 498 coronary angiograms conducted on two different imaging systems (Philips and Siemens) over a 13 month period has been collated. The dosimetric parameters of Kerma Air Product (KAP), Entrance Air Kerma (EAK) and fluoroscopy time were correlated with the extent of coronary disease as per the cardiologist’s report. From the dosimetric parameters two new terms, the effective beam area at the reference point (KAP/EAK) and the effective examination dose rate (EAK/sec), were derived. For a subset of this dataset the number of acquisition runs conducted (n = 230) and where available the patients weight (n = 179) was recorded. This dataset was then reviewed to establish what the expected ranges of the dosimetric terms might be for patients with differing levels of coronary disease and weight range. Results A review of this dataset shows that in general terms 67% of patients are male with an interquartile weight range of 71–96 kg. Of these patients 25% are considered to have no coronary disease and for these patients the interquartile ranges of 20.3–40.8 Gy cm2 (KAP), 308–616 mGy (EAK), 133–362 s fluoroscopy time and 8–9 acquisition runs could be expected. In this context it would be expected that the fluoroscopy component of the exam contributes 16–19% of the total examination KAP. Conclusion These ranges provide a useful baseline for cardiologist education and for DRL and optimisation assessments.

Topics

Radiation Oncology Related

Imaging Patient outcomes

Types of research Radiation Safety

MO23 Radiation oncology medical physicists: Perceptions of research M. A. Ebert1,2, G. K. B. Halkett3, M. Berg3, D. Cutt4, M. Davis4, M. House2, R. Kearvell1, S. Maresse5, J. McKay5

IT Related Workforce Department efficiency

Sub-Topics Dosimetry New technology Planning Brachytherapy Stereotacc radiotherapy (including SABR and SBRT) Monte Carlo Treatment techniques VMAT/IMRT Adapve radiotherapy Synchrotron Total Body Irradiaon Electrons Tomotherapy Parcle therapy Advanced imaging techniques PET/CT/MRI/Nuclear Medicine Clinical research - outcomes Clinical trials Radiobiology Experimental research QA Radiaon protecon Developing a Database Soware engineering Workforce Issues Educaon and training Workflow

Fig. 1 Areas of research in radiation oncology of interest to ROMPs

1

Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia. ([email protected]). 2School of Physics, University of Western Australia, Western Australia. ([email protected] [Presenting author]), ([email protected]). 3School of Nursing, Midwifery and Paramedicine, Faculty of Health Sciences, Curtin University, Perth, Western Australia. ([email protected]), ([email protected]). 4Genesis Cancer Care Western Australia. ([email protected]), ([email protected]). 5Discipline of Medical Radiation Science, Faculty of Science and Engineering, Curtin University, Perth, Western Australia. ([email protected]), ([email protected]) Introduction As part of a program to understand and improve workforce support in Radiation Oncology, Radiation Oncology Medical Physicists (ROMPs) who have been employed in Australia were surveyed regarding their attitudes to and perceptions of undertaking research. Method A content validity-tested survey was made available online and promoted via professional societies and contacts. The instrument was used to collate responses regarding demographics, educational, work experience levels, country/state of origin, employment history, satisfaction and intentions, research participation, opportunities and perceptions. Descriptive statistics were used to summarise responses for each question. The survey included open ended questions and responses were evaluated qualitatively to identify key themes. Results Eighty-eight ROMPs completed the survey with 65 indicating previous involvement in research activities. Of those 65, the majority (89%) had been involved in department-level projects with 23%

involved in international studies. One quarter (25%) had received funding as a chief/principal investigator. Two percent of ROMPs indicate dislike of research participation, and 55% indicated their involvement in research encouraged them to remain in the profession. Figure 1 lists research areas that ROMPs indicated they were interested in. Multiple benefits of research were identified, both personal and patient-care related. The personal barriers of conducting research were predominantly related to a lack of time. Conclusion This survey has provided a snapshot of the interests and attitudes of ROMPs to conducting research. The perceived benefits of research are many and diverse, though similarly there are many perceived barriers which should be addressed in order to facilitate what is likely to be an important aspect of the ROMP workforce. Acknowledgements We acknowledge funding support from the Commonwealth Department of Health Better Access to Radiation Oncology program, and the Department of Health Western Australia, Cancer

MO24 Challenges, triumphs and expectations of TEAP in NSW: From 2004 and beyond! L. Wilfert1,2, S. Howlett3 1

Department of Radiation Oncology, Calvary Mater Newcastle, NSW Australia. 2School of Physics & Mathematical Science, University of Newcastle, NSW Australia. ([email protected] [Presenting author]). 3ACPSEM, Mascot, NSW Australia.

123

Australas Phys Eng Sci Med Introduction Based on the recommendations of a national enquiry into Radiotherapy [1], the NSW and Federal Governments injected significant funding into the NSW Public Radiation Oncology Medical Physics (ROMP) profession. This funding included dedicated training positions, with the goals of having a permanent workforce of certified ROMPs, to ensure patient safety is maintained, and achieving a reduction in vacancy rates. The need to minimise the impact on clinical services during workforce shortages was a high priority and therefore supporting infrastructure was provided. As a result of this infrastructure, NSW has played a lead role in the implementation and development of ACPSEM’s structured training program, TEAP, over the past 12 years. The objective of the periodic review of Registrar data is to identify the outcomes of the financial support, to improve retention, and review our capacity to meet RANZCR’s workforce projections [2] with varying radiotherapy utilisation rates. Method and Results NSW public Registrar data has been collected, including demographic data, funding source, regional spread, postgraduate degree status on entry, retention, exam pass rates, time to certification and post-certification employment. An overview will be presented, including a comparison with NSW ROMP workforce projections [2]. Conclusion NSW public TEAP participants have met many challenges, including changes to requirements, access to brachytherapy training and significant workload of training, assessment and documentation. However, these challenges are far outweighed by the significant benefits, including having a steady supply of qualified ROMPs to fill vacancies, a more versatile qualified ROMP with increased scope of training, state-wide coordination by the NSW Clinical Program Coordinator position, annual feedback from ACPSEM, and a reduction in the reliance on slow overseas recruitment. The NSW public sector has significant training capacity to cope with additional training positions if needed to increase the utilisation rate of Radiotherapy across Australia. References 1. Baume, P (2002) A Vision for Radiotherapy, Commonwealth of Australia, Canberra. 2. Philip Munro, private communication (2014) based on Faculty of Radiation Oncology, (2014) ‘‘Projecting the Radiation Oncology Workforce: 2013 Update’’. RANZCR, Sydney.

Fig. 1 Example images with major tracts labelled (CS = corticospinal tracts, ML = medial lemnisci, PF = transverse pontine fibres, CP = decussation of the superior cerebellar peduncles, OT = optic tracts, OR = optic radiations, AC = anterior commissure, LF = superior longitudinal fasciculi, CC = corpus callosum, UF = subcortical U-fibres) directions. Scan data were transferred in DICOM format to an independent computer and analysed using MRtrix software with a standardised CSD analysis: white matter masking, averaging, tensor calculation, determining of Fractional Anisotropy (FA), thresholding to generate single fibre pixel populations, calculation of the single fibre response function and spherical deconvolution. A whole brain probabilistic method was used to generate the streamline tractogram images. A QC programme was devised to investigate data integrity. Results Clinically useful tractograms were obtained in 10/12 patients (Fig. 1). Excessive movement and too few diffusion directions were the reason for the two undiagnostic scans. Higher b-value and directionality were the principal determinants of image quality, although QC provided additional assurance of the results. Conclusion MR tractography can be implemented on standard clinical scanners, and is capable of producing useful clinical data which may aid diagnosis or patient management. A relevant QC programme is an important component of the process. The service can be set up with modest resources. References 1.

O015 Setting up a clinical MR tractography service: Initial results and observations

Tournier JD, Yeh CH, Calamante F et al. (2008) Resolving crossing fibres using constrained spherical deconvolution: Validation using diffusion-weighted imaging phantom data. NeuroImage 42:617–625.

D. W. McRobbie, M. J. Agzarian SA Medical Imaging, Flinders Medical Centre, Bedford Park, South Australia. ([email protected] [Presenting author]), ([email protected]) Introduction Research-based developments in multi-directional diffusion MRI have not yet found widespread clinical application. Reasons include the degree of uncertainty associated with streamlining algorithms and the ability of diffusion tensor imaging (DTI) to resolve crossing fibres. Constrained Spherical Deconvolution (CSD) offers a means of overcoming both these issues, making it a suitable clinical tool. The presentation will focus on the practical steps to establish such a service, methodologies, outcomes and pitfalls. Method Scanning was performed on Siemens Trio (3T) and Aera (1.5T) scanners using b = 1,000 or 3,000 smm-2 with 30 or 64

123

O016 Fast and robust 3D T1 mapping sequence for quantification of Iron Oxide Nanoparticles concentrations in vivo Muna Adhikari-Paudel, Neha Konjoo, Charles Castets, Sylvain Miraux, Emeline J. Ribot Centre de Re´sonance Magne´tique des Syste`mes Biologiques, UMR 5536 CNRS/Universite´ de Bordeaux, France. ([email protected] [Presenting author]), ([email protected]) Background Iron Oxide Nanoparticles (IONPs) are contrast agents mostly used for angiography and cell tracking. It appears crucial to

Australas Phys Eng Sci Med Table 1 Measurements of the r1 relaxivity of IONPs with 3D UTE LL T1 and the 2D IR sequences r1 (mM-1 s-1) r1 (mM-1 s-1) % Errors 3D UTE LL T1 sequence 2D IR

IONPs

Sinerem

2.84 ± 0.1

2.8 ± 0.1

1.4

P904

1.44 ± 0.18

1.27 ± 0.03

12.9

Ferumoxytol 2.53 ± 0.01

2.46 ± 0.33

References 1. Castets CR., Ribot EJ, Lefrancois W, Trotier AJ, Thiaudiere, E, Franconi, J-M, Miraux S (2015) Fast and robust 3D T1 mapping using spiral encoding and steady RF excitation at 7T: application to cardiac manganese enhanced MRI (MEMRI) in mice. NMR in Biomedicine 28: 881–889.

2.84

O017 Reducing administered activity in FDG oncology PET CT – How low can we go? 100 80

D. Tout1, L. Du2, J. Richards2

Ferumoxytol

C ( g/ml)

1

60

Biomedical Technology Services, Gold Coast University Hospital, QLD, Australia. ([email protected] [Presenting author]). 2Gold Coast Hospital and Health Service, QLD, Australia. ([email protected]), ([email protected])

P904

40

Sinerem

20 0 0

200

400

600

Ti me (minutes)

Fig. 1 Measured Ferumoxytol, P904, Sinerem concentrations in blood as a function of time in mice

quantify IONPs to evaluate the amount of therapeutic IONP-labeled cells reaching a tumor. Transversal relaxivity measurements remain to date inadequate in vivo. Therefore, the quantification of their effects on the longitudinal relaxation time (T1) was developed. Methods A 3D T1 mapping sequence combining a Look-Locker (LL) module and Ultra-short Echo Time (UTE) acquisitions was developed by using stacks of spirals [1]. The slice selection gradient was removed in order to achieve a UTE of 0.5 ms. Two spiral interleaves were acquired at each Saturation Time in order to reduce acquisition time.The longitudinal relaxivity (r1) of tubes containing increasing concentrations of Ferumoxytol, Sinerem or P904 (0, 0.5, 0.75, 1, 1.5, 2 mM) were measured at 7T using different LL parameters (number of points: 10, 18, 23, 50; delay time: 30, 80, 140, 200 ms; flip angle: 5, 10, 15), and compared with a 2D Inversion Recovery (IR) sequence. C57/Bl6 mice were intravenously injected with 100 lmol/ kg of these IONPs. T1 maps were acquired before, right after, 1, 2, 4, 6 and 8 h post-injection. Results Optimal parameters were 18points separated by 140 ms with a flip angle of 10, in 15 min. The measured r1 of Ferumoxytol, P904 and Sinerem were 2.53 ± 0.01, 1.44 ± 0.18 and 2.84 ± 0.1 mM-1s-1, respectively, with a [85% accuracy (Table 1). The half-lives were measured in the jugular veins (3.49 h ± 0.07, 2.75 h ± 0.57 and 2.22 h ± 0.5 for Ferumoxytol, Sinerem and P904 respectively (Fig. 1), and are in accordance with literature. Conclusions The developed 3D UTE-LL-spiral T1-mapping protocol allowed to rapidly and accurately quantifying T1 of a wide range of IONPs concentrations, enabling the clearance measurement of IONPs non-invasively. Keywords: Magnetic Resonance Imaging, T1-mapping, iron-oxyde nanoparticle, quantification, mouse.

Introduction With advances in PET technology, such as extended axial field-of-view and time-of-flight (TOF) imaging giving increased sensitivity and signal-to-noise ratio (SNR), are recommended administered activities for 18F-FDG oncology still providing the optimal balance between image quality and radiation safety? The PET service at Gold Coast University Hospital uses an administered activity of 4 MBq/kg 18F-FDG in line with national diagnostic reference activities. List-mode PET data was used to investigate the relationship between administered activity and image quality. Method 30 oncology patients (14 male, weight 47-102 kg) administered with 4 MBq/kg 18F-FDG were imaged on a Siemens Biograph mCT at 2.5 min per bed position from vertex to mid-thigh. PET data were acquired as list-mode. Reconstruction was performed using 3D iterative reconstruction with point-spread-function (PSF) modelling and TOF for the full acquisition time, as well as 3 reduced imaging times simulating administered activities of 3.5, 3.0 and 2.5 MBq/kg. Results in each case were compared to the full reconstruction (4.0 MBq/kg). The relationship between administered activity and noise equivalent counts is complex, and the required reduction in acquisition time to simulate reduced activity was estimated based on data acquired from a decaying phantom with scatter and randoms added. Results Quantitative analysis of lesion SUVmax, lesion volume (40% of SUVmax) and liver SNR are shown in Table 1. Conclusion Changes in lesion SUVmax and volume with reduced administered activity were small and not significant. The reduction in liver SNR was significant (p \ 0.001) for all reduced activities. Further qualitative analysis by experienced observers will demonstrate whether this reduction in SNR reduces diagnostic image quality Table 1 Change in lesion SUVmax, lesion volume and liver SNR with reduced activity Parameter

n

Mean (SD) change compared to 4.0 MBq/kg 3.5 MBq/kg

3.0 MBq/kg

2.5 MBq/kg

Lesion SUVmax

67

+0.4% (2.0%)

+1.1% (5.6%)

+1.0% (4.8%)

Lesion Volume

67

-0.8% (4.7%)

-0.7% (6.5%)

-1.3% (10.0%)

Liver SNR

30

-3.4% (2.0%)

-7.8% (2.3%)

-12.1% (2.8%)

123

Australas Phys Eng Sci Med and help to determine optimal administered activity to provide patients with the lowest radiation exposure level whilst maintaining diagnostic accuracy.

O018 Time-of-flight performance assessment on clinical PET scanners D. Tout Biomedical Technology Services, Gold Coast University Hospital, QLD, Australia. ([email protected] [Presenting author]) Introduction Advances in PET technology and crystal detectors enabled the first commercially available time-of-flight (TOF) PET scanner in 2006. Today all major PET manufacturers offer systems capable of TOF imaging. Despite it’s documented impact on clinical imaging, current NEMA standards [1] do not specify tests for TOF performance assessment. Annihilation photons occurring from a point source at the isocentre of a PET scanner will have identical path lengths (and therefore TOF) to all detectors in the same 2D plane. Annihilation photons occurring from point sources offset from the centre will have path lengths (and TOF) which vary with angle of emission. The difference in the path lengths of 2 annihilation photons has been shown previously to be expressed as a simple sinusoidal function [2], and can be converted to TOF difference by c, the speed of light. Method Offset point source acquisitions will be performed on 3 Siemens Biograph mCT scanners, which have 13 time-bins (width 312.5 ps). The maximum intensity plane from each time bin sinogram can be used to determine the total counts for each acquisition angle. Results The TOF bin width can be converted to a distance measurement, and applied to the theoretical model for the difference in path lengths to generate angular ranges for correctly assigned counts in each TOF bin sinogram for each of the PET scanners. Conclusion Results from this investigation can be used as a performance measure for timing resolution of TOF PET scanners. A comparison between systems can be made, as well as establishing baseline values for annual performance assessment. References 1. NEMA NU 2-2012 Performance Measurements of Positron Emission Tomographs. NEMA 2013, Rosslyn, VA 22209. 2. Armstrong IS, Tout D, Williams HA. PET time-of-flight performance using analytic modelling and offset point-source measurements. IEEE Nucl Sci Symposium 2010;3389–3392.

(b) the development of corrections for the non-ideal behaviour of the beam line optics and data acquisition system. These outputs (a) allow for comparison against diagnostic reference levels for medical CT [2] and enable balancing the trade-off between reconstructed image quality and radiation dose delivered during the scan [1], and (b) deliver CT reconstructions that are free from systematic errors manifesting as artefacts. Accurate CT data can be utilised for dual energy methods of materials analysis including [3] K-edge subtraction (KES) and dual energy x-ray energy analysis (DEXA) [4, 5]. Method Experimental studies utilised the Australian Synchrotron Imaging and Medical beam line. The incident primary beam was monitored with an ion chamber, and CsI flat panel array. Beam quality was investigated for transmission through copper. Scatter profiles were measured using the beam stop method. Detector after glow was investigated at maximum frame rate using a fast shutter to open/close the beam. Dosimetry utilised perspex phantoms with diameters 35 mm (mouse) 50 mm (rat), 100 mm (child), 160 (adult head) and 320 mm (adult abdomen), plus an ionisation chamber calibrated to measure the CT dose index (CTDI). Results Measurements were repeated at photon energies 30-100 keV, with results summarised graphically. Conclusion The beam quality is near mono-energetic modified by nonlinearities in the detection system. The scatter to primary ratio (SPR) for CT with the CTDI phantoms is quantified as a function of beam energy and sample size. Dosimetry measurements deliver conversion coefficients that scale the ionisation chamber measurements during each CT scan to an estimate for CTDI for a range of sample sizes. Acknowledgements This research was undertaken on the Imaging and Medical Beam line at the Australian Synchrotron, Victoria, Australia. Travel funding was provided by the New Zealand Synchrotron Group. References 1. ARPANSA (2008) 14.1 Radiation protection in diagnostic and interventional radiology. 2. Australian National Adult DRL for MDCT. http://www.arpansa. gov.au/services/ndrl/adult.cfm. 3. Y Zhu et al. (2014). Spectral K-edge subtraction imaging Phys. Med. Biol. 59, 2485–2503. 4. Midgley SM (2013) Feasibility study for DEXA using synchrotron CT at 20-35 keV Phys Med Biol 581085–1205 5. Midgley S & Schleich N (2015) DEXA using synchrotron computed tomography at 35 and 60 keV J Synch Rad22 807–818

IS06 The role of MR physics in the public health service D. W. McRobbie

O019 Dosimetry and pre-processing corrections for synchrotron computed tomography

SA Medical Imaging, Flinders Medical Centre & Royal Adelaide Hospital, Adelaide, South Australia ([email protected] [Presenting author])

S. M. Midgley1, N. Schleich2

Introduction Whilst medical physics in Australasia has been mainly concerned with the radiation sciences (radiology, nuclear medicine, radiation oncology), there is a growing need for medical physics expertise in Magnetic Resonance (MR). Such expertise exists within the university sector, but little has translated out into clinical service. Method The presentation will consider the role of MR physics in a number of areas:

1

School of Physics and Astronomy, Monash University. ([email protected] (Presenting author]). 2Department of Radiation Therapy, University of Otago, Wellington. ([email protected]) Introduction Synchrotron computed tomography studies with large samples requires (a) an estimate of the radiation dose delivered, and

123

• •

MR safety expertise Teaching/training

Australas Phys Eng Sci Med • • • • • •

Procurement Quantitative imaging Quality control Support for advanced applications Sequence design and optimisation Supporting clinical trials

Results Examples will be shown in advanced neuro imaging (fMRI, DTI, tractography), radiation oncology (MR treatment planning, quantitative diffusion, PET-MRI) together with an overview of quantitative QC methods and the role of the MR Safety Expert as advised by international bodies1,2. Conclusion There is a significant role for greater involvement of DIMPs and ROMPs in clinical MRI services. This presents and opportunity and a challenge for the ACPSEM to develop medical physics training and credible expertise in this area. References 1. Institute of Physics and Engineering in Medicine (2013). Policy statement: Scientific Safety Advice to Magnetic Resonance Imaging Units that Undertake Human Imaging. http://www.ipem. ac.uk/Portals/0/Documents/Publications/Policy%20Statements/ IPEM_MRSafetyExpert_PolicyStatement_04102013_SK.pdf 2. Calemante et al (2016). The Inter-Society Working Group on MR Safety: Recommended responsibilities for management of MR safety. J Magn Reson Imag, On-lin early publication 3 June 2016. http://onlinelibrary.wiley.com/wol1/doi/10.1002/jmri.25282/ abstract.

KS05 IMRT dose delivery in practice: What are we actually delivering? S. Kry The University of Texas MD Anderson Cancer Center ([email protected] [Presenting author]) IMRT has become the standard of care for the treatment of many disease sites because dose distributions generated with this technique are far superior to those from conventional radiotherapy. However, in order to realise the potential benefits of IMRT it is essential that treatments be delivered correctly. It is paramount that the intended dose distribution is that which is actually delivered. This talk will highlight the poor rate with which IMRT plans are accurately delivered. This will be based primarily on the IROC Houston IMRT anthropomorphic phantom program, where institutions treat the phantom as a patient. Disconcertingly, despite relatively loose criteria, there remains a relatively poor rate at which institutions deliver the dose they believe they are delivering. The underlying causes of this problem will be shown to most often involve the treatment planning system dose calculation, which has broad implications for all treated patients. The poor results of the IROC Houston phantom program will be contrasted with ubiquitous measurement-based patient specific IMRT quality assurance, which typically shows very good results. The sensitivity and specificity of common measurement-based patient specific quality assurance systems will be reviewed to explore the apparently inconsistent results between these two systems. These findings will highlight the need for further study on how the field should be verifying IMRT treatment integrity.

O020 The sensitivity and specificity of patient-specific QC at Wellington Blood and Cancer Centre B. E. W Scarlet, A. J. Williams, R. J. W. Louwe Wellington Blood and Cancer Centre, Wellington, NZ. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Three different patient-specific QC methods are available in our department: time resolved point dose measurements (TRPD)[1], EBT3 gafchromic film verification, and usage of an ArcCheck device[2]. This study aims at quantifying the sensitivity and specificity of these QC methods for clinically relevant errors during VMAT. Method Intentional errors simulating incorrect linac output, MLC position, dosimetric leaf gap (DLG), focal spot size, (FSS) and output variation with gantry angle were introduced to five H&N plans in Eclipse[3]. Measurements were carried out using a TrueBeam[4]. Sensitivity metrics were SA: change in QC result over change in conformity index (CI); and SB: change in QC result over introduced error magnitude. In this study, CI95 is the ratio of volume receiving 95% of prescribed dose and the PTV volume receiving that dose. TRPD specificity was assessed by comparing the average deviation per segment in a specific range of the detector distance to field edge (DTFE) for different intentional errors. Results The median CI95 over all patients and all introduced errors was 1.11 (1.00–1.71). Integral point dose deviations showed a linear correlation (r2 = 0.80) with SA. Film and ArcCheck results showed poor correlation between gamma pass rate changes and SA (r2 = 0.30 and 0.48 respectively). Similar results were obtained for SB, although r2 depended more on the type of error. In addition, the ArcCheck measurements resulted in increased numbers of false negatives. TRPD analysis showed that different error modes could be resolved; MLC errors resulted in increased deviations in region II, while output errors increased deviations in region III (Fig. 1; Table 1). Conclusion ArcCheck and film showed poor correlation between QC results and CI95. Furthermore, ArcCheck results had a high false negative rate. TRPD analysis showed good correlation and can resolve different error modes, but may require averaging over multiple measurements.

Fig. 1 Average contribution to the fraction dose deviation for the orginal plan and three intentional error plans. The solid line represents the average over five patients. Region I = DTFE \ -0.5 cm; Region II = -0.5 cm, Region III = DTFE [ 0.5 cm

123

Australas Phys Eng Sci Med Table 1 Sensitivity of QC methods for various error modes.Gamma analysis was performed using a 2%, 2 mm gamma criterion Error Type

Sensitivity Metric

PD (% change in Film (change in c ArcCheck (change in integral dose) pass rate) c pass rate) Medin Range

MU

MLC

Medin

Range

%/DCI95

32.9

17.945.0

491.1

120.0–639.6

376.4 203.6–603.9

%/DMU

1.0

1.0–1.0

-15.6

-20.0 to -6.9

-12.0 -15.2 to -10.9

%/DCI95

39.8

11.088.4

33.4

-846.3 to 548.2

329.8 147.0–693.4

3.8

1.7–4.7

-37.3

-72.1 to 4.0

-24.3 -71.8 to 50.2

-32.7 to -11.6

52.7

-194.1 to 159.6

-213.2 -322.8 to -68.3

2.8–5.5

-6.4

-27.4 to 32.6

-72.8 to 25.6

11.3

-151.7 to 308.7

0.0

-1.4 to 0.6

%/mm DLG

Medin Range

%/DCI95 %/mm

-24.7 4.0

FSS

%/DCI95

0.1

0.1 to 0.3

MU with gantry angle

%/DCI95

36.8

22.4–60.9

%/DMU

0.3

%/mm

-8.1

0.3–0.4

34.5 54.3–15.5

and compared against EBT3 GafChromic film QA results to evaluate PSQA performance. Once it was maintained that the Octavius 1000 SRS was effective at measuring SBRT plans, a rigorous protocol was developed for using the equipment for PSQA. This included devising a method which utilized the Elekta HexaPod robotic couch and iGuide infrared tracking system to position the array accurately to the machine isocentre. Results The detector array was found to perform as expected in commissioning tests. The Octavius system provided patient specific QA results comparable to results achieved using film. Finally, the procedure for positioning the phantom has been shown to provide sub-millimetre set-up accuracy—which aids the reduction of uncertainty in SBRT QA. Conclusion The Octavius system, along with the PSQA methodology proposed in this presentation, has been demonstrated to be an excellent tool for patient specific measurements.

-48.6 -335.1 to 189.3 0.4 -0.9 to 1.3 242.9 195.5–496.3

O022 Clinical validation of a dynamic anthropomorphic phantom for SBRT and DIBH QA

2.1 1.1–4.1

Chuan-Dong Wen References 1. R.J.W. Louwe et al., Time-resolved dosimetry using a pinpoint ionization chamber as quality assurance for IMRT and VMAT. Med Phys. 2015; 42(4):1625–39. 2. Sun Nuclear Corporation, Melbourne FL, USA 3. Eclipse, Version 11, Varian Medical Systems, Palo Alto CA, USA 4. TrueBeam, Version 2.5, Varian Medical Systems, Palo Alto CA, USA

O021 Using the Octavius 1000 SRS 2D array and 4D rotational phantom for patient specific SBRT measurements. P. O’Connor1, P. H. Charles2 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]). 2 Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected]) Introduction Due to the severe impact any delivery errors can have on Stereotactic Body Radiotherapy (SBRT), physicists devote a lot of time doing patient specific quality assurance (PSQA) measurements to ensure safe treatment. PSQA, The PTW Octavius 1000 SRS is a new instrument which aims to improve measuring SBRT dose distributions. This high resolution array was investigated as a replacement for radiochromic film for PSQA measurements. A proposed PSQA methodology, including a solution to the problem of sub-millimetre positioning accuracy with a 2D array is presented in detail. Method The 1000 SRS array consists of 977 liquid-filled ionization chambers (diameter = 2.3 mm) in an 11 9 11 cm2 area. The detector spacing is 2.5 mm in the inner 5.5 cm and 5 mm outside this. Initially the array system was characterised with a range of standard commissioning tests. A set of past patient SBRT plans was then delivered and analysed (using the popular gamma analysis technique)

123

INWENTECH, Melbourne, VIC 3150 Australia. ([email protected]) Introduction A novel and innovative dynamic anthropomorphic phantom has been designed and prototyped to assist the clinical implementations of SBRT and DIBH techniques for treating lung and breast cancers. It can be used for equipment commissioning and patient specific pre-treatment QA purposes. This paper describes the preliminary clinical validation of this device using 4DCT studies. Method The concept of an innovative breathing phantom has been investigated for clinical applicability with current available technologies. The architecture of a unique electromechanical driven, anthropomorphic device was designed and a functioning prototype was constructed. In order to confirm the availability of intended functions and features, a preliminary clinical validation was carried out on the prototype. A GE 4DCT (Advantage Windows v6.4) system was used to scan the device for verifications of organs available and their image qualities, motion ranges, speeds and frequencies as well as dimensions and relative distance variations. A series CT images and 4DCT videos were produced for fine analysis. Results It has demonstrated that the prototype has the designed features of organ motion and deformity of chest, lung, heart and skin during a breathing cycle. Furthermore, it can demonstrate the complex tumour motion locus in lung and the heart position shift due to the DIBH effect, and therefore it can simulate the processes of human respiratory motions occurring during SBRT and DIBH treatments. Of importance, it can represent the true clinical complexity of above treatment techniques. Therefore, the clinicians can use this device for final end-to-end testing during system commissioning, on individual clinical pre-treatment QA and for staff training purposes. Conclusion A comprehensive and purposely-built dynamic phantom has been constructed for the complex motion management in SBRT and DIBH treatments. This device can physically and radiographically represent and re-produce the human thorax in respiratory process. References 1. Keall et al. (2006) The management of respiratory motion in radiation oncology report of AAPM Task Group 76 Med. Phys. 33 10, 3874–900

Australas Phys Eng Sci Med 2. Seppenwoolde et al. (2002) Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy Int. J. Radiat. Oncol. Biol. Phys. 53 822–34 3. Hayden et al. (2012) Deep inspiration breath hold technique reduces heart dose from radiotherapy for left-sided breast cancer J of Med Imag and RadOnc 56 464–472 4. P Steidl et al. (2012) A breathing thorax phantom with independently programmable 6D tumour motion for dosimetric measurements in radiation therapy Phys. Med. Biol. 57 2235–2250

IS07 Pre-clinical evaluation of the integral quality monitor for real-time beam verification Jurgen Oellig iRT Systems, Koblenz, Germany. ([email protected]) Introduction Quality Assurance of complex Radiation Therapy is critical to ensure patient safety but can be time consuming for the Medical Physicist. The Integral Quality Monitor (IQM) System is designed to be a real-time beam verification system that monitors the accuracy of radiation delivery throughout each patient treatment without any user interaction. IQM detector characteristics such as signal reproducibility, linearity sensitivity and dose rate independence are fundamental for the successful application of the IQM principle of operation. Methods The detector characteristics were thoroughly evaluated during the extensive IQM pre-clinical tests at some of the world’s leading radiation therapy departments. Over twenty clinical centers performed a variety of pre-clinical evaluation measurements to verify signal reproducibility (short-term and long-term), signal linearity, dose rate dependency and sensitivity. The results outlined are based on measurements performed on various types of Linear Accelerators. Results Every IQM installed at each RT department passed the criteria for short and long term reproducibility, independent of field size and field position, with a coefficient of variation under 0.5% and 1% respectively. The results indicate a high reproducibility with low variation. Every installed IQM passed the criteria for linearity. The results confirm that IQM Signal changes linearly with applied dose output. The variation in the IQM Signal measured for a given dose delivered at different dose rates are well within ±0.5%. Conclusion Both short-term and long-term reproducibility measurements show high signal reproducibility of the IQM signal independent from field size and field position. The IQM signal varies linearly with applied dose. The IQM Signal is virtually independent of dose rate. Discussion Multiple clinical test partners are now evaluating the process of integrating the IQM system into their clinical workflow and the risks associated with this integration. Some test partners have evaluated the capability of the IQM System to verify new and highly complex treatment delivery techniques like hypo-fractionated SBRT treatments and for the purpose of daily Linear Accelerator Quality Assurance tasks.

O023 Dose to medium or dose to water: Commentary from ARPANSA J. Lye1, M. Shaw1, F. Gibbons1, S. Keehan1, A. Alves1, I. Williams1, D. Butler2

1

Australian Clincial Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2 Australian Radiation Protection and Nuclear Safety Agency, ARPANSA ([email protected]) Introduction Since the introduction of Monte Carlo based treatment planning systems, calculation based on dose to water with differing densities (Dw) or dose to true medium (Dm) has been debated1. One key argument for using the Dw option in newer algorithms such as Monaco MC or Eclipse Accuros or iPlan MC is to match to historical prescriptions. However, the Dw option in Monaco, Accuros, and iPlan does not calculate Dw in the same manner as older algorithms such as Xio Superposition/Convolution and Eclipse AAA. The concern with Dw calculations in Monaco, Accuros, and iPlan is that they do not match historical calculations of Dw. Method An investigation was conducted at ARPANSA using Monaco and Xio TPS. Plans were completed in both systems using a solid water phantom with 30 cm geometry. Dense bone material (ED = 2.0) of various shapes and sizes was contoured in the centre of the phantom, and the same plan was replicated in all cases using Monaco Dm, Monaco Dw and Xio. Results The investigations showed that for a simple slab of dense bone up to 10 mm, Monaco Dm matches historical calculations (Xio Dw) to within 5%. Monaco Dw calculations with the slab geometry show a *10% difference compared to Xio Dw calculations. Similar results were seen using a simple cuboid of dense bone. Conclusion When clinical plans call for prescription dose in dense bone, MC based TPS should use Dm calculations in preference to Dw calculations to more closely resemble historical planning. For other tissue types the difference between Dm and Dw options is minimal. References 1. Andreo P. Phys. Med. Biol. 2015; 60: 309–337

O024 Clinical implementation of a monte carlo-based electron dose calculation algorithm for radiotherapy treatment planning – Dosimetric and therapeutic considerations Andrew Kovendy1, Aisling Haughey2 1

North Coast Cancer Institute (NCCI), Coffs Harbour, New South Wales. ([email protected] [Presenting author]). 2Radiotherapy Unit, Altnagelvin Hospital, Londonderry, UK. ([email protected])

Introduction The introduction of Monte Carlo (MC) computational methods into commercial radiation therapy planning systems has a potentially significant impact to the accuracy of dose calculations for both photon and electron treatments and an improvement to treatment outcomes1,2,3. However the implementation of electron MC dose calculations in a clinical environment requires careful evaluation of the algorithms and options used and the implications of statistical noise and dose to medium calculations for treatment prescription3. Method The electron Monte Carlo (eMC) algorithm recently incorporated into the Elekta Monaco4 radiotherapy treatment planning system (RTPS) was commissioned for routine clinical use at the

123

Australas Phys Eng Sci Med NCCI. Beam models provided by the vendor were validated against profiles scanned in water and point doses measured in both water and tissue-equivalent material. Radiochromic film embedded in anthropomorphic phantoms containing lung and bone-equivalent material was used to assess 2D dose distribution for typical clinical treatments such as for breast and head/neck. Calculations of dose to water vs dose to medium and the effects of statistical noise were also evaluated. Results eMC calculations were in acceptable agreement5 with measured scan data down to field sizes of 2 cm 9 2 cm although eMC tended to underestimate surface dose by up to 5%. For planning purposes an optimal grid size of 2 mm and a limit of 750 K histories were used for all eMC calculations. Radiochromic film measurements using anthropomorphic phantoms also resulted in acceptable agreement5 with eMC calculations. Output factors calculated for dose to medium deviated from measurements in water by up to 3.4%. Statistical noise from eMC calculations also had a significant effect on the clinical prescription. Conclusion While dose to medium calculations provided by a validated eMC algorithm can significantly enhance the accuracy of a treatment plan, the therapeutic impact must be carefully assessed for use by clinicians.

Fig. 1 The average Parotid mean dose for VMAT, ssIMRT and Auto-plan VMAT

References 1. Curran BH, Balter JM, Chety IJ (2006) Integrating new technologies into the clinic: Monte Carlo and image-guided radiation therapy. Medical Physics Publishing, Madison, WI 2. Ding GX, Duggan DM, Coffey CW, Shokrani P, Cygler JE (2006) First macro Monte Carlo based commercial dose calculation module for electron beam treatment planning-new issues for clinical consideration. Phys Med Biol. 51(11):2781–99. 3. Indrin J. Chetty et al (2007) Report of the AAPM Task Group No. 105: Issues associated with clinical implementation of Monte Carlo-based photon and electron external beam treatment planning. Med. Phys. 34, 12: 4818-4853 4. Elekta AB, Stockholm, Sweden 5. Van Dyk J, Barnett RB, Cygler JE, Shragge PC (1993) Commissioning and quality assurance of treatment planning computers. Int J Radiat Oncol Biol Phys. 26(2):261–73.

O025 A robustness assessment of Nasopharynx ssIMRT, VMAT and Autoplan VMAT to simulated machine errors E. M. Pogson1, C. R. Hansen2, S. J. Blake1, S. Arumugam3, L. Holloway3, D. I. Thwaites1 1

Institute of Medical Physics, School of Physics, The University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Laboratory of Radiation Physics, Odense University Hospital, Denmark Institute of Clinical Research, University of Southern Denmark ([email protected]). 3 Liverpool and Macarthur Cancer Therapy Centres, Australia ([email protected]), ([email protected]) Introduction Robustness of treatment plans is important as conformity increases. Step-and-Shoot IMRT (ssIMRT) optimised manually

123

Fig. 2 The average dose difference for various plans (VMAT, ssIMRT and Auto-plan VMAT) simulated errors compared to the original plan. Solid red lines indicate tolerance limits and VMAT optimised manually and automatically were compared for nasopharynx patients. Method Ten nasopharynx patients were planned utilising 7 beam ssIMRT and both manual and automatic one-arc VMAT. All plans were completed in Pinnacle version 9.10 for Elekta VersaHD with Agility MLCs. Plans were modified by increments using Python to create simulated error plans; -5 to 5 for gantry and, collimator angles and -5 and +5 mm for MLC shift and MLC field size, considering each parameter separately. Plans were reviewed against acceptable tolerance limits of [10% deviation in parotid mean dose, [5% deviation in PTV D95% and [5% deviation for D1cc for Spinal Cord and Brainstem. Results VMAT errors were slightly higher for the Brainstem and PTV differences. The dose difference exceeded tolerance for mean dose to the parotids for even small errors in the MLC field size (Fig. 1). The gantry angle displayed the smallest source of dose deviation for the Brainstem and Spinal Cord (Fig. 2). PTV Dmean uncertainties, for a large MLC field size of 5 mm, for ssIMRT, VMAT and autoplan were *10, *18 and 23% respectively. ssIMRT was less sensitive to MLC shift errors compared to VMAT and auto-plan VMAT. Conclusion The field size error displayed the largest impact. ssIMRT was least sensitive to MLC shift errors, possibly due to relatively smaller dose in the penumbra region for ssIMRT compared to VMAT fields. The plan robustness of these plans to simulated errors needs to be assessed in context with the optimal plan. Acknowledgements The authors acknowledge support from the Cancer Council NSW project grant RG 14-11.

Australas Phys Eng Sci Med

O026 Variance-based sensitivity analysis study of a prototype beam delivery check system using Monte Carlo simulation O. M. Oderinde, F. C. P. du Plessis Department of Medical Physics, University of the Free State, South Africa. ([email protected] [Presenting Author]), ([email protected]) Introduction Advanced radiotherapy (RT) techniques which have improved the quality of RT have accompanied complexities in their pretreatment quality assurance (QA) and there is huge interest to monitor the real-time treatment. This study applied the EGSnrc/ BEAMnrc Monte Carlo (MC) codes to model an online dose monitoring system, the Integral Quality Monitoring (IQM) device (iRT Systems, Germany) and to study its signal response using a variancebased sensitivity analysis (VSA) technique for small alterations in multileaf collimator (MLC) leaf settings. The IQM is an independent real-time beam treatment verifying system which verifies the integrity and accuracy of the delivered treatment plan. It can also be used as a pretreatment QA tool for radiotherapy e.g. beam output monitoring. Methods The TCL/TK and MORTRAN codes were utilized for coding the IQM component module (CM) which was incorporated into the BEAMnrc interface. The EGSnrc/BEAMnrc MC codes were used to model the Elekta Synergy linac equipped with Agility 160-leaf MLC as well as the newly created CM model of the IQM. Segments of 1 9 1, 3 9 3, 5 9 5 and 10 9 10 cm2 were created and were altered by 1, 2 and 3 mm in terms of its individual leaf positions as defined at the isocenter for 10 MV beams. Simulated output signal was analysed using VSA. Results IQM model was successfully developed. VSA for in to and out of field for 1, 3, 5, and 10 square field was between 0.32–1.79, 0.78–1.19, 0.84–1.14 and 0.93–1.09 for shifts considered. Conclusion IQM signal response per field decreases with increase in field size. This shows that the IQM can be utilized for online dose monitoring for advanced radiotherapy such as IMRT, VMAT and SRT. Acknowledgements This research was sponsored by the Medical Research Council of South Africa, the MRC’s Flagships Awards Project SAMRC-RFA-UFSP-01-2015/ HARD.

O027 The effects of hypoxia on head and neck cancer growth in a 4D cellular growth model with angiogenesis J. C. Forster1,2, M. J. J. Douglass1,2, W. M. Harriss-Phillips1,2, E. Bezak1,3 1

Department of Physics, University of Adelaide, North Terrace, Adelaide, South Australia 5005, Australia. 2Department of Medical Physics, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) 3International Centre for Allied Health Evidence and Sansom Institute for Health Research, Division of Health Sciences, University of South Australia, Adelaide, South Australia 5001, Australia. ([email protected]) Introduction Tumour hypoxia affects tumour growth rate and its response to radiotherapy. The effects of hypoxia on tumour growth have been investigated in an in-house developed computational Monte Carlo model of 4D cellular tumour growth with angiogenesis.

Method A tumour consisting of semirealistic cells and blood vessels is modelled using Matlab. Tumour cell oxygenation is a function of distance to the nearest blood vessel and hypoxic cells have an increased cell cycle time. Cell quiescence is simulated at oxygen tensions \ 1 mmHg. Cells may also become necrotic and be resorbed. To simulate head and neck cancers, a cell hierarchy of stem cells, transit cells and differentiated cells is considered and differentiated cell loss is included. Simulations were performed on the Phoenix supercomputer using as many as 32 cores to observe the effects of hypoxia and necrosis on the rate of tumour growth. Results In accordance with clinical data reported in literature for head and neck cancers [1-4], values of relative vascular volume between 0.7-14%, blood oxygenation between 20 and 100 mmHg and vesselto-necrosis distance between 80–300 lm were considered. This resulted in values of HF10 (fraction of cells with oxygenation \10 mmHg) ranging from 0 to 0.91, values of HF2 (fraction with oxygenation \2 mmHg) from 0 to 0.54, mean oxygenation from 2.0 to 67 mmHg and relative necrotic volume from 0 to 38%. With a probability for stem cell symmetric division of 0.02 and 80% loss of differentiated cells, the doubling time increased from 47 ± 4 days to 209 ± 10 days with increasing amounts of hypoxia and necrosis. Conclusion In work ongoing, tumours grown are finely voxelised and imported into Geant4 for irradiation using track structure methods. By taking into account cellular oxygenation and the formation of hydroxyl radicals, tumour response to photon radiotherapy will be explored in this 4D tumour model for hypoxic tumours versus well oxygenated tumours. References 1. Amelink A, Kaspers OP, Sterenborg HJ et al (2008) Non-invasive measurement of the morphology and physiology of oral mucosa by use of optical spectroscopy. Oral Oncol 44:65–71 2. Pazouki S, Chisholm DM, Adi MM et al (1997) The association between tumour progression and vascularity in the oral mucosa. J Pathol 183:39–43 3. Beasley NJ, Wykoff CC, Watson PH et al (2001) Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 61:5262–7 4. Wijffels KI, Kaanders JH, Rijken PF et al (2000) Vascular architecture and hypoxic profiles in human head and neck squamous cell carcinomas. Br J Cancer 83:674–83

O028 Individualised effective dose for radiotherapy imaging B. Jordan, J. Morton Adelaide Radiotherapy Centre, Adelaide, South Australia, Australia. ([email protected] [Presenting author]), ([email protected]) Introduction Calculation of effective dose requires information of the dose given and the organs within the body receiving this dose. Traditionally, this has been done through NRPB and Monte-Carlo modelling. With modern treatment planning systems, the X-ray source can be modelled in the treatment planning system to calculate dose to the patient for imaging. This combined with auto-contouring with anatomically based systems means that effective dose can be calculated for an individual. Method A kilovoltage photon source model for an Elekta XVI cone beam CT (CBCT) unit was generated in the Philips PinnacleTM

123

Australas Phys Eng Sci Med treatment planning system and validated. Using built-in contouring tools to delineate anatomical structures in the imaging field and recommendations outlined in ICRP 103, individual effective dose calculations have been made for several patients who have undergone kilovoltage CBCT imaging. Results Effective doses due to cone beam CT imaging were calculated for six cases: three (3) head and neck and three (3) pelvic radiotherapy patients, with varying effective dose values for each case despite the use of anatomical site-specific imaging protocols. Conclusion Using built-in tools and a commercial treatment planning system, individualised effective dose can be calculated for patients. This will be more accurate than estimates based on NRPB as the actual organs are considered in the calculation. References 1. Song WY, Kamath S, Ozawa S, Al Ani S, Chvetsov A, Bhandare N, Palta JR, Liu C, Li JG (2008) A dose comparison study between XVI and OBI CBCT systems. Med Phys 35(2):480–486 2. Alaei P, Spezi E (2012) Commissioning kilovoltage cone-beam CT beams in a radiation therapy treatment planning system. J Appl Clin Med Phys 13(6):19–33 3. International Commission on Radiological Protection (2007) The 2007 Recommendations of the International Commission on Radiological Protection. ICRP Publication 103. Ann. ICRP 37:1–332

O029 Improving linear accelerator treatment room shielding with 3D computer aided design G. B. Warr, P. Christiansen, M. West, T. Aland Genesis CancerCare Queensland, PO Box 585, Albion DC, QLD 4010, Australia. ([email protected] [Presenting author]) Introduction Shielding design flaws inherited in an existing linac treatment room limited the use of the linac to ensure equivalent doses in adjacent full-occupancy offices remained within legislated limits of 10lSv per week [1]. The original shielding design did not account for additional expansion of the primary radiation beam of the linac at oblique treatment angles and was performed when higher dose limits were in place. This resulted in safe use of the linac being limited for oblique gantry angles. The purpose of this work was to remediate the shielding to enable unhindered use of a new linear accelerator installed in the treatment room. Method A fully three-dimensional model of the existing shielding design was constructed. Rays were traced from the linac source position to points of interest in adjacent full-occupancy offices outside the treatment room. Shielding calculations were performed with additional shielding added in the ray path to determine the thickness of steel required to ensure equivalent doses at the points of interest remained below legislated limits. Based on the design, 1–1.5 m wide slabs of 150 mm thick steel were added to the edge of the primary barrier. Equivalent dose measurements were made following the guidelines set out in [2], using the 1 m 9 1 m measurement grid scheme described in [3] where appropriate. Results Measurements of the remediated shielding gave equivalent dose results within the legislated limits. Conclusion Three-dimensional CAD tools should be used in linac treatment room shielding design. To ensure oblique angle primary radiation beams are considered, particular attention should be given to primary barrier design in the corners of treatment rooms.

123

References 1. Queensland Government. Radiation Safety Standard PR100: 2010. http://www.health.qld.gov.au/radiationhealth/documents/ pr100.pdf. 2. NCRP Report No. 151. Bethesda, MD 20814-3095, USA, 2005. ISBN-13: 978-0-929600-87-1. 3. Wootton, P., et al., Medical Physics, 2 (1975) 110.

O030 Design of the optimal core material for use in protective garment manufacture Johnny Laban Photon Physics Ltd, Christchurch, New Zealand. ([email protected] [Presenting author]) Introduction Recently published international standards [1–3] for the determination of lead equivalence of protective garments now provide for a broad beam type measurement set-up. As a result, the observed performance of conventional non-lead based materials reduces to levels more reflective of their actual performance when in use. Nonlead materials assembled using a ‘‘laminated’’ structure can still offer significant weight savings, however, if bismuth is used as the exit layer [4]. The purpose of this work is to determine the theoretical optimal construction in terms of elements used and layer thicknesses to achieve maximum lead equivalence for minimum mass. Method A range of elements have been selected which have desirable mass attenuation coefficient characteristics (ie well-spaced photoelectric k-edges), and a spreadsheet based calculation tool has been developed that calculates air kerma attenuation for up to 3 layers of these elements of variable thicknesses, with the final attenuation result expressed in terms of lead equivalence. With this spreadsheet, the problem of establishing the combination of elements that produces the maximum attenuation over the required range of X-ray beams for the minimum of mass can be solved. Results Optimal performance is predicted for a layered combination of gadolinium, tin, and bismuth with lead equivalence varying from 0.35 to 0.42 mm over the range 60–120 kVp for a combined mass equal to only 0.30 mm of lead. Tungsten, despite being in commercial use is found to be unsuitable, principally because its photoelectric k-edge sits just above the characteristic x-rays present in the test x-ray spectra from 70 kVp and above. Conclusion Improvements to the products currently available for protective garment manufacture are theoretically possible. Availability, chemical suitability, and cost of raw materials remains the next hurdle in terms of implementing these improvements in actual manufacturing. References 1. DIN 6857-1 (2009). Radiation protection accessories for medical use of X-radiation – Part 1: Determination of shielding properties of unleaded or lead reduced protective clothing. 2. ASTM F3094-14 (2014). Standard Test Method for Determining Protection Provided by X-ray Shielding Garments Used in Medical X-ray Fluoroscopy from Sources of Scattered X-rays. 3. IEC 61331-1 (2014). Protective devices against diagnostic medical X-radiation – Part 1: Determination of attenuation properties of materials. 4. McCaffrey JP, Mainegra-Hing E, Shen H (2009) Optimizing nonPb radiation shielding materials using bilayers. Med. Phys. 36 (12) 5586–5594.

Australas Phys Eng Sci Med

O031 Considerations of radiation shielding in nuclear medicine facilities – A technical review F. Salehzahi, J. Tse Medical Physics and Radiation Engineering, Canberra Hospital, Canberra, Australia. ([email protected] [Presenting author]), ([email protected]) Introduction While radiation shielding methodologies for PET/CT systems are well established [1], we recognised a need for standardised approaches in Australia for shielding consideration for Nuclear Medicine facilities including SPECT/CT system and gamma camera rooms, hot laboratory and patient waiting areas. The aim of this review is to summarise existing literatures on this topic and share the authors’ experience and views of shielding considerations in Australia and other countries, and also to highlight the impact of cultural considerations in shielding design. Method A thorough literature search was conducted to summarise and compare the radiation dose rate constants (G) and attenuation information (i.e. HVL and TVL of lead) for a range of commonly used radionuclides in Nuclear Medicine departments. An iterative application was also proposed to determine and streamline the optimal radiation shielding thickness needed for SPECT/CT camera room which can be further modified to account for shielding other areas in a Nuclear Medicine facility. The second section of the review would focus on the ‘patient flow and design concepts’ which must include PET, per authors’ experiences where examples of practical shielding which aims for a cost-effective design will be provided. Discussion and Conclusion The literature survey has revealed that there are various definitions of radiation dose rate constants and attenuation information where their values can differ significantly depending on their experimental or computational methods. These differences have to be appreciated by medical physicists to avoid propagating systematic errors in the shielding calculations. In addition, practical examples have demonstrated that the proposed iterative method is capable of producing optimum shielding results. Comparisons with other conventional methods were also made. Last but not least, the review has also provided an overview of the different radiation safety cultures, patient flow and economic considerations and their influences on designing a cost-effective Nuclear Medicine premises through a number of practical examples. References 1. Madsen M. T. et al. (2006) AAPM Task Group 108: PET and PET/CT Shielding Requirements. Med Phys 33(1): 4 – 15

the age of 75, and at least 30 years after cessation of work resulting in occupational exposure. The ANRDR is administered by ARPANSA as part of its role to protect the health and safety of people and the environment from the harmful effects of radiation, and to promote national uniformity in radiation protection practices across all Australian jurisdictions. Method The Dose Register has been open to receive dose data from the Australian uranium mining and milling industry since 2010 and is now in the process of expanding coverage to the mineral sands mining and processing industry, Commonwealth organisations and the aviation sector through targeted engagement. In the near future, ARPANSA will commence engagement with the medical sector and will aim to work collaboratively with industry in a pilot phase of the ANRDR’s expansion program, prior to launching to the entire sector. Results The ANRDR provides a single uniform national approach to the management of radiation dose records for workers and ensures the longevity of records for the long-term. In turn, this will ensure that records remain available to workers and that analysis of data in the ANRDR will facilitate optimisation of radiation protection programs. Conclusion This presentation will provide an analysis of the data contained in the ANRDR in the context of dose limits, and discuss the capabilities, benefits and challenges in establishing and expanding the ANRDR to the medical sector. The expansion of the ANRDR to cover all occupationally exposed workers in Australia would be the desired outcome and would ensure that operation of a national dose register is consistent with international best practice.

KS06 Accreditation of radiation protection experts and medical physics experts in the UK medical sector D. G. Sutton Medical Physics Dept, Ninewells Hospital & Medical Scholl, Dundee, DD19SY, UK. ([email protected] [Presenting author]) Outline UK regulation is informed by EU Directives. These require the appointment of Radiation Protection Experts and also Medical Physics Experts. To take RPEs first, the three professional societies that are concerned with Radiation Safety in the UK (IPEM, SRP, AURPO) jointly established an organisation - RPA2000 - to devise and implement a competence based scheme to recognise experts in Radiation Safety and Radioactive Waste Management across all sectors of industry, including Medicine and Universities. The presentation describes how the scheme is implemented in terms of (a)

O032 Australian National Radiation Dose Register (ANRDR) in review B. Paritsky Australian Radiation Protection and Nuclear Safety Agency (ARPANSA). ([email protected]) Introduction The ANRDR is a centralised database designed for the collection, storage and dissemination of dose records for occupationally exposed workers. In alignment with international best practice, these records should be maintained until the worker attains

The assessment of: (i) knowledge of theory and facts; (ii) skills; (iii) competence; (iv) experience

(b) Alignment with regulatory requirements (c) Approval and audit by regulators. (d) The recruitment and training of suitable assessors. (e) The provision of clear procedures for applicants, assessors and the governance process. (f) Impartiality and proportionality. (g) The process for assessing continued competence of experts. (h) Complaints process.

123

Australas Phys Eng Sci Med The role of the Medical Physics Expert involves the optimisation of Medical Exposures and has to date been less well defined than that of the RPE. However, the revised and harmonised EU BSS requires member states to put in place processes for recognising Medical Physics Experts as well as the Radiation Protection Experts described above. An MPE is an ‘‘individual or a group of individuals, having the knowledge, training and experience to act or give advice on matters relating to radiation physics applied to medical exposure, and whose competence in this respect is recognised by the competent authority’’. This means that a formal recognition scheme has to be put in place. At the time of writing a formalised scheme is being developed and there is considerable debate as to the shape it will take, but it will need to be established by the end of 2017.

PDD (%)

350 300

Open 1.5 T

250

Open 0 T

200 150 100 50 0 0

50

100 Depth (mm)

150

200

Fig. 1 Open field PDD measured in 0 and 1.5-T magnetic fields. Ormalisation point is at 50 mm depth

O033 The Australian MRI-Linac Program: Commissioning of the mega voltage X-ray source inside a 1.5 T MRI J. Begg1,2,3, A. George1,2, S. Alnaghy2,4, T. Causer2,4, T. Alharthi2,5, B. Dong1,2, G. Goozee1,2, G. Liney1,2, L. Holloway2,6,7,8,9, P. Keall2,8

spread; however response was within IEC criteria. No difference in beam quality was observed between 0 and 1.5-T. PDDs in a 1.5-T field show an over response near the surface relative to 0-T due to electron contamination (Fig. 1). Absolute dosimetry output at 1.5-T was approximately 3% lower relative to output at 0-T. Conclusion Dosimetric measurements of the Linatron were performed on a 1.5-T ex-clinical MRI. This work will be used to refine monte-carlo models and for comparison to 1.0-T measurements made with the permanent bespoke magnet developed for the Australian MRI-Linac program.

1

Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Ingham Institute for Applied Medical Research, Australia, ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3 South Western Sydney Clinical School (SWSCS), University of New South Wales, Australia. 4Centre for Medical Radiation Physics (CMRP), University of Wollongong (UOW), Australia. 5Institute of Medical Physics, School of Physics, University of Sydney, Sydney, NSW, Australia. 6Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Australia CMRP, UOW, Australia. 7Institute of Medical Physics, University of Sydney, Australia. 8Sydney Medical School, University of Sydney, Australia 9 South Western Sydney Clinical School (SWSCS), University of New South Wales, Australia Introduction The Australian MRI-Linac Program[ 1] aims to develop a 1-T open-bore MRI/6-MV linac. Commissioning of a Varian Linatron MP (Varian Medical Systems, Palo Alto, USA) outside of a magnetic field was previously presented [2]. The Australian MRILinac program obtained an ex-clinical 1.5-T Siemens Sonata as a temporary magnet. Commissioning of the 6-MV X-ray source inside a high strength magnetic field presented a number of challenges compared to a standard clinical linear accelerator. This work aimed to commission the Linatron within the magnetic field and compare results to dosimetric measurements acquired in a 0-T field. Method Commissioning measurements inside the magnetic field included monitor chamber reproducibility and linearity with dose, beam quality (TPR20/10), Total Scatter Factors, Percentage Depths Doses (PDD), Beam Profiles and absolute dosimetry. PDD and beam profiles were acquired using gafchromic film in solid water. Absolute dosimetry measurements at 0 ad 1.5-T were normalised using a farmer chamber positioned at a low magnetic field strength point. Results Monitor chamber reproducibility was predominately below the 0.5 % IEC criteria for both 0 and 1.5-T measurements. Monitor chamber response with dose had a larger uncertainty and greater

123

References 1. Keall PJ, Barton M, Crozier S (2014) The Australian Magnetic Resonance Imaging–Linac Program. Seminars in Radiation Oncology 24(3):203–206. doi:10.1016/j.semradonc.2014.02.015 2. Begg J, George A, Alnaghy S, Causer T, Alharthi T, Glaubes L, Dong B, Goozee G, Liney G, Holloway L, Keall P (2015) The Australian MRI-Linac Program: Commissioning of the mega voltage X-ray source. Engineers, Physicists and Scientists in Medicine.

O034 Experimental results from the Australian MRI-Linac G. P. Liney1, B. Dong1, J. Begg2, K. Zhang1, P. Vial, F. Lee3, L. Holloway2, M. Barton1, S. Crozier4, P. Keall3 1

Ingham Institute for Applied Medical Research. ([email protected] [Presenting author]), (Bin.dong@ sswahs.nsw.gov.au), ([email protected]), ([email protected]). 2Liverpool Cancer Therapy Centre. ([email protected]), ([email protected]), ([email protected]). 3University of Sydney. ([email protected]), ([email protected]) 4 University of Queensland, Brisbane ([email protected]) Introduction The pursuit of real-time image guided radiotherapy using optimal tissue contrast has seen the development of several hybrid MRI-treatment systems, high field and low field, and inline and perpendicular configurations. As part of a new MRI-Linac program, an MRI scanner was integrated with a linear accelerator to enable investigations of a coupled inline MRI-Linac system. This work describes results from a prototype experimental system to demonstrate the feasibility of a high field inline MR-Linac.

Australas Phys Eng Sci Med

Fig. 1 Schematic diagram of experimental set-up

Fig. 1 Photograph of the phase#1 system with the treatment unit at the middle position. The magnet can be seen through the recess in the cage wall (covers removed). Inset: first beam on image of a kangaroo steak Methods A 1.5 T magnet (Sonata Siemens) was located in a purpose built RF cage enabling shielding from and close proximity to a linear accelerator with inline (and future perpendicular) orientation. A portable linear accelerator (Linatron, Varian) was installed together with a multi-leaf collimator (Millennium, Varian) to provide dynamic field collimation and the whole assembly built onto a stainless-steel rail system. A series of MRI-Linac experiments was performed to investigate: (1) image quality with beam on measured using a macropodine (kangaroo) ex vivo phantom; (2) the noise as a function of beam state measured using a 6-channel surface coil array and; (3) electron contamination effects measured using Gafchromic film and an EPID. Results (1) Image quality was unaffected by the radiation beam with the macropodine phantom image with the beam on being almost identical to the image with the beam off. (2) Noise measured with a surface RF coil produced a 25% elevation of background intensity when the radiation beam was on. (3) Film and electronic portal image device (EPID) measurements demonstrated electron focusing occurring along the centerline of the magnet axis (Fig. 1). Conclusion A proof-of-concept high-field MRI-Linac has been built and experimentally characterized. This system has allowed us to establish the efficacy of a high field inline MRI-Linac and study a number of the technical challenges and solutions.

O035 Incorporating an EPID on the Australian MRI- Linac: First experiments F. Lee1, J. Begg3, G. Liney2, B. Dong2, P. Vial1 1

Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected]), ([email protected] [Presenting author]). 2Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute, Australia. ([email protected]), ([email protected]). 3 South Western Sydney Clinical School (SWSCS), University of New South Wales, Australia. ([email protected]) Introduction This work presents preliminary experiments to assess the feasibility of integrating an EPID X-ray imaging system with the Australian MRI-Linac [1].

Fig. 2 Qualitative image tests of MRI (middle) and EPID (right) images acquired simultaneously on the MRI-LINAC using a multimodality imaging phantom (left) Method The first prototype Australian MRI- Linac consisted of an exclinical Siemens MRI (1.5T Sonata) and a Varian Linatron accelerator, aligned in the ‘in-line’ configuration with the X-ray beam parallel to the MRI’s magnetic field. A Perkin Elmer EPID and software were used for X-ray imaging. The EPID was positioned on the beam exit side of the MRI bore (Fig. 1). The simultaneous operation of the MRI and EPID were tested, and the mutual impact on image quality was assessed for both imaging systems. Procedures for using the EPID to verify the alignment of the X-ray beam and MRI bore axis, and the MLC position calibration were developed. Results The magnetic field at the EPID position was measured to be approximately 0.08 T. The EPID FOV projected at the MRI bore centre was limited to 26 cm 9 26 cm, similar to conventional linacs. The SNR of MRI images was approximately constant with and without the EPID operating, after some initial problems with noise from the EPID power supply was addressed. The MRI had no discernable impact on EPID functionality or image quality (Fig. 2). The coincidence of the X-ray beam and MRI bore central axis was confirmed to within 3 mm. Preliminary tests confirm the EPID can be used to verify MLC position calibration. Conclusion The feasibility of integrating a standard EPID system with the Australian MRI- Linac has been demonstrated. Acknowledgements This work was supported from Cancer Institute NSW Australia (Research Equipment Grant 10/REG/1-20) References 1. Keall PJ, Barton M, Crozier S (2014) The Australian Magnetic Resonance Imaging-Linac Program. Seminars in Radiation Oncology 24(3):203–206

O036 MRI in radiotherapy treatment planning: Geometric accuracy considerations A. Walker1, P. Metcalfe2,3, G. Liney1,2,3,4, V. Batumalai1,2,5, K. Dundas1,2,3, G. Delaney1,5,6,7,

123

Australas Phys Eng Sci Med

M. Boxer1,5, M. Yap1,2,5,6,7, J. Dowling8, D. Rivest-Henault8, E. Pogson1,2,3,4, L. Holloway1,2,3,4,5

Table 2 Dosimetric impact of distortions on breast plans Dose criteria

Mean difference ± 2 r (phantom)

1

Liverpool and Macarthur Cancer Therapy Centres, Liverpool, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Ingham Institute for Applied Medical Research, Liverpool Hospital, Sydney, Australia. ([email protected]), ([email protected]). 3 Centre for Medical Radiation Physics, University of Wollongong, Wollongong, Australia. 4Institute of Medical Physics, School of Physics, University of Sydney, Sydney, Australia. 5South Western Clinical School, University of New South Wales, Sydney, Australia. ([email protected]). 6Collaboration for Cancer Outcomes Research and Evaluation, Liverpool Hospital, Liverpool, Australia. ([email protected]). 7School of Medicine, University of Western Sydney, Sydney, Australia. ([email protected]). 8Commonwealth Scientific and Industrial Research Organisation Computational Informatics, Australian E-Health Research Centre, Brisbane, Australia. ([email protected]), ([email protected]) Introduction Inclusion of MRI in radiotherapy is increasing with improved image quality compared to CT allowing improvements in anatomical delineation. Previous work highlighted the need to quantify systematic distortions on scanners used for radiotherapy treatment planning [1,2]. Distortions vary within scanners and vary between anatomical sites. This work investigates the impact of systematic and patient-induced distortions on whole breast radiotherapy planning [3]. Method CT and MRI datasets for 18 breast radiotherapy patients were acquired. Patient CTs were deformably registered to the corresponding patient MRI. This deformation field quantified geometric distortion from the system and patient, plus any setup errors. An inhouse phantom was scanned on CT and a 3T Siemens Skyra. The images were deformably registered to quantify systematic distortions [2]. Original patient CTs were deformed by this deformation field, simulating systematic distortions. Whole breast radiotherapy treatment plans were then generated on the patient CT: – deformed simulating the systematic distortions – deformed simulating the combined systematic and patient distortions (with setup errors).

Table 1 Anatomical sites location ranges and the expected systematic distortions Anatomical site

Radial distance from scanner isocenter (mm)

Distortion range (mm)

Brain stem

0–75

0–1

Cervix

0–80

0–1

Prostate Larynx

0–85 0–95

0–1 0–1.5

Oral Cavity

0–145

0–2

Lung

0–180

Breast

10–300

* * Note: These distortion values are scanner dependent

123

0–2.5 0–4

Mean difference ± 2 r (patient)

Right PTV

V47.5 Gy

0.2 ± 1.2 %

0.3 ± 2.5 %

Ipsilateral Lung

V20 Gy

-0.1 ± 0.4 %

0.8 ± 2.7 %

CB

Max 1 cc

-0.0 ± 0.1 Gy

0.1 ± 0.6 Gy

PTV

V47.5 Gy

-0.8 ± 1.6 %

-0.7 ± 1.9 %

Ipsilateral Lung CB

V20 Gy Max 1 cc

-0.2 ± 0.5 % -0.0 ± 0.1 Gy

-0.3 ± 1.8 % 0.3 ± 2.4 Gy

Plan pass/fail

Overall

18 4/07

10 4/87

Left

CB = Contralateral Breast, 4 plan met acceptable criteria, 7 plan did not meet acceptable criteria

These plans were copied onto the original patient CT, assessing the impact of distortions on the treatment planning process. Breast was investigated as previous work indicated distortions may be worse in this region of the scanner (Table 1) [1]. Results Table 2 shows variations in dose criteria between plans generated on the distorted images to the distorted plan copied to the original CT image. Conclusion Systematic and patient distortions combined with setup errors may result in unacceptable dosimetric variations for whole breast IMRT if considering MRI-only radiotherapy treatment planning. Residual system distortions alone had minimal impact on plan dosimetry. Impact of system distortion is expected to be less for other anatomical sites, however further work is required. Further investigations quantifying patient distortions are needed to ensure minimal dosimetric impact for MRI-only planning, particularly around airtissue interfaces (lung) where susceptibility variations may produce larger distortions. References 1. Walker A, Liney G, Metcalfe P, Holloway L (2014) MRI distortion: considerations for MRI based radiotherapy treatment planning. Australas Phys Eng Sci Med 37 (1):103-113. doi: 10.1007/s13246-014-0252-2 2. Walker A, Liney G, Holloway L, Dowling J, Rivest-Henault D, Metcalfe P (2015) Continuous table acquisition MRI for radiotherapy treatment planning: Distortion assessment with a new extended 3D volumetric phantom. Medical Physics 42 (4):19821991. doi:10.1118/1.4915920 3. Walker A, Metcalfe P, Liney G, Batumalai V, Dundas K, GlideHurst C, Delaney GP, Boxer M, Yap M, Dowling J, RivestHenault D, Pogson E, Holloway L (2016) MRI geometric distortion: Impact on tangential whole breast IMRT. Journal of Applied Clinical Medical Physics (In press)

O037 MRI-compatible patient rotation system – Design, construction, and first use Brendan Whelan1, Gary Liney2, Jason A. Dowling3, Robba Rai4, Rob Wilkins5, Lois Holloway2, Leigh McGarvie2, Ilana Feain1, Michael Barton6, Paul Keall1

Australas Phys Eng Sci Med 1

Radiation Physics Laboratory, University of Sydney, Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Ingham Institute for Applied Medical Research, Liverpool, Australia. ([email protected]), ([email protected]), ([email protected]). 3 Australian E-Health Research Centre, CSIRO, Australia. ([email protected]). 4Liverpool Cancer Therapy Centre, Liverpool, Australia ([email protected]). 5Biomech engineering, Sydney, NSW. ([email protected]). 6Institute of Clinical Neurosciences, Royal Prince Alfred Hospital, Sydney, Australia. ([email protected]) Introduction Patient rotation could greatly simplify radiotherapy delivery engineering by enabling fixed beam systems, with particularly important ramifications for hadron, MRI-Linac, and low cost therapy. A substantial barrier to clinical implementation of patient rotation is quantifying and mitigating the dosimetric impact of rotation induced organ motion. To date, very little data exists which can address this issue. In this work, we present the design and first use of a novel MRI-compatible decubitus patient rotation system (PRS).

Imaging (MRI)-Alone External Beam Radiation Therapy From Standard MRI Sequences. International Journal of Radiation Oncology* Biology* Physics, 2015. 93(5): p. 1144-1153.

KS07 MRI guided radiation therapy Jan-Jakob Sonke Department of Radiation Oncology, The Netherlands Cancer Institute, Amsterdam, The Netherlands. ([email protected]) Image guided radiotherapy aims to improve the accuracy and precision of treatment delivery. Various technological evolutions have emerged over the past few decades ranging from port films, electronic portal imaging devices, ultra-sound, kV planar imaging and cone beam CT. The latest innovation is the integration of MRI with radiation delivery devices. In this presentation our initial experience with such a MRI guided system will be discussed.

O038 Quantification of lung tumor translation and rotation during lung SABR V. Caillet1,2, J.-H. Kim2, N. Hardcastle1, P. J. Keall2, J. T. Booth1,3 1

Northern Sydney Cancer Centre, Royal North Shore Hospital, Australia. ([email protected]), ([email protected]). 2School of Medicine, University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) 3 School of Physics, University of Sydney, Australia

Method The PRS was designed and constructed in-house to be compatible with our 3T MRI scanner (Skyra, Siemens). Turbo-spinecho weighted prostate images of a healthy male volunteer were taken at 45 increments over 360. Each image took around two minutes to acquire. The body, prostate, bladder, and rectum were auto-contoured using the algorithm of Dowling, et. al. [1]. Images were registered to the 0 degree image in two steps; first, a rotation step which compensated for the rotation angle; second a deformable registration which accounted for the resultant organ motion. The deformation vectors within each organ were extracted and analysed. Results The maximum deformation compared to the 0 degree image occurred at 90, with mean deformation in prostate, bladder and rectum of 13.1 ± 2.3, 11.7 ± 4.6 and 14.3 ± 2.8 mm respectively. The minimum deformation occurred at 180; here the deformation was 6.1 ± 1.2, 4.2 ± 3.9, and 5.6 ± 1.7 mm. The small deviations relative to the means show that the principle component of organ deformation was rigid. High quality images were obtained with the PRS, verifying MRI compatibility. Conclusion An MRI-compatible PRS has been designed, installed and successfully implemented at Liverpool hospital. This device will enable data collection necessary to develop decubitus patient rotation strategies, which could globally impact the practice of radiotherapy. References 1. Dowling, J.A., et al., Automatic Substitute Computed Tomography Generation and Contouring for Magnetic Resonance

Introduction During lung SABR treatment, variation in respiratory motion may lead to geographic miss and underdosing the target. Understanding the magnitude of tumour translation and rotation contributes to reliable ITV and PTV margins. The purpose of this study is to analyze lung tumor motion to determine the magnitude of the tumor translation and rotation during treatment and evaluate the adequacy of PTV margin. Method Tumor motion was obtained from six patients treated with implanted Calypso beacons as part of the LIGHT SABR clinical trial. Patients had three Calypso beacons implanted. We simulated delivery of standard, no tracking, and treatment delivery for this study. The coordinates of each beacon were used to calculate tumour rotation using a closed-form least squares method [1]. Baseline position was defined as the average position of the tumor over *5 breathing cycle taking the planned position as reference. The motion patterns of the lung tumour for each patient and across the cohort are investigated. We assume the Calypso beacons are a good estimate of tumour position and pose. Results Mean position across the cohort was 1.1 mm ± 0.5 mm. Tumour translation across the cohort showed superior-inferior (SI), anterior-posterior (AP) and left-right (LR) motion greater than 2 mm 18%, 9% and 8% of the treatment time, respectively (Fig. 1). The amount of time with motion greater than 5 mm, representing the PTV margin, was less than 1% of the treatment time and would not have led to target being underdosed without tracking. Rotation around SI, AP and LR shows that the tumor is subjected to rotation above 5 4.2%, 3.2% and 2.6% of the time. Fig. 2 provides an example of rotation and translation during treatment.

123

Australas Phys Eng Sci Med

Fig. 1 Probability density functions of the tumor rotational and translation component during lung SABR summarized for six patients. Each graph is read as the probability for the tumor to have a rotation superior to 5, or translation from baseline superior to 2 mm. Treatment times were in average 3 min per fraction and per patient

Fig. 2 Example of a patient tumor motion during one fraction (two arcs) for the rotation (a) and tumor motion (blue) with baseline position (red) (b). Rotation around the AP direction (left) is shown 20 s of treatment with rotation up to 8 during breathing. Tumour translation in the SI direction over the course of an entire fraction shows that tumour translation can be quite erratic Conclusion Analysis of tumor motion suggests that the 5 mm margins associated with the PTV were sufficient to cover the magnitude of motion experienced during treatment. Rotations above 5 were observed and may affect the target dose distribution suggesting the potential need to define margins to account for rotation at planning. References 1. Horn, Berthold KP. ‘‘Closed-form solution of absolute orientation using unit quaternions.’’ JOSA A 4.4 (1987): 629–642.

O039 A respiratory phantom-based comparison of measured and Pinnacle3 lung-tumour dose in conformal SBRT beams S. Zawlodzka, J. Baines, M. Parfitt Radiation Oncology Mater Centre, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Stereotactic Body Radiation Therapy (SBRT) has recently been introduced to our clinic for the treatment of non-small

123

cell lung cancer. Pinnacle3 SBRT treatment plans are calculated using ten non-coplanar 6 MV conformal beams with a free-breathing CT data set and 4DCT images used for ITV delineation. The purpose of this work is to compare Pinnacle3 and measured dose in moving lungtumour surrogates. Method A QUASARTM Respiratory Motion Phantom (Modus Medical Devices Inc., London, CA) with moving split lung inserts containing GAFCHROMICTM EBT3 film (Ashland Inc., Covington, KY) and either a 1 or 3 cm diameter split tumour were used to obtain coronal plane dose distributions. Programmable respiratory motion in the craniocaudal direction, amplitudes of 5 and 15 mm and periods of 4 or 8 s each, simulated clinical tumour motions in lung. Pinnacle3 and measured dose profiles were compared for single jaw defined fields of 5 9 8 cm2 (3 cm tumour) and 3 9 6 cm2 (1 cm tumour) that are representative of our clinical stereotactic conformal fields. Additionally dose profile comparisons were obtained for a clinically representative non-coplanar SBRT treatment consisting of 10 beams with the treatment isocentre coincident with the tumour centre. Results For moving tumours in single jaw defined fields, S-I and L-R dose profiles within the tumour are in accord with Pinnacle3. However differences outside the tumour in the S-I direction are evident with 15 mm amplitude motion showing the largest, independent of the period of respiratory motion. Composite profiles for the SBRT treatments demonstrate features similar to the rectangular fields with 15 mm amplitude motions for both tumour sizes showing the greater discrepancy from Pinnacle3 in lung. Conclusion EBT3 film based measurements of tumour dose are consistent with Pinnacle3 calculated dose. However differences in lung dose distribution are evident between Pinnacle3 and film.

O040 Density overrides for the target volumes in lung SBRT planning G. Healy, A. T. Cousins Medical Physics and Bioengineering Department, Canterbury District Health Board, New Zealand ([email protected] [Presenting author]) Introduction SBRT is used at Christchurch Hospital to treat appropriate early stage lung tumours following a clinical protocol. Patients receive a phase binned 4DCT scan. A Maximum Intensity Projection (MIP) is used to determine the ITV, with the dose distribution calculated on an average or typical phase CT scan. Accurately modelling dose to the target volumes is a problem due to the effects of target motion and the difficulties in modelling dose deposition to relatively high density tumour targets located in low density lung material. Overriding the densities of the ITV or PTV may help improve dose calculation accuracy for moving targets, with the aim of achieving the desired dose within the GTV. Method A QUASAR Respiratory Motion Phantom was used to produce standard breathing patterns and reproduce patient breathing traces collected by a Varian RPM. Custom-made cylindrical cork inserts with embedded Perspex hemispheres of different diameters to mimic tumours were designed for use with EBT3 film and a CC13 Ion Chamber. 4DCT scans were collected for each insert and breathing pattern combination. MIPs were produced on the CT scanner then exported to the Monaco TPS along with a typical phase scan for treatment planning. Results Plans generated on each 4DCT scan include: a standard density plan, an ITV density overridden plan, a PTV density overridden plan, and a hybrid plan where the ITV was set to water density and the PTV to a density between lung and water.

Australas Phys Eng Sci Med The table shows representative results for three Conformality Indices (CI) and a Heterogeneity Index (HI) for each type of plan and tumour size. CI (Monaco)

CI (100%)

CI (50%)

HI

30 mm Typical Phase

0.85 ± 0.02

0.95 ± 0.02

3.79 ± 0.06

1.33 ± 0.03

ITV ED = 1.00

0.85 ± 0.02

0.93 ± 0.01

3.67 ± 0.09

1.34 ± 0.02

PTV ED = 1.00

0.91 ± 0.01

1.09 ± 0.01

3.72 ± 0.08

1.24 ± 0.02

ITV ED = 1.00

0.92 ± 0.02

1.06 ± 0.01

3.70 ± 0.10

1.26 ± 0.02

PTV ED = 0.650 15 mm Typical Phase

0.70 ± 0.06

0.70 ± 0.05

4.28 ± 0.16

1.38 ± 0.02

ITV ED = 1.00

0.65 ± 0.05

0.66 ± 0.03

4.01 ± 0.13

1.44 ± 0.03

PTV ED = 1.00

0.91 ± 0.01

0.97 ± 0.03

3.78 ± 0.17

1.27 ± 0.02

ITV ED = 1.00

0.89 ± 0.02

0.91 ± 0.04

3.85 ± 0.20

1.30 ± 0.02

PTV ED = 0.650

Conclusion The optimal value for the density override is dependent on the tumour size and the PTV size. This work will determine the most appropriate ITV and PTV densities to be set in the TPS for future SBRT patients at Christchurch Hospital.

Fig. 1 Histogram distribution of treatment plans for varying magnitude of errors in a JW, b CS, c GP, and d GSP. Red dashed bars and green lines show the machine interlock thresholds and machine QA thresholds for each error type respectively

References 1. D. Wiant, C. Vanderstraeten, J. Maurer, J. Pursley, J. Terrell, and B. J. Sintay, ‘‘On the validity of density overrides for VMAT lung SBRT planning,’’ Med. Phys. 41(8) 081707 (2014)

O041 Clinical significance of treatment delivery errors for helical TomoTherapy lung stereotactic ablative radiotherapy plans - a dosimetric simulation study S. Deshpande1,2, M. W. Geurts3, C. R. Hansen4,5, P. Metcalfe1,2, P. Vial1,4, D. Thwaites1,4, L. Holloway1,2,4,6 1

Liverpool and Macarthur Cancer Therapy Centre and Ingham Institute for Applied Medical Research, Sydney, Australia. 2Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. ([email protected]), ([email protected] [Presenting author]). 3Department of Human Oncology, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA. ([email protected]). 4Institute of Medical Physics, School of Physics, University of Sydney, NSW, Australia. ([email protected]), ([email protected]). 5Laboratory of Radiation Physics, Odense University Hospital, Denmark. ([email protected]). 6 South West Sydney Clinical School, School of Medicine, University of NSW, Australia. ([email protected]) Introduction This study simulated potential helical tomotherapy (HT) delivery errors for lung stereotactic ablative radiotherapy (SABR) plans and quantified the clinical impact. Method Ten HT lung SABR plans were optimised using a fixed 2.5 cm jaw width(JW), 0.286 pitch and 1.6–1.8 modulation factor using the clinical HT planning system. Each plan was edited using TomoTherapyV5.0 GPU dose calculator to introduce systematic errors in JW, couch speed(CS), gantry period(GP), gantry start

position(GSP), one MLC leaf stuck open (either leaf 32, 42 or 52) or random errors in multi leaf collimator leaf open time(MLC-LOT) with SD of 1–10% with a mean MLC LOT error of zero. Plans modified for each error were compared to clinical thresholds as outlined RTOG 0915 [1]. Errors causing violations in clinical thresholds were compared to machine interlock thresholds and quality assurance (QA) tolerances as outlined in AAPM report 148 [2]. Results Clinical thresholds were exceeded with errors in JW [ 0.25 mm, CS [ 0.5%, GP [ 0.2 s and GSP [ 1 (Fig. 1). Clinical impact of MLC-LOT random error up to 4% and MLC leaf 32 stuck open were not consistent across treatment plans. All errors compromised PTV conformity. Clinically relevant errors other than CS and the secondary interlock for JW position and GP are prevented by machine interlocks. Maintaining the JW and GSP accuracy below recommended QA [2] tolerances would prevent clinically relevant error. Conclusion This work has established a method to characterize HT machine delivery errors and their clinical significance. This approach can be used to validate robustness of institution-specific planning methods for a given machine QA tolerance. References 1. G.M. Videtic, C. Hu, A.K. Singh, J.Y. Chang, W. Parker, K.R. Olivier, S.E. Schild, R. Komaki, J.J. Urbanic, H. Choy, ‘‘A Randomized Phase 2 Study Comparing 2 Stereotactic Body Radiation Therapy Schedules for Medically Inoperable Patients With Stage I Peripheral Non-Small Cell Lung Cancer: NRG Oncology RTOG 0915 (NCCTG N0927),’’ International Journal of Radiation Oncology* Biology* Physics 93, 757–764 (2015). 2. K.M. Langen, N. Papanikolaou, J. Balog, R. Crilly, D. Followill, S.M. Goddu, W. Grant III, G. Olivera, C.R. Ramsey, C. Shi, ‘‘QA for helical tomotherapy: Report of the AAPM Task Group 148a),’’ Medical physics 37, 4817–4853 (2010). Acknowledgements NSW cancer council grant project number 1067566. The authors also acknowledge Accuray Incorporated for providing the research TPS used for this work.

123

Australas Phys Eng Sci Med

O042 VMAT-FFF Lung-SBRT QA: A method for routinely assessing the dosimetric impact of the interplay effect Flavio Enrico Nelli, Jeffrey Roy Harwood Andrew Love Cancer Centre, University Hospital Geelong, Geelong VIC Australia. ([email protected]) [Presenting author]) ([email protected]) Introduction The interplay effect in FFF-VMAT SBRT deliveries produces noticeable discrepancies between planned and delivered doses [1]. A quality assurance program (QA) needs to incorporate procedures for ensuring these discrepancies are within clinically acceptable limits. The impact of motion during delivery can be experimentally assessed on a moving phantom. Alternatively, a simple solution exists for modelling the dose distribution within the treatment planning system for the case of cylindrically symmetric phantoms moving axially. This solution provides a versatile tool for interplay effect management. We use this novel methodology for assessing the impact of the interplay effect on target coverage on VMAT-FFF-based Lung-SBRT treatment deliveries. Method For VMAT-FFF treatments planned on a cylindrically symmetric phantom comprising a target moving parallel to the symmetry axis and perpendicular to the VMAT arcs’ plane, dose distributions calculated moving the arcs’ plane with respect to a fixed target are equivalent to those calculated with a moving target and fixed arcs’ plane. Using a moving arcs’ plane to evaluate dynamic deliveries on such phantoms enables assessment of the interplay effect on dynamic deliveries. We developed software to split each VMAT arc into partial arcs synchronized with the moving target [2]. Subsequently, dose deposition on a moving phantom is calculated positioning each partial arc on the 3DCT image set of the phantom (fixed target) at different axial positions corresponding to the target movement pattern. VMAT-FFF SBRT plans, delivering 48 Gy in 4 fractions or 28 Gy in 1 fraction, were planned on a Quasar phantom [3] fitted with a cedar insert comprising a 3 cm or 1 cm diameter spherical target. PTV was defined as a 5 mm external margin from the ITV, which was defined as the superposition of all target positions [4]. Realistic 14BPM breathing traces with 1, 2 and 3 cm amplitude were considered. The impact of the interplay effect on target coverage was analyzed by using 4 different starting points evenly distributed on a breathing period. Results Beam profiles corresponding to each delivery synchronism where compared with target dimensions (GTV+5 mm). The interplay effect produced some noticeable distortions of high doses, however, these distortions did not compromise target coverage even in the extreme case of 3 cm amplitude movement. Conclusion The methodology presented here provides an easy and reliable way of assessing the impact of the interplay effect on dose distributions which can easily be incorporated in a treatment QA program. This method has been implemented as part of the Andrew Love Cancer Centre’s QA program for VMAT-FFF Lung-SBRT treatments.

123

References 1.

X. Li, Y. Yang, T. Li, K. Fallon and D. E. Heron, ‘‘Dosimetric effect of respiratory motion on volumetric-modulated arc therapy–based lung SBRT treatment delivered by TrueBeam machine with flattening filter-free beam,’’ Xiang Li,a Yong Yang, Tianfang Li, Kevin Fallon, Dwight. E Heron, vol. 14, no. 6, pp. 195-204, 2013. 2. https://github.com/flavio-au/4d-vmat. 3. Modus Medical Devices Inc. http://modusqa.com/radiotherapy/ phantoms/respiratory-motion. 4. TROG 13.01 / ALTG 13.001 ‘‘Stereotactic Ablative Fractionated Radiotherapy versus Radiosurgery for Oligometastatic Neoplasia to the Lung: A Randomised Phase II Trial (SAFRON II): Radiotherapy Planning, Delivery and Quality Assurance’’. TROG, 2014. http://www.trog.com.au/SiteFiles/trogcomau/RT_Planning Delivery_and_QA_Guidelines_SAFRON_II_v2_18_Nov_2015.pdf.

O043 Assessment of surface dose from a TrueBeam linear accelerator in the context of breast radiotherapy P. Lonski1, R. D. Franich2, T. Kron1 1

Peter MacCallum Cancer Centre, Melbourne, Australia ([email protected] [Presenting author]). 2RMIT University, Melbourne, Australia Introduction Flattening filter free (FFF) beams are fast becoming popular and their use has some potential benefits in breast radiotherapy. The high dose rates associated with FFF allow for more comfortable treatment delivery for patients undergoing deep inspiration breath hold owing to the shorter beam-on time and the peaked beam shape may be exploited to improve dose homogeneity within the PTV. The consequences for skin dose however are not well understood and this work aims to quantify the surface dose in FFF beams. Method A thin window parallel plate chamber with a nominal effective measurement depth of 70 microns was used to measure surface dose in a solid water phantom irradiated using a TrueBeam linear accelerator. Measurements were conducted at various points in a 10 9 20 cm2 half beam blocked field (Fig. 1a) using 6 MV, 6FFF, 10 MV, 10 FFF and 18 MV. Surface doses beneath the shielding jaw were also measured at each energy for comparison.

1a

X2 7cm off axis 5cm off axis Y2

field centre

2cm 5cm OOF OOF

iso Y1

X1

Australas Phys Eng Sci Med

35 surface dose (%)

30

centre of field 5 cm off axis 7 cm off axis beneath jaw, 2 cm OOF beneath jaw, 5 cm OOF

1b

25 20

observed range of LDs. The increased risk compared to the original plan was determined using the linear relationship found by Darby et al. [2]. The potential MHD was also calculated assuming the treatment delivery was done with the highest LD found for the whole course.

15 10 5 0 6 MV

6 FFF

10 MV 10 FFF 18 MV beam

Results Compared to the flat beam, 6 FFF was associated with a lower surface dose out-of-field (OOF) but a higher surface dose in the field centre for a half beam blocked 10 9 20 cm2 field. 10 FFF however gave a lower surface dose in the field centre as well as out-of-field. The surface dose at the field centre, off-axis, and beneath the shielding jaw for all energies are summarised in Fig. 1b. Conclusion Surface dose measurements between flattened and FFF beams have been compared for a typical breast treatment field, both in-field and out-of-field. Surface dose from FFF has been shown to differ from flat beams of the same nominal energy. Skin reactions in breast radiotherapy can be of concern and surface dose should be considered during treatment planning.

Results The figure shows the range of LDs for all fractions of an average patient (Patient 4) as well as the ELD (red line). While for all patients the increase in risk for the delivered MHDs remained well below 2% the maximum risk increase was found to be 5.5% (Patient 10, see table). Patient

0044 How does the RPM control of DIBH treatments influence the risk of major coronary events? A MV cine imaging study 1,2

1

M. Doebrich , J. Downie , J. Lehmann 1

1,2,3

Calvary Mater Newcastle, Newcastle, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2The University of Newcastle, Newcastle, Australia. 3 The University of Sydney, Sydney, Australia Introduction Deep inspiration breath hold (DIBH) techniques aim to reduce the radiation dose to the heart during left breast radiotherapy treatments. A recent retrospective study has shown the possibility of using MV cine images to quantify the quality of the breath hold during DIBH treatments by measuring the distance between the internal alignment of the field edge and the breast-lung interface (lung depth, LD, [1]). As the risk of major coronary events increases linearly with the mean heart dose [2] the aim of the present study was to quantify the change in risk due to deviations of the LD from the distance measured during treatment planning (expected LD, ELD). Method MV cine images were obtained for all fractions of 10 left sided breast tumour patients treated with DIBH using the Varian RPM system. The LD was determined in each MV cine image ([15000 images total). The mean dose to the heart (MHD) that was delivered was estimated by creating different plans reflecting the

Planned MHD (cGy)

Delivered MHD (cGy)

Change in risk-delivered MHD compared to plan (%)

Potential MHD (cGy)

Change in risk-potential MHD compared to plan (%)

Patient 1

71.4

64.9

-0.5

81.7

Patient 2

74.4

86.2

0.9

114.3

0.8 3.0

Patient 3

89.3

101.0

0.9

130.0

3.0

Patient 4

57.7

65.8

0.6

80.1

1.7

Patient 5

109.4

114.4

0.4

158.2

3.6

Patient 6

91.7

91.1

0.0

122.8

2.3

Patient 7

80.1

77.8

-0.2

95.8

1.2

Patient 8

74.0

72.5

-0.1

98.3

1.8

Patient 9

77.5

81.1

0.3

103.0

1.9

Patient 10

72.6

90.9

1.4

146.7

5.5

Conclusion In this study we presented a method to estimate the change in risk of major coronary events during RPM-controlled DIBH treatments. This can be used to adapt the acceptable range of breath hold depth prior to treatment. References 1. Doebrich M, Downie J, Stanton C, Dempsey C, Lehmann J. Feasibility of MV cine imaging for monitoring of DIBH treatments. EPSM 2015, Wellington, New Zealand. 2. Darby SC, Ewertz M, McGale P, et al. Risk of ischemic heart disease in women after radiotherapy for breast cancer. N Engl J Med 2013; 368:987.

123

Australas Phys Eng Sci Med

O045 Inter- and intrafraction motion during deep inspiration breath-hold associated with Elekta Active Breathing Coordinator (ABCTM) in left-sided breast radiotherapy. P. O’Connor, N. Beaton Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]) Introduction Reducing heart dose in left-sided breast cancer radiotherapy is key to avoiding long term cardiac complications. One strategy for reducing heart dose is deep inspiration breathhold (DIBH), which in effect moves the heart out of the field but retains the full target coverage. One commercial system for respiratory gating is the Elekta Active Breathing Coordinator (ABC). As part of clinical introduction of this equipment, the reproducibility and stability of breath-hold and any inter- and intrafraction motion associated with using ABC with left-sided breast radiotherapy patients was investigated. Method The Elekta iView GT electronic portal imaging device (EPID) was used to take portal images of left breast treatment fields 1-3 times per week. Typically, an image was taken during the initial breath-hold, and another two images were taken during beam delivery whilst under a second breath-hold. The first two images were compared to assess the reproducibility of the system, whilst comparison between the second and third images gave an indication of the stability of the breath-hold. The image processing software ImageJ was used to measure the position of the chest wall on each image. Chest wall position was compared between images from a single treatment fraction and across treatment fractions to measure intra- and interfraction motion respectively. Results There was minimal movement of the chest wall between breath-holds administered by ABC in different treatment fractions; maximum motion was 1.4 mm, with 0.8 mm measured on average. Mean intrafraction motion was found to be 0.56 mm. Sub millimetre intrafraction motion was detected when analysing multiple images of the same breath-hold (i.e. DIBH was stable using ABC) Conclusion The results collected give confidence in the reproducibility of the DIBH technique using the Elekta ABC equipment. Results also provide assurance that there is minimal intrafraction motion associated with the equipment. Acknowledgements Thanks to Harish Sharma, Elizabeth Brown, Patricia Browne and Cathy Hargrave for their work on DIBH at the PAH.

O046 An automated EPID-based system for VMAT linac QA and commissioning: preliminary results and clinical implementation

([email protected]). 2School of Maths and Physical Sciences, University of Newcastle, Newcastle, Australia. ([email protected]). 3School of Maths and Physical Sciences, University of Newcastle, Newcastle, Australia. ([email protected]). 4Radiation Physics Laboratory, University of Sydney, Sydney, Australia. ([email protected]) Introduction Commissioning and quality assurance (QA) of linear accelerators (linacs) for Volumetric Modulated Arc Therapy (VMAT) is both technically challenging and resource intensive. The main difficulty is associated with gantry-resolved measurement and assessment of the dynamic components of the delivery; i.e. dose rate, gantry speed, and MLC-defined beam shape. In this work, we report on preliminary results and clinical implementation of a system for streamlining automatic gantry-resolved linac commissioning and QA for VMAT, using an electronic portal imaging device (EPID) and frame grabber assembly. Method The QA system relies solely on the analysis of EPID image frames acquired without the presence of a phantom at 8.41 frames per second using a dual-channel frame grabber and ancillary acquisition computer. A commissioning set of nine arc-dynamic QA plans were designed to assess the performance of individual linac components including MLC positional accuracy, MLC speed constancy, MLC acceleration constancy, MLC-gantry synchronisation, beam profile constancy, dose rate constancy, gantry speed constancy, dose-gantry angle synchronisation and mechanical sag. A subset of these 9 plans was used for ongoing linac QA. The components listed above were automatically extracted from image frames using an in-house designed software and user interface. The measured parameters were compared to DICOM plan files and, where appropriate, machine log files. Results The Figs below show a sample of the preliminary results attained at Central Coast Cancer Centre using a Varian iX linear accelerator. Data is also being acquired at two additional centres for (a) determination of realistic thresholds for error detection and (b) long term comparison between EPID-based measurements and machine log files (Fig.s 1, 2).

Fig. 1 Sample of preliminary results for dynamic dose rate and speed QA

B. J. Zwan1,2, M. Barnes1,2, J. Hindmarsh1, E. Seymour1, C. G. Lee1, D. J. O’Connor3, P. Keall4, P. B. Greer1,2 1

Department of Radiation Oncology, Central Coast Cancer Centre, Gosford, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]),

123

Fig. 2 Sample of preliminary results for MLC speed and MLC positional accuracy

Australas Phys Eng Sci Med Conclusion The proposed system has been designed, implemented and tested for automatic linear accelerator QA and commissioning for VMAT radiotherapy. This system allows direct measurement and assessment of each dynamic component of VMAT as a function of gantry angle.

O047 Development of a next-generation EPID for radiotherapy imaging and dosimetry: A review of progress to date S. J. Blake1,2, Z. Cheng1,2, S. Atakaramians3, M. Lu4, S. Meikle5, Z. Kuncic1, P. Vial1,2,6 1

Institute of Medical Physics, School of Physics, University of Sydney, Sydney, Australia. ([email protected]). 2 Ingham Institute of Applied Medical Research, Sydney, Australia. ([email protected] [Presenting author]), ([email protected]). 3Institute of Photonics and Optical Science, School of Physics, University of Sydney, Australia. ([email protected]). 4PerkinElmer Medical Imaging, United States of America. ([email protected]). 5 Faculty of Health Sciences & Brain and Mind Centre, University of Sydney, Australia. ([email protected]). 6Liverpool and Macarthur Cancer Therapy Centres, Sydney, Australia. ([email protected]) Introduction New methods of treatment verification that are in keeping with advances in radiotherapy treatment technology are desirable. Standard amorphous silicon electronic portal imaging devices (a-Si EPIDs) are widely used for patient imaging and have also been applied for dose verification by a few groups. However, the non water-equivalent dose response of standard EPIDs complicates their use for dosimetry. Our group is developing alternative EPID technologies that are optimised for simultaneous imaging and dosimetry applications. Method Prototype technologies currently being developed include: a standard EPID modified into a direct x-ray detection configuration [1]; a dual detector combining a standard EPID and ionisation chamber array [2]; an indirect-detection EPID comprising an array of plastic scintillator fibres (PSF) on the photodetectors of an x-ray detector (PerkinElmer, Santa Clara CA, USA) [3]. Each prototype’s performance for dosimetry and imaging of 6 MV photon beams was evaluated relative to a standard EPID and reference dosimeters. Monte Carlo (MC) simulations of detector performance using validated models are being used to inform the development of nextgeneration systems [4]. Results All prototype technologies exhibited near water-equivalent non-transit and transit dose response. Open and intensity-modulated dose measurements made using reference dosimeters consistently demonstrated greater agreement to prototype EPID dose measurements than standard EPID dose measurements. Although the direct detection EPID exhibited a reduced image quality relative to the standard EPID, the dual detector and PSF-EPID configurations improved upon this deficiency. MC simulations suggest that further gains in image quality may be realised by optimising the length and optical cladding properties of the PSF-EPID prototype without affecting dose response. Conclusion The feasibility of developing a next-generation EPID for simultaneous imaging and dosimetry has been demonstrated. Optimisation and clinical implementation studies suggest that such

technologies may be better suited for dosimetry applications than standard EPIDs while maintaining comparable image quality. References 1.

Vial P, Gustafsson H, Oliver L et al (2009) Direct-detection EPID dosimetry: investigation of a potential clinical configuration for IMRT verification. Phys Med Biol 54:7151-7169 2. Deshpande S, McNamara A, Holloway L et al (2015) Feasibility study of a dual detector configuration concept for simultaneous megavoltage imaging and dose verification in radiotherapy. Med Phys 42(4):1753-1764 3. Blake S, McNamara A, Deshpande S et al (2013) Characterisation of a novel EPID designed for simultaneous imaging and dose verification in radiotherapy. Med Phys 40(9):091902 4. Blake S, Cheng Z, Vial P et al (2016) In silico optimisation of the imaging performance of a novel water-equivalent EPID. In preparation for submission to Phys Med. Acknowledgements The authors acknowledge funding from the Australian Research Council (ARC) Linkage Project grant LP150101212, Perkin-Elmer Pty Ltd, Cancer Institute NSW (Research Equipment Grant 10/REG/1-20) and Cancer Council NSW (Grant ID RG 11-06). ZC also thanks The University of Sydney and the Australian Government for scholarship support in the form of an Australian Postgraduate Award. SA acknowledges a support of ARC funding DE140100614.

O048 Low dose paediatric Cone Beam Computed Tomography (CBCT) protocol development Julie-Anne Miller Radiation Oncology Mater Centre (ROMC), South Brisbane QLD, Australia. ([email protected]) Introduction The aim of this project is to develop reduced dose Cone Beam Computed Tomography (CBCT) protocols for small paediatric patients. In reducing patient dose, the protocols must maintain suitable image quality to setup the patient for treatment accurately. ROMC has four Varian Clinac iX linear accelerators equipped with On-Board Imager (OBI) kV imaging system, with OBI software v1.6. Method The two protocols assessed were Low Dose Head (LDH) and Low Dose Thorax (LDT) with a pulse width per projection of 10 ms and 20 ms1. The effect of the reconstruction kernel (smooth, sharp and standard) was evaluated. The CIRS 062 phantom, with and without rice as an expansion material, was used to assess the Contrast to Noise Ratio (CNR) and qualitative visibility of the inserts against the background. The SPoRT Pediatric Anthropomorphic Training Phantom (CIRS CI1-715 Series)2 was used to assess the suitability of the images acquired for bony image matching. SPoRT represents a typical 5-year old with materials that mimic attenuation properties of human tissue for diagnostic energies. Results The CNR for selected inserts in the CIRS (with rice expansion), for the LDT protocol shown in Graph 1 and LDR protocol in Graph 2. The smooth kernel had the best CNR, however fine detail in the scan was blurred out. All inserts were discernible against the background. Image quality for bony matching was subjectively assessed by a Radiation Therapist. All the scans were found to have sufficient image quality to allow bony matching. The limitation to this study is that SPoRT is not part of a range of different sized anatomical phantoms.

123

Australas Phys Eng Sci Med

Fig. 1 Weighted average dose values for Pelvis and Thorax CBCT scanning modes on the OBI and XI systems

Conclusion The reduced dose protocols can be adopted for patients with a maximum separation in the anterior-posterior or left-right of 20 cm, only where bony landmarks are used for image matching. References 1. Varian Medical Systems, Quick tip CBCT Acquisition Modes, 2014 2. Christine Rodgerson, ‘‘ALARA and Pediatric Imaging in Radiation Therapy’’, 71st Annual General Conference of the Canadian Association of Medical Radiation Technologists, May 2012 Acknowledgements Thanks to John Baines and Tim Markwell for all their assistance.

O049 Evaluation of dose from kV cone-beam computed to during radiotherapy: A comparison of methodologies J. Buckley1, D. Wilkinson2, A. Malaroda1, P. Metcalfe1 1

Centre of Medical and Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Illawarra Cancer Care Centre, Wollongong Hospital, Australia. ([email protected])

123

Introduction In this study, four different protocols are evaluated for measuring CBCT dose on two different linear accelerator models. The protocols compared are: the Computed Tomography Dose Index (CTDI), the Cone-Beam Dose Index (CBDI) [1], the International Atomic Energy Agency Report 5 (IAEA) [2] and the American Association of Physicists in Medicine Task Group 111 (TG111) report [3]. Method The four methodologies were evaluated for Pelvis and Thorax kV CBCT scanning protocols on the Varian 2100iX OBI and Varian Truebeam XI imaging systems. The CTDI, CBDI and IAEA protocols were evaluated using a 100 mm pencil ionisation chamber placed inside a CTDI phantom with a diameter of 32 cm and length of 16 cm. The TG111 was measured with a 0.6 cc Farmer ionisation chamber and required a custom built PMMA phantom with the same radial dimensions as the CTDI phantom, but of length 45 cm to achieve scatter equilibrium at the centre of the phantom. Results The weighted average doses for the OBI and XI systems are presented in Fig. 1. The TG111 measured the highest dose due to a longer, full scatter phantom. CBDI and IAEA gave the second and third highest dose due to insufficient phantom scatter compared with TG111. CTDI gave the lowest dose due to insufficient phantom scatter and an underestimation of dose due to the ionisation chamber not having the required length to measure the full dose profile. Conclusion The CBDI, IAEA and TG111 address the underestimation of CBCT dose by the current CTDI paradigm. The TG111 method gives a more realistic estimate of dose to the patient compared with the other protocols. The CBDI and IAEA protocols provide a method for characterising CBCT dose without having to build a customised phantom and are useful for comparing relative doses between machines and various anatomical CBCT protocols. Acknowledgements The author wishes to thank Martin Carolan and the Illawarra Cancer Care Centre for the use of their equipment and Simon Downes from Prince of Wales Hospital for lending of the CTDI phantom. References 1. A. Amer et al., Br. J. Radiol. 80, 476–482 (2007). 2. IAEA Human Health Reports 5, I. A. E. A, (2011). 3. AAPM Task Group Report 5, A.A. Phys. Med, (2010)

Australas Phys Eng Sci Med

O050 Optimising image quality of 4D-CBCT on a Varian TrueBeam Linear Accelerator A. Briggs, N. Hardcastle Northern Sydney Cancer Centre, Royal North Shore Hospital, St Leonards, NSW. ([email protected] [Presenting author]), ([email protected]) Introduction Clinical implementation of 4D-CBCT on Varian TrueBeam linear accelerators requires image quality optimisation. 4D-CBCT is used to validate the setup position for moving targets and verify tumour motion and motion management strategies. 4DCBCT suffers from poor soft-tissue contrast due to lack of projections throughout the full breathing cycle at each projection angle. Gantry speed and frame rate can be adjusted and these variables were investigated to individualise image acquisition based on tumour motion characteristics. Method Image quality was assessed using the uniformity module of the Catphan 503 phantom, comparing corresponding slices on a 4DCBCT to a reference 3D-CBCT. The interplay of image quality with number of projections and gantry speed was investigated qualitatively, and quantitative comparisons were performed using Mean Absolute Pixel Difference (MAPD) and Structural Dissimilarity (DSSIM) [1]. Imaging dose was measured using methods outlined in AAPM TG-111 [2]. Results Qualitatively, image quality can be improved by reducing the gantry speed for a given frame rate (Fig. 1). This increases the projections at a given angle. In practice, we have limited the acquisition time to 180 s. Quantitative comparisons using DSSIM and MAPD suggest the diminishing returns with increasing the frame rate, facilitating a pathway of dose reduction for a cohort of slower breathing patients (Fig. 2). Imaging dose has a linear relationship with total frames per scan. The 4D-CBCT dose ranges from 2–14 cGy (840 c.f. 5400 frames). Conclusion The best compromise in image quality and dose was achieved with a 180 second scan (360 rotation) at 11 fps for slow breathers and 15 fps for fast breathers. Selection of frame rate should be guided by tumour motion characteristics, where a higher frame rate is typically required for faster moving targets.

Fig. 1 Improving image quality by increasing total scan time for a constant frame rate

Fig. 2 DSSIM (higher value is worse correlation with reference image) for a 180 s scan with three frame rates and two breathing periods References 1. Wang Z, Bovik AC, Sheikh HR, Simoncelli EP (2004) Image quality assessment: from error visibility to structural similarity. Image Processing, IEEE Transactions on 13(4):600–612 2. American Association of Physicists in Medicine (2012) Dose calculation for photon-emitting brachytherapy sources with average energy higher than 50 keV: Full report of the AAPM and ESTRO. College Park, MD

KS08 Implementing dose accumulation in standard practice: Is it clinically feasible and evidence driven? Kristy K. Brock University of Michigan. ([email protected] [Presenting author]) The use of deformable registration for dose accumulation and potentially subsequent adaptive treatment has been investigated in the retrospective research setting for many years. The recent integration of this technology into commercial treatment planning systems has begun to make the transition to clinical use possible. However, as with the introduction of any new technology, initial commissioning, an overall quality assurance program, and patient specific QA procedures must be established. In addition, the incorporation of new technology may require costs due to training, equipment, and additional personnel. Therefore, the dissemination of evidence to support the proposed clinical benefits of the integration of new technology is required. These topics will be addressed. Several methods have been reported to validate image registration, including the use of mathematical and physical phantoms, as well as metrics that can be performed on clinical images, such as target registration error, mean distance to agreement, and Dice similarity coefficient. The combined use of these techniques play a vital role in the commissioning of image registration algorithms. Once commissioned, an efficient and effective patient specific QA technique must be established. These steps are critical to the safe use of image registration, as limitations exist. The limitations and uncertainties in the clinical setting will be illustrated, as well as proposed QA and QC procedures.

123

Australas Phys Eng Sci Med Finally, as a field, we must strive to obtain evidence that new technology improves our ability to safely and effectively treat patients with radiation therapy. Early studies will be presented that provide preliminary evidence of the impact of including deformable registration and dose accumulation in clinical practice. Acknowledgements The author would like to acknowledge the AAPM Task Group 132 for their recommendations on commissioning and QA for deformable registration.

O051 Recommendations for implementation of deformable image registration based on TG100 J. Yuen, L. Foessel, A. Ralston St George Hospital Cancer Care Centre, Sydney Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Deformable registration can address challenges in image registration for tumor localization [1] and its potential has been demonstrated in multi-institution accuracy studies [2]. However, its use for diagnosis and treatment planning without considerations of associated risks could introduce serious errors such as under-dosing tumours or over-dosing critical organs. The aim was to create guidelines for safe implementation and use of deformable image registration software. Method Our risk/benefit and process-based framework, based on risks assessments and failure mode effects analysis (FMEA) recommended by TG-100 [3, 4], manages all processes relevant to image registration, including when to use rigid and/or deformable registrations (see Table 1). Specific quality control strategies covering all stages in the patient path from presentation to treatment were identified. Iterative risk evaluation and process adaptation with quality controls were performed until residual risks were acceptable. Results The process map for our department (Fig. 1) identified nine quality improvement projects:

Table 1 An image registration scheme for selection of registration method based on image registration risks, risk factors, and relative risk assessment of rigid and/or deformable registration. Red colour indicates potential weaknesses Image Regis- Option 1: Rigid Option 2: Option 3: tration Risks registration Deformable reg- Resample data only istration only with option 1&2 Error due to user or Processes potentially processes optimised

Processes

Processes

potentially suboptimal

potentially suboptimal

Setup and immobilisation error

Uncorrected errors

Correction reduces this error

Correction reduces this error

Error due to imager unit differences

Uncorrected errors

Correction reduces this error

Correction reduces this error

Error due to organ size change

Uncorrected errors, size change visible

Correction reduces error, Correction reduces error, size change size change not visible visible

Errors due to patient preparation variations

Uncorrected errors, changes visible

Correction reduces error, Correction reduces changes not visible error, size change visible

Errors due to intrafraction effects

Uncorrected errors, changes visible

Correction reduces error, Correction reduces changes not visible error, size change visible

Uncorrected algorithm errors

Simple translation and Limited validation rotational results, error data algorithm requires specialised QA (e.g. vector maps)

Data storage and workload requirements

Simple workload, data More complex workload, Most complex increased data increased workload, data increased more

Conceptual approach and accounting for random and non-random errors

Manual visual estimation of PTV OAR contours, depends on site registration strategy

Comparison of deformable and rigid resampled data

Corrected registration for Uncorrected and corrected estimation of PTV registration for OAR contours, estimation of PTV depends on site OAR contours, registration strategy depends on site registration strategy

1. Risk-based framework: each major site-function undergoes risk management before clinical release 2. DICOM connectivity: minimise unavailability of datasets and transfer time 3. Preventable scan differences: minimise differences such as setup-immobilisation 4. Image registration aims: including aims in prescription 5. Image registration scheme: adoption of a scheme (Table 1) to achieve prescribed aims 6. Image registration QA: registration errors with action thresholds 7. Auto-contouring: standardised naming and increased efficiency 8. Contour and margins: registration errors accounted for systematically 9. Image registration management: technical and non-technical issues logged and used to improve quality Conclusion A risk-based framework specific for image registration has been developed to systematically consider human and process factors, including variations in case studies. This framework can be used to clinically implement functions in a staged manner based on acceptable risks vs. benefits. An optimised dataset registration scheme individualised to specific image registration aims was developed.

123

Fig. 1 Processes identified that relate to image registration with quality control measures within a risk-based framework that will be unique to each department’s hardware/software/human condition

Australas Phys Eng Sci Med Acknowledgements The authors thank Amy Walker from Liverpool Hospital for providing us with CT and MRI phantom datasets. References 1. Sharpe, M., and K. K. Brock (2008), Quality assurance of serial 3D image registration, fusion, and segmentation: Int J Radiat Oncol Biol Phys, v. 71, p. S33-7. 2. Brock, K.K. and Deformable Registration Accuracy Consortium (2010). Results of a multi-institution deformable registration accuracy study (MIDRAS). International Journal of Radiation Oncology* Biology* Physics, 76(2), pp.583–596. 3. Huq, M. S., B. A. Fraass, P. B. Dunscombe, G. J. J. P., G. S. Ibbott, A. J. Mundt, S. Mutic, J. R. Palta, F. Rath, B. R. Thomadsen, J. F. Williamson, and E. D. Yorke (2016), The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management: Medical Physics. 1. ISO (2009), ISO 31000:2009 Risk management–Principles and guidelines: International Organization for Standardization, Geneva, Switzerland.

O052 Feasibility study for deformable image registration in cervical cancer HDR brachytherapy E. Flower1,2, V. Do1,3, J. Sykes4, C. Dempsey5,6,7, L. Holloway8,9,10,11, K. Summerhayes1, D. Thwaites11 1

Crown Princess Mary Cancer Centre, Westmead, NSW, Australia. ([email protected]), ([email protected]), ([email protected]). 2Institute of Medical Physics, School of Physics, University of Sydney, NSW, Australia. 3 Sydney Medical School Nepean, University of Sydney, NSW, Australia. 4Blacktown Cancer & Haematology Centre, Blacktown, NSW, Australia. ([email protected] (Presenting author)). 5Department of Radiation Oncology, University of Washington, WA, USA. ([email protected]). 6Department of Radiation Oncology, Calvary Mater Newcastle Hospital, NSW, Australia. 7School of Health Sciences, University of Newcastle, NSW, Australia. 8Ingham Institute and Liverpool and Macarthur Cancer Therapy Centres, NSW, Australia. ([email protected]). 9Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. 10South Western Sydney Clinical School, University of New South Wales, NSW, Australia. 11Institute of Medical Physics, School of Physics, University of Sydney, NSW, Australia. 11Institute of Medical Physics, School of Physics, University of Sydney, NSW, Australia. ([email protected]) Introduction The purpose of this study was (1) to determine the reproducibility of CT-based DIR for cervical cancer brachytherapy, (2) to compare the anatomic location of hot spots for individual fractions and accumulated across multiple fractions for CT based planning and (3) to determine the feasibility of using DIR for dose accumulation for MRI-based planning. Method DIR reproducibility was assessed for different implementation based on adding contour biases added to the DIR algorithm, assessed using dose volume histogram parameters. VolD2cc and VolD0.1cc were created from the overlap of the D2 cc and D0.1 cc isodoses and the bladder or rectum and the overlap was calculated using the Dice similarity coefficient (DSC).

Results DIR summed D2 cc and D0.1 cc decreased 2.9 and 4.2% for bladder and 5.08 and 2.8% for rectum compared with no DIR. The reproducibility increased with the addition of bladder contour biases to the DIR algorithm for the bladder. The average DSC was 0.78 and 0.61 for the bladder Vol2 cc and Vol0.1 cc, and 0.83 and 0.62 for rectal Vol2 cc and Vol0.1 cc. For the MRI DIR the average HR-CTV D90 increased 0.6% (range -10.5 to 11.5%) and the average D98 decreased 3.7% (range -16.7 to 17.6%). Conclusion Dose decreases were observed for summed DVH parameters using DIR, except for HR-CTV D98. The anatomical positions of the VolD2cc and VolD0.1cc were not stable. More studies with a larger cohort are required to determine any potential clinical effects of DIR accumulated dose.

O053 Commissioning of diagnostic radiology imaging equipment at the new Royal Adelaide Hospital C. Boyd, D. McRobbie, Y. Matyagin, S. Midgley Medical Physics and Radiation Safety, SA Medical Imaging, Adelaide, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]) Introduction The new Royal Adelaide Hospital is a $2bn, 130,000 m2, 800 bed facility which represents a significant upgrade for Adelaide and wider South Australia. The facility is a private public partnership, with the builder responsible for infrastructure, including shielding design and verification, whilst health is responsible for the imaging equipment. The hospital layout uses a ‘distributed imaging’ design, meaning radiology is dispersed throughout multiple departments. As an essential part of preparation for clinical operation, all diagnostic radiology imaging equipment requires both acceptance testing and regulatory compliance testing. The new equipment requiring testing is summarised in Table 1 below: In addition to the apparatus listed above, a number of existing equipment will be transferred and require significant testing. Method Given the limited number of physicists available to conduct testing and the timeframe for completion, a significant number of machines were required to be tested during the final phase of construction. To allow this, a full day Work Health and Safety construction training course was required to be completed, along with

Table 1 List of new equipment for the Royal Adelaide Hospital Machine type

Quantity

1.5T MRI

2

3.0T MRI

1

OPG X-Ray

1

Fluoroscopy Suites (single & bi plane)

7

Mobile Fluoroscopy

6

Radiography Suites

12

Mobile Radiography

6

Computed Tomography

6

SPECT-CT

1

PET-CT

1

123

Australas Phys Eng Sci Med a half day site induction. Additionally, a detailed electronic testing protocol was developed to assist in collecting relevant information in a time efficient manner. Results The process, challenges, outcomes and lessons learned will be individually examined and discussed. Conclusion Whilst requiring additional care and planning, both prior and during testing, the largescale commissioning of radiology equipment during building construction is feasible. It is recommended however that whenever possible, acceptance testing is conducted immediately prior to clinical use.

O054 A Radiation Safety Audit and Optimisation of Dental Radiography within a Large Centre for Oral Health A. J. Pascoe, A. Jones, R. K. Grewal Medical Physics Department, Westmead Hospital, Sydney, Australia. ([email protected] [Presenting Author]), ([email protected]), ([email protected]) Introduction The Westmead Centre for Oral Health (WCOH) is a large centre with more than 50 intra-oral, 3 panoramic and 2 conebeam computed tomography (CBCT) units. This project was initiated by the Medical Physics Department at Western Sydney Local Health District (WSLHD) with an aim to improve current radiation safety practices across the WCOH. The project consists of three components: a radiation safety audit, dental x-ray dose standardisation; and subsequent optimisation of image quality and patient dose. Method A questionnaire was developed to identify key issues related to current work and radiation safety practices across WCOH. The data from 64 staff members was used identify the clinical exposure factors used for 30 intra-oral units. The typical entrance surface air kerma (ESAK) was measured for a typical adult mandibular molar (AMM) exposure. These values were standardised to 1.15 mGy which represents the UK National Diagnostic Reference Level for a typical AMM x-ray using a digital detector [1]. An image quality audit [2] is currently being undertaken with the support of 5 senior dental officers. The aim of this audit is to establish the optimal ESAK required to produce an image with sufficient clinical image quality. Results Key factors identified from the survey included the role of dental assistants in taking x-rays, inconsistent lead apron usage, increased staff assistance for special care patient imaging, and the wide variation in exposure settings selected for an AMM intra-oral X-ray. Dose standardisation has resulted in a significant reduction in variance between the initial (median = 0.815 mGy, r = 0.438 mGy) and standardised ESAK datasets (median = 0.812 mGy, r = 0.170 mGy). Conclusion The early results of this work have highlighted the need for dose optimisation programs and radiation safety education within dental centres. This ongoing project is anticipated to include panoramic and CBCT units in the near future, and will also be extended to other oral health sites across the WSLHD.

dental practice. Radiation Protection 136. European Commission, Belgium. ISBN 92-894-5958-1.

O055 To validate a dedicated Monte Carlo simulation system for performance evaluation of anti-scatter grids Abel Zhou, Graeme L. White, Rob Davidson Faculty of Health, University of Canberra, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Many investigations of grid performance are undertaken using Monte Carlo simulations. Monte Carlo simulations provide flexibility in designs of phantom and irradiating systems, which are either very expensive or demanding to setup experimentally [1-2]. Monte Carlo simulations also have the advantages to replicate clinical situations and allow validation of technologies before implementation [3]. A commonly used Monte Carlo simulation is PENELOPE (Penetration and energy loss of positrons and electrons). PENELOPE uses FORTRAN computer language and it has a set of subroutines to perform simulations of electron-photon coupled transport for a wide energy range in arbitrary materials and geometries [4]. For improved efficiency of simulations of photon transport in biological tissues for the purpose of anti-scatter grid evaluation, the simulation code from PENELOPE was previously modified. The modified code kept the linear attenuation coefficients and physical interaction processes, i.e. Rayleigh scattering, Compton scattering, and photoelectric absorption, however the subroutines to perform simulations were re-written in Matlab (release 2015b, Simulink, Massachusetts, United States). The purpose of this work was to validate the improved Monte Carlo simulation system against experimental results. Method A 12:1 ratio focused grid (Philips Medical Systems, Germany) and various thicknesses of PMMA phantoms (IBA Dosimetry GmbH, Germany) were obtained. The details of this grid and various thicknesses of PMMA were used in the revised PENELOPE simulation code. The simulated X-ray beams used were according to IEC 60627 International Standards [5]. The output calculations of the simulation were the transmissions of primary radiations (Tp), the transmissions of scatter radiations (Ts) and the transmissions of total radiations (Tt) at various PMMA thicknesses and for the 12:1 grid. In the experiment a general x-ray unit and tube (Philips Medical Systems, Germany) was used as the source of the X-rays. Consistency of the output, HVL and various dose measurements were recorded using a dosimeter (MagicMax Universal, IBA Dosimetry GmbH, Germany). Measurement of Tp, Ts and Tt were made using various x-ray beams according to IEC 60627 International Standards [5]. Results and Conclusions Measurement of Tp, Ts and Tt, under various x-ray beams and PMMA thicknesses, were obtained using simulation and experimental methods and then compared. The results show a high level of agreement between simulation and experiment. The revised PENELOPE simulation code using Matlab can be used to validate existing and/or new grid designs.

References 1. Hart D, Hillier MC, Shrimpton PC (2012) Doses to Patients from Radiographic and Fluoroscopic X-ray Imaging Procedures in the UK – 2010 Review. HPA-CRCE-034. Health Protection Agency, Oxfordshire. ISBN 978-0-85951-716-4. 2. European Commission (2004) European guidelines in radiation protection in dental radiology: The safe use of radiographs in

123

References 1. Chan, H.P., Y. Higashida, and K. Doi, Performance of antiscatter grids in diagnostic radiology: experimental measurements and Monte Carlo simulation studies. Medical Physics, 1985. 12(4): p. 449–54.

Australas Phys Eng Sci Med 2. Cunha, D.M., A. Tomal, and M.E. Poletti, Evaluation of scatterto-primary ratio, grid performance and normalized average glandular dose in mammography by Monte Carlo simulation including interference and energy broadening effects. Physics in Medicine & Biology, 2010. 55(15): p. 4335–59. 3. Del Lama, L.S., D.M. Cunha, and M.E. Poletti, Validation of a modified PENELOPE Monte Carlo code for applications in digital and dual-energy mammography. Radiation Physics and Chemistry, 2016. 4. Salvat, F., J. Fernndez-Varena, and J. Sampau, Penelope-2011, A code system for Monte Carlo simulation of electron and phonon transport, OECD. NEA Data Bank, Issy les Moulineaux, France, 2011. 5. International Electrotechnical Commission, Diagnostic X-ray imaging equipment : characteristics of general purpose and mammographic anti-scatter grids. 2013, IEC: Geneva.

O056 Development of a patient specific quality assurance technique for the Brainlab Multi-Metastases treatment planning system J. Poder1,2, J. Morales1,3, R. Hil1,4, M. Butson1 1

Department of Radiation Oncology, Chris O’Brien Lifehouse, Sydney, Australia. 2Centre of Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 3School of Chemistry, Physics and Mechanical Engineering, Queensland University of Technology, Australia. 4Institute of Medical Physics, School of Physics, University of Sydney, Australia Introduction Linear accelerator based stereotactic radiosurgery (SRS) is used to treat brain metastases in the Department of Radiation Oncology, Chris O’Brien Lifehouse. Traditionally, the SRS program was limited to treating multiple lesions with multiple isocentres, and patients with more than 3 brain metastases were treated using whole brain radiation therapy. The Brainlab Elements Multiple Metastases (BEMM) (v1.0.1.97, Brainlab AG, Feldkirchen, Germany) module was recently commissioned in Lifehouse. The development of a patient specific quality assurance technique for BEMM treatments is described. Method The lack of a dedicated physics quality assurance (QA) module in BEMM required the use of third party systems to develop a patient specific QA protocol. The Radcalc MU check program Lifeline Software) and the Varian Portal Dosimetry (Varian Medical Systems) were commissioned to achieve this purpose. BEMM arcs were planned on a slab phantom geometry with field sizes ranging from 6–30 mm, number of metastases ranging from 2–10 and off axis distances of 2–6 cm. The dose to each metastasis was individually checked in Radcalc using an independent dose calculation and fluence checks were analysed in Portal Dosimetry using gamma analysis. Subsets of these plans were then calculated on an anthropomorphic phantom and the process repeated. Results Mean dose differences between BEMM and Radcalc for the slab geometries was 0.4% (max 13.3%, min -4.9%) where as for the anthropomorphic cases, the mean difference was 2.8% (max 11.6%, min -0.6%). Fluence checks using Portal Dosimetry yielded an average pass rate of 98% for a gamma criteria of 3%/3 mm and 96.5% for 2%/2 mm. Conclusion Despite the limitations of the BEMM v1.0.1.97 for physics QA, a patient specific QA protocol has been developed using

both independent dose calculations and fluence measurements with an electronic portal imager. Acknowledgements The support of Brainlab staff is acknowledged.

O057 Testing the limits of stereotactic radiosurgery for multiple brain metastases S. Zolfaghari1, S. B. Crowe1,2, D. Papworth3, M. West3, T. Kairn1,3 1

Queensland University of Technology, Brisbane, Australia. ([email protected]), ([email protected] [Presenting author]). 2Royal Brisbane & Women’s Hospital, Brisbane, Australia. 3Genesis Cancer Care Queensland, Brisbane, Australia. ([email protected]), ([email protected]), ([email protected]) Introduction Brain metastases are the most common form of intracranial tumours, and a significant cause of morbidity and mortality in many cancer patients. The most widely recommended treatment technique is the use of stereotactic radiosurgery, for which there is a perceived limitation in applicability to patients with multiple lesions. The volume of brain receiving 12 Gy has been suggested as the most important factor in predicting the risk of neurological impairment [1]. This study evaluated the relationship between number and combined volume of metastatic lesions and both dosimetric plan quality and treatment deliverability. Method Five phantom metastatic lesion combinations were generated on a CT of a CIRS head phantom, containing 4, 6, 8, 10 and 12 lesions, with combined total lesion volumes of 4.4 to 8.0 cc. Static conformal arc treatments were planned for a Varian iX accelerator with attached BrainLab m3 micro-multileaf collimator system, with a prescription dose of 24 Gy to 90% isodose (taken from RTOG 0320). Treatment delivery was verified with Gafchromic EBT3 film measurements performed at a transverse plane intersecting the centre of the largest simulated lesion, and compared against planning system dose calculations using SNC MapCheck software. Results The volume of healthy brain receiving 12 Gy increased linearly with total treatment volume. For the RTOG prescription of 24 Gy, no plans resulted in a V12 less than 10 cc. Gamma pass rates

Fig. 1 CIRS head phantom with EBT3 film at isocentre of largest metastatic volume

123

Australas Phys Eng Sci Med (using 3% / 1.5 mm criteria) were lowest for the treatments of 10- and 12-metastases, due to increased disagreement in out-of-field regions. Area of film receiving 12 Gy was not a strong predictor of V12 (Fig. 1). Conclusion Treatment deliverability was not a limiting factor in the treatment of multiple metastases. Recommended dose constraints could not be met with the prescribed dose. References 1.

Lawrence, Y. R. et al. (2010). Radiation dose-volume effects in the brain. Int J Radiat Oncol Biol Phys 76(3): S20-S27.

O058 A retrospective study of PTV size reduction with gating for hypofractionated SBRT liver treatments

Results Figure 1 shows decreasing dose prescription with increasing PTV size; 7/20 patients had a reduction in prescribed dose due to normal tissue constraints. As indicated by the pseudo-PTV series, for volumes 10–30% of liver size there is potential for the BED = 100 Gy to be achieved reducing the PTV size. For larger volumes, although the target BED = 100 Gy value may not be reached, there is a significant reduction in surrounding liver volume included in the target. For an example case, a PTV 63% of the size of the liver saw a reduction in liver volume included in the PTV of 12% with gating. Conclusion For PTV sizes greater than 20% of the liver volume, reduction in PTV size through gating or MLC tracking, can potentially increase the BED for liver tumours, and decrease (prescriptionlimiting) dose received by the surrounding liver volume. The degree of reduction in PTV size is governed by the GTV size and degree of motion of the tumours. References

M. Gargett, V. Caillet, C. Haddad, J. Booth, N. Hardcastle

1. Radiation Oncology Therapy Group (RTOG) 1112 Protocol, available at https://www.rtog.org/ClinicalTrials/ProtocolTable/ StudyDetails.aspx?study=1112

Northern Sydney Cancer Centre, Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected])

O059 Predicting the PTV margin for liver SABR treatment

Introduction Due to the proximity of the liver to the diaphragm, anterio-inferior movement of liver tumours is accounted for in planning target volumes (PTVs) that include both setup uncertainties and tumour motion. Liver SBRT plans can struggle to reach target dose (50 Gy (BED = 100 Gy)[1] in our institution) without compromising OAR tolerances, hence have isotoxic prescriptions. We investigate the potential gains of respiratory-gating SBRT delivery to reduce PTV size and facilitate maintenance of high tumour doses and reduction of dose-limiting toxicities. Method A retrospective study of 20 liver SBRT patients has been performed. The PTV size inclusive of setup uncertainties and tumour motion was compared to a pseudo-gated-PTV. The pseudo-gatedPTV was obtained by applying a 5 mm expansion on the gross tumour volume (GTV) contoured on the exhale phase of the 4D-CT; that is, tumour motion is not included in the pseudo-gated-PTV.

120 PTV

100

pseudoPTV

BED (Gy)

80 60 40 20 0 0%

10%

20%

30%

40%

50%

60%

70%

PTV size (% of liver volume)

Fig. 1 The biologically effective dose prescribed to 20 patients as a function of PTV size (as a percentage of the liver volume), where the green segment represents volume sizes expected to receive a higher target BED with gating. The pseudo-PTV series represents the estimated reduction in PTV size with gating

123

J. Foo1, S. Yau1,2 1

Nepean Cancer Care Centre, Nepean Hospital, Australia. ([email protected] [Presenting author]). 2Crown Princess Mary Cancer Centre, Westmead Hospital, Australia. ([email protected])

Introduction Stereotactic ablative body radiotherapy (SABR) of the liver has many challenges mainly attributed to organ deformation, respiratory motion and poor cone beam CT (CBCT) contrast in the liver. This causes difficulty in choosing a PTV margin. A PTV margin that is too small may not adequately cover the target because of treatment setup errors thus leading to undesired replanning with larger margins. In this study, CBCT data was acquired during treatment for position verification and the adequacy of the current PTV margin was assessed. The setup data was also put in a margin recipe calculation to help predict a PTV margin for future patients. Method Five patients were treated using a breath-hold technique with the Elekta ABC system. Each patient was planned for five fractions. Initial PTV margins of 5 mm were set by the oncologist. For each fraction, the patient typically receives four CBCTs. The positional error during treatment was analysed. The van Herk1 margin recipe formula was applied to determine the appropriate PTV margin using the systematic and random error from the setup. Results The positional error was plotted against time from the first CBCT. An example is shown in Fig. 1 for one patient in the S-I direction. Fig. 1 shows a steady increase in intrafraction motion as the treatment time increased. It is clear that a 5 mm margin is inadequate; this resulted in replanning with the margin increased to 10 mm. The PTV margin was also validated against the margin recipe calculation. As liver SABR patients are less common, the predictive power of this margin recipe calculation can be improved as patient numbers increase. Conclusion Intrafraction motion of liver SABR patients has shown to exceed initial PTV margins even with breath-hold. CBCT data was used to predict the appropriate margin required for future patients to improve target delivery accuracy.

Australas Phys Eng Sci Med 0.80 0.60

PTV Margin Posion Error (cm)

0.40 0.20 0.00 0:00:00 -0.20

0:14:24

0:28:48

0:43:12

0:57:36

Conclusion An opportunity exists to improve education and the profile of medical physics in the region through AMPLE to organise and intensify training activities. Acknowledgements IAEA regional projects are a team effort with counterparts from 16 or more countries as well as staff from IAEA and consultants. I would like to thank them all. Those from Australia currently include Brendan Healy, Anne Perkins and Caroline Irle amongst others.

-0.40

PTV Margin -0.60 -0.80

Time elapsed from 1st CBCT (hh:mm:ss) Fxn 1

Fxn 2

Fxn 3

Fxn 4

Fxn 5

Fig. 1 Position error during treatment for each fraction for one liver SABR patient in the S-I direction

O061 Recent events in radiation therapy in Papua New Guinea S. J. Downes Nelune Comprehensive Cancer Centre, Prince of Wales Hospital, Randwick NSW. ([email protected])

References 1.

van Herk, M., Remeijer, P., Rasch, C., Lebesque, J.V. (2000) The probability of correct target dosage: dose-population histograms for deriving treatment margins in radiotherapy, Int. J. Radiat. Oncol. Biol. Phys., 47(4):1121-1135.

O060 Clinical training of medical physicists in Asia through IAEA regional projects I. D. McLean Medical Physics and Radiation Engineering (MPRE), Canberra Hospital, Australia. ([email protected] [Presenting author]) Introduction In 2003 it was identified by John Drew and the International Atomic Energy Agency (IAEA/ RCA) that the most effective way to strengthen clinical medical physics capacity in the Asian Pacific region was through the introduction of clinical medical physics training. The IAEA regional project RAS6038 developed and piloted methodology to achieve this in a number of locations enjoying a symbiotic relationship with the development of clinical training in Australian through the ACPSEM. The current project (RAS6077) is extending this concept through a focus on professional issues including cooperation with AFOMP and the uses of an e-learning platform called AMPLE (advanced medical physics learning environment). Method AMPLE has been developed as a vehicle to collect and deliver the resources needed in clinical training and also as a tool for communication and effective administration of national clinical training programs. The role of the project is to establish the platform, and assist in its implementation in pilot countries in the region. Additionally the increasingly important issues of accreditation of training institutions and certification of medical physicists in a regional context will be examined. Results Pilot testing of the electronic platform has begun and is continuing through 2017. A new project is in design phase for 2018 onwards to specifically investigate and address implementation in developing countries. This will need development of strategies for remote supervision. The potential to use AMPLE to define, sustain and record progress in short term attachments, such as IAEA fellowships, or volunteer placements, is seen as a direction that should be explored.

Introduction Radiotherapy in Papua New Guinea (PNG) has continued to treat patients at the PNG National Treatment Centre using the same radiotherapy equipment that was purchased in 2008 along with additional treatment units. These include a Cs-137 low dose rate brachytherapy afterloader (2012) and Pantak kilovoltage unit (2016). The Australian and PNG government agreed to co-finance the redevelopment of the ANGAU Memorial hospital site on which the PNG National Treatment Centre resides. The construction of the new ANGAU hospital development requires the current site of the treatment centre to move. The age of the Cobalt-60 teletherapy source is now approaching 7 years and the purchase of a new source is also needed. This purchase has been stalled as a result of a lack of any regulatory framework for radiation safety in PNG. Method With the aid of the International Atomic Energy Agency (IAEA), the PNG government has commenced the establishment of a national regulatory framework and regulatory body in radiation safety. Ensuring the continued delivery of radiation therapy to patients during the relocation of the national treatment centre will be a logistical and financial challenge for the hospital. Results A potential site for a new national cancer centre was found on the existing ANGAU hospital campus which shows significant promise as permanent location and would allow for significant future expansion of cancer services. Conclusion Due to a number of issues that will be discussed, the development of radiotherapy in PNG continues to be uncertain.

O062 Can the anisotropic analytical algorithm be used to plan total body irradiation treatments at 460 cm SSD? G. Fejes1, J. C. Crosbie2, V. Panettieri3, J. Droege3 1

RMIT University, Australia. ([email protected] [Presenting author]). 2School of Science, RMIT University, Australia. ([email protected]). 3Alfred Health Radiation Oncology, Alfred Hospital Commercial Rd, Melbourne, Australia. ([email protected]), ([email protected]) Introduction Total Body Irradiation (TBI) is used in conjunction with chemotherapy to prepare patients for white blood cell transplantation. Accurate planning is essential due to the high doses and

123

Australas Phys Eng Sci Med large volumes being treated. However, most treatment planning systems do not support the extreme geometries used in TBI. Here, the Eclipse Analytical Anisotropic Algorithm (AAA) is assessed for treatment planning TBI at 460 cm SSD. Method This planning study commenced by placing TBI fields (40x40 cm jaws, 460 cm SSD, 18MV) on a large water equivalent phantom including two lung equivalent blocks. Dose distributions generated by AAA and the current clinical algorithm, Pencil Beam Convolution (PBC), were compared. Next, fourteen patient plans were recalculated using AAA. Dose distributions for both algorithms were compared, and exit point doses from the plans were matched with in-vivo dosimetry measurements taken during treatment. Finally, BioSuite radiobiological software was used to assess risk of pneumonitis for both algorithms using an LKB model. Results Phantom measurements showed *8% discrepancies between PBC and AAA at 460 cm SSD on both beam profiles and percentage depth doses, particularly noticeable in heterogeneous regions. These appeared as circular isodose patterns, disappearing as the SSD was reduced to 100 cm. The exit point doses calculated by AAA were systematically lower than both PBC and measurement, statistically significant in the pelvis and abdomen regions only. PBC doses were generally higher than measurement. The difference in lung dose was deemed not clinically significant for risk of pneumonitis, with a change in risk of 0.01%. Conclusion The AAA algorithm has been shown to systematically calculate lower doses than the PBC algorithm and in-vivo measurements at 460 cm SSD. This was statistically significant for abdomen and pelvis measurements but still within clinical tolerances. If AAA is used for TBI planning, these discrepancies should be taken into account.

compared using the two-sided t-test and paired Wilcoxon signed rank test with a significance level of 0.05 with false discovery rate of less than 0.05. Results For IMRT, 1-arc VMAT and 2-arcs VMAT prostate plans, there was a significant reduction in the planning time with vectormodel-supported optimisation by 1.99 hours, 5.44 hours and 2.91 hours, respectively. Similarly, the number of iterations was significantly reduced with vector-model-supported optimisation (4 iterations for IMRT, 8 iterations for 1-arc VMAT and 4.87 iterations for 2-arcs VMAT). From the first optimisation plans comparison, CTV D99 of IMRT and 1 arc VMAT with vector-model-supported optimisation was 0.7 Gy higher than that of conventional manual optimisation. The volume receiving 35 Gy in the femoral head for 2-arcs VMAT plans was reduced by 10% with the vector-model-supported optimisation compared to the traditional manual optimisation approach. Otherwise, the quality of plans from both approaches was comparable. ConclusionVector-model-supported optimisation was shown to expedite the optimisation of IMRT/VMAT for prostate while maintaining plan quality.

O064 Knowledge based assessment of radiotherapy treatment plan quality for audit and quality improvement J. Simpson1,2, L. Wilton1, S. Bhatia1, P. Rouse3, A. Raith4 1

O063 Vector-model-based case retrieval approach in IMRT/VMAT prostate plan optimisation Eva Sau Fan Liu1, Vincent Wing Cheung Wu2, Benjamin Harris1, Margot Lehman1,3, David Pryor1,3, Lawrence Wing Chi Chan2 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]). 2Department of Health Technology and Informatics, Hong Kong Polytechnic University, Hong Kong. ([email protected]), ([email protected]). 3School of Medicine, University of Queensland, Australia. ([email protected]), ([email protected]) Introduction Lengthy time periods consumed in traditional manual optimisation can limit the use of intensity modulated radiotherapy / volumetric modulated radiotherapy (IMRT / VMAT). A vector model was therefore created using a combination of features extracted from several reference cases’ CT images and structure contours. To test this optimisation approach, planning parameters were retrieved from the selected most similar reference case and applied to test case to bypass the gradual adjustment of planning parameters. Method Each IMRT VMAT prostate reference database was comprised of 100 previous treated cases. Prostate cases were replanned with both conventional optimisation and vector-model-supported optimisation based on the oncologist’s clinical dose prescriptions. A total of 360 plans (30 cases of IMRT, 30 cases of 1-arc VMAT, 30 cases of 2-arcs VMAT plans including first optimisation and final optimisation with/without vector-model-supported optimisation) were

123

Department of Radiation Oncology, Calvary Mater Newcastle, Australia. 2School of Mathematical and Physical Sciences, University of Newcastle, Australia. ([email protected] [Presenting author]). 3Department of Accounting and Finance, Faculty of Business and Economics, University of Auckland, New Zealand. 4Department of Engineering Science, Faculty of Engineering, University of Auckland, New Zealand Introduction Institutional audit of treatment plan quality is rarely undertaken outside of clinical trials and yet as a complex process, continuing quality improvement should be a necessary requirement for treatment planning. A major reason for this is the lack of an efficient method for the assessment of plan quality. In this work we use knowledge-based methodology to analyse a database of treatment plans in order to provide a plan quality ranking. Methods For our knowledge-based methodology we use the operation research tool called data envelopment analysis (DEA). DEA is widely used in fields such as economics to determine the relative efficiency of processes and is well suited to treatment plan assessment because of its ability to determine a Pareto optimal frontier from the knowledge database. Previously we have shown that DEA can be used to assess prostate treatment plan quality [1]. Here we demonstrate the ability of DEA for plan audit by retrospectively analysing 41 IMRT prostate plans and assessing differences in 12 plan quality influencing factors between the plans ranked highest and those ranked lowest. Results Four influencing factors were found to be statistically significant between the highest and lowest ranked plans; the volume of PTV, the number of structures used in the optimiser, the number of rectal objectives used in the optimiser, and the number of overall objectives used in the optimiser. All planning technique influences demonstrated that the optimal plans were associated with the use of fewer structures or objectives whilst the smaller PTV’s were also associated with the optimal plans.

Australas Phys Eng Sci Med Conclusion DEA is an efficient tool for treatment plan quality audit. It can readily determine the Pareto optimal plan frontier and rank plans according to their relative quality that allows for the assessment of factors that influence plan quality including differences in planning technique. References 1. Lin K-M, Simpson J, Sasso G, Raith A and Ehrgott M. 2013. Quality assessment for VMAT prostate radiotherapy planning based on data envelopment analysis. Phys Med Biol. 58 5753-5769.

(p \ 0.05) and 299.46 ± 57.58 (noRF) and 267.25 ± 57.83 (RF) for the bladder (p \ 0.05). Conclusion All expert observers found the use of a centralised system of contouring for atlas development convenient and effective. This allowed for expedited collection of the data required to generate the multi-atlas and removed the uncertainties and bias introduced but using multiple different planning systems and data transfers. Acknowledgements The authors would like to acknowledge the assistance of Monica Harris and Olivia Cook from TROG and Laura O’Connor from Calvary Mater Newcastle.

O066 Changes to medium-energy kV X-ray calibration services in Australia O065 Automated individual case review for radiation oncology trials: The Radiotherapy Atlas Contouring Tool (TRAC) for PROMETHEUS M. G. Jameson1,2,3,4, J. A. Dowling5, K. Lim1,2,4, M. Berry1, J. de Leon6, D. Pryor7,8, J. M. Martin9, M Sidhom1, L. C. Holloway1,2,3,4,10 1

Liverpool and Macarthur Cancer Therapy Centres, Australia. Ingham Institute for Applied Medical Research, Liverpool, Australia. 3 Centre for Medical Radiation Physics, University of Wollongong, Australia. 4South West Clinical School, University of New South Wales, Australia. ([email protected] [Presenting author]). 5CSIRO Health and Biosecurity, The Australian e-Health Research Centre, Herston, Australia. 6Illawarra cancer Care Centre, Wollongong, Australia. 7Princess Alexandra Hospital, Brisbane, Australia. 8Queensland University of Technology, Brisbane, Australia. 9Calvary Mater Newcastle, Cancer Therapy Centre, Newcastle, Australia. 10School of Physics, University of Sydney, Australia 2

Introduction Automation of assessment of dosimetric objectives in clinical trials is feasible but automated evaluation of contouring accuracy, remains elusive. With appropriate trial specific knowledge, atlas based segmentation may provide a method of automating contouring QA in radiotherapy clinical trials (The RadioTherapy Atlas Contouring (TRAC) tool). This work will present the system description and infrastructure for the TRAC tool using the PROstate Multicentre External beam radioTHErapy Using Stereotactic boost (PROMETHEUS) trial as a test case Methods Contouring for multiatlas development was run in a centralised framework utilising the Trans-Tasman radiation oncology (TROG) quality management system. Rectal separation devices are recommended in the PROMETHEUS trial protocol. Ten exemplar patient MRI data sets (5xRectaFixTM and 5xSpaceOAR) were contoured by five expert observers. The organs to be contoured were the prostate, bladder and rectum. Atlas testing for both the RectaFixTM and SpaceOAR groups of MR images involved an MR image from each patient was selected as a target and all remaining MR images (n = 4) were registered against this target using symmetric rigid followed by diffeomorphic demons registration. Accuracy was asess using the dice similarity coefficient and the mean average surface distance. Results The DSC and MASD for the prostate and rectum were significantly improved with the RF in situ. The deformed volumes for the prostate, rectum and bladder were (mean ± r) 48.24 ± 19.48 cm3 (noRF) and 49.09 ± 19.2 (RF) for the prostate (p = 0.78), 54.81 ± 13.19 (noRF) and 91.66 ± 23.85 (RF) for the rectum

D. J. Butler, T. E. Wright, C. P. Oliver Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction During 2015 the ARPANSA Seifert X-ray system was replaced with a Gulmay Comet model NDI-320-26 tube and associated generators. A new series of beam qualities was commissioned. The primary standard is unchanged (a free-air chamber), and new correction factors were calculated for the new beam qualities. Methods New beam qualities were commissioned based on the German DIN 6809-4 standard therapy beams. These beams are available from the German standards laboratory PTB under the beam codes TH 50 to TH 320. We present results from commissioning and validation of the new beams, including a comparison with PTB. We suggest a method of interpolating from the standard beams to clinical beams, based on both kVp and HVL. Results Results of the calibration of ARPANSA calibration chambers on the previous system and on the new system agree to within 1.5%. Figure 1 shows the kVp and HVL of the new ARPANSA beams plotted over the clinical beams, indicating good coverage. The comparison with PTB indicated agreement to better than 0.2% over the range 100–280 kVp. Conclusion The new calibration beam qualities provide a more extensive set than the previous Seifert qualities. Uncertainties in ionisation chamber calibrations are of the same order as for the Seifert

Fig. 1 Comparison of new calibration qualities (open squares) against the existing clinical beam qualities in Australia [1]. The data is plotted for Al and Cu HVLs. The 70 kVp mark, below which Farmer-type chambers should not be used [2], is indicated with a dashed line

123

Australas Phys Eng Sci Med beams. As well as comparing the constancy of secondary standard chambers, two overseas comparisons support the validity of the new service. Users may experience small shifts in dosimetry (within the standard uncertainties of two calibrations) when adopting the new beams. References 1. Results of the 2009 dosimetry survey of Australian radiotherapy centres, R L Brown and D J Butler, Australas. Phys. Eng. Sci. Med., 33 No. 3 (2010) 285–297. 2. AAPM protocol for 40–300 kV x-ray beam dosimetry in radiotherapy and radiobiology, C.-M. Ma, C. W. Coffey, L. A. DeWerd, C. Liu, R. Nath, S. M. Seltzer, J. P. Seuntjens, Med. Phys. 28 (2001), 868–893.

Results For all three chamber types, the standard deviation in the ND,w coefficients for all four beams, within each type of chamber, is approximately 0.5%. The main source of variation is differences in chamber volume, although the Type A uncertainty in the calibration also contributes. The kQ values for 2571 chambers are shown in Fig. 1. Here the standard deviation is much smaller, 0.13%, as volume differences between the chambers cancel. The variation is similar for 30013 and FC65-G chambers. Also shown are the values from TRS398, and the values calculated by Muir and Rogers [2] using Monte Carlo methods. Conclusion The consistency of chamber calibrations has been observed over 2 years. The results are consistent with those of the field trial of this service in 2014. We note that adoption of the service entails a change in clinical dosimetry [1], within the combined uncertainties of the two methods but significant nonetheless. References

O067 Linac-based MV calibration service at ARPANSA: The first two years C. P. Oliver, G. Ramanathan, P. D. Harty, V. Takau, T. E. Wright, D. J. Butler Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction ARPANSA’s direct megavoltage calibration service for linac photon beams [1] has been running for over 2 years, with about two-thirds of facilities opting for the direct service in 2015/16. Here we present the collected results for some 24 thimble chambers calibrated using the service. Methods The measured energy correction factor kQ is compared for three common thimble chambers: the NE 2571, the PTW 30013 and the IBA FC65-G. The standard deviation of the ND,w calibration coefficients at 60Co, and 6, 10 and 18 MV linac beams are given, as well as that in the ratio to 60Co (i.e. kQ).

Fig. 1 Measured kQ for ten different 2571 Farmer chambers. Also shown is the kQ value from TRS-398, and that calculated using Monte Carlo methods by Muir and Rogers [2]. The standard uncertainty in the TRS-398 values is 1%

123

1. Direct megavoltage photon calibration service in Australia, D. J. Butler, G. Ramanathan, C. Oliver, A. Cole, J. Lye, P.D. Harty, T. Wright, D. V. Webb, Australas. Phys. Eng. Sci. Med. 37 No 4 (2014) pp. 753–761 (2014) 2. Monte Carlo calculations of kQ the beam quality conversion factor, B. R. Muir and D. W. O. Rogers, Med. Phys., 37, 5939 (2010)

O068 Development of a mail-out kV dosimetry audit using OSLDs F. Kadeer1, S. Keehan1,2, A. Alves2, M. Shaw2, F. Gibbons2,3, R. Franich1, I. Williams1,2, J. Lye1,2 1

School of Science, RMIT University. ([email protected] [Presenting author]), ([email protected]). 2Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3 Physical Sciences, Peter MacCallum Cancer Centre. ([email protected]) Introduction The ACDS currently provides a mail-out dosimetry audit for MV photon beams. The dose is measured using OSLDs in a perspex phantom and converted to dose to water under full scatter conditions. At present, there are no audits available for kV dosimetry. Methods Al2O3 OSLDs are known to over-respond to low energy x-rays. OSLDs were exposed to 1 Gy by a selection of kV beams to measure the energy dependence of the OSLD response and determine an energy correction factor. Monte Carlo modelling of the dose to the OSLD and the dose to water from kV beams of similar characteristics was performed using EGSnrc. Results The measured and Monte Carlo calculated KE are shown in Fig. 1. The differences are shown in Fig. 2 with one series of nonclinical filtered kV beams excluded. There is a trend in the agreement between the measured and Monte Carlo KE as a function of the energy, possibly due to a mismatch between the amount of Al2O3 in the model and the actual amount in the OSLDs. With the trend removed from the data, the standard deviation is 2.6%. Conclusion A standard deviation of approximately 2.5% would be acceptable for kilovoltage audits, where an action level could be set at twice the standard deviation, 5%, and an out of tolerance level at three times the standard deviation 7.5%. This is a higher tolerance than

Australas Phys Eng Sci Med 1.2 Measured

1.1

Monte Carlo

1

KE

0.9 0.8 0.7 0.6 0.5 0.4 0.3 50

100

150

200

250

300

Beam Energy (kV)

Fig. 1 The energy correction factor relative to a 100 kV beam as determined by measurement and by Monte Carlo modelling

KE % difference (MC to Measured)

14% y = 0.0005120x - 0.04761 R² = 0.6872

12% 10% 8% 6% 4% 2% 0% -2% 50

100

150 200 Beam energy (kV)

250

Introduction Over six years of operations the Australian Clinical Dosimetry Service, (ACDS) has developed and deployed a multilevel independent audit service which has gained international recognition. A summary of audit outcomes to-date will be presented and the process by which the ACDS will transition from a federallyfunded and voluntary service to one which is user-pays. Method The ACDS audit structure was, and is designed around, mutually supportive audits, each with a reference type measurement. Additionally, the ACDS recognises that it is neither a research nor clinical department at the cutting edge of treatment technology nor technique. The ACDS utilises a second-mover advantage to adapt existing approaches for its purposes. As such, the ACDS has deployed a three level audit service for conformal radiotherapy which is presently being adapted to also encompass IMRT and VMAT modalities. The assurance which the ACDS brings to radiotherapy providers is recognised through the efforts by a Jurisdictional Working Group, JWG, appointed through the Australian Health Ministers Advisory Council, (AHMAC) and Hospital Principals Committee, (HPC) to resolve long term future for the ACDS. This process will be presented. Results Over 90% of all Australian radiotherapy departments have been audited by the ACDS. A summary of all audit outcomes and recommendations will be presented. The National Dataset of audit data will be described, and how this information is being included in all higher level audit reports to provide clinics with technology specific audit outcomes. Additionally, the mechanism for moving to a user-pays system, approved by AHMAC, should be ready for tabling in November. Conclusion Since 2011 the ACDS pilot has developed an audit program which is internationally comparable. In 2017 the ACDS will move to a user-pays system following a model approved by the combined Australian jurisdictions. Acknowledgements The Australian Clinical Dosimetry Service is a joint initiative between the Department of Health and the Australian Radiation Protection and Nuclear Safety Agency.

300

Fig. 2 The difference between the Monte Carlo calculated and the measured KE for kV beams. One series of non-clinical kV beams has been excluded

used for the MV audits, however the overall accuracy required for kilovoltage radiotherapy is less strict than that needed for megavoltage therapies.

O070 Predictive power of the ACDS national data set: A case study in the AAA algorithm A. D. C. Alves1, I. Williams1,2, J. Lye1, M. Shaw1, S. Keehan1, F. Gibbons1, J. Lehmann3,4, L. Dunn5, J. Kenny5 1

O069 The ACDS: Summary of outcomes and (probable) future (How we got here, what we have, and where, maybe, we’re going) I. Williams1,2, J. Lye1, A. Alves1, M. Shaw1, S Keehan1, F. Gibbons1, J. Lehmann3,4, L. Dunn5, J. Kenny5 1

Australian Clinical Dosimetry Service (ACDS), ARPANSA. 2School of Sciences, RMIT University. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3Newcastle Calvary Mater Hospital. 4School of Medical Sciences, Sydney University. ([email protected]). 5Epworth Healthcare. ([email protected]), ([email protected])

Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2School of Sciences, RMIT University. ([email protected]). 3Calvary Mater Hospital Newcastle. 4Institute of Medical Physics, School of Physics, University. ([email protected]). 5Epworth Healthcare. ([email protected]), ([email protected]) Introduction The ACDS has conducted 65 end-to-end dosimetry audits of Australian radiotherapy facilities. The national data set provides powerful insight into how different treatment planning systems (TPS) are likely to perform under audit. Of interest is the Eclipse TPS using the AAA algorithm; known to over-calculate dose to tissue behind a low density inhomogeneity, leading to under-dose upon delivery [1]. Method The ACDS has examined national data in comparison with each facility’s own audit results. We have compiled data from 20

123

Australas Phys Eng Sci Med 6%

6%

6%

6%

4%

4%

4%

4%

2%

2%

2%

2%

0%

0%

0%

0%

-2% -4%

Dosimetric Offset (%)

-6% 1.560 6%

-2%

C2P1

-4%

1.580

1.600

1.620

-6% 1.560 6%

-2%

C2P4

-4%

1.580

1.600

1.620

-6% 0.420 6%

-4%

0.440

0.460

0.480

-6% 0.430 6%

4%

4%

4%

2%

2%

2%

2%

0%

0%

0%

0%

-4% -6% 1.130 6%

-2%

C4P2

-4%

C4P2fit 1.150

1.170

1.190

4%

-6% 1.850

-4%

C4P3fit 1.150

1.170

1.190

-4%

C5P2fit 1.900

1.950

2.000

-6% 1.850

-2%

C4P4

-4%

C4P4fit 1.150

1.170

1.190

-4%

C5P3fit 1.900

1.950

2.000

-6% 1.850

-6% 1.130 6%

0.450

0.470

0.490

Overriding machine parameters in R&V system

3

C4P5 C4P5fit 1.150

1.170

1.190

Incorrect transfer of plan to R&V system

2

Patient (phantom) positioning error - with image verification Patient (phantom) positioning error - without image verification

2 2

2%

0% -2%

C5P3

No. observed

C3B3P5

4%

2%

0% -2%

C5P2

-6% 1.130 6% 4%

2%

0%

-4%

-2%

C4P3

4%

2%

-2%

-6% 1.130 6%

Procedural error

-2%

C3B2P5

4%

-2%

Table 1 Summary of procedural errors in ACDS audits

0% -2%

C5P4

-4%

C5P4fit 1.900

1.950

2.000

-6% 1.850

C5P5 C5P5fit 1.900

1.950

Incorrect CT-ED table

1

Inhomogeneities OFF in TPS

1

2.000

Rao of MUs in the reference case to MUs in the behind lung cases

Fig. 1 The correlation between planned MUs and dosimetric offset for 12 behind lung points audits using Eclipse and have calculated the average and standard deviation of the dosimetric offset arising from the AAA algorithm. Results The average offset for AAA behind lung is 2.3% for 12 points in three separate phantom setups relative to the ACDS reference case. The average standard deviation in the dose offset for these 12 points is 1.0% leading to 0.3% uncertainty in the average offset. However, the dosimetric offsets observed from facility to facility range from 0.2 to 3.7%. This range, well outside the uncertainty, indicates that site specific implementations of the algorithm give rise to differing dosimetric offsets. Further, when the treatment planned monitor units are compared with the audit results there is a correlation (Fig. 1). Therefore, the reason for the differing offsets must lie deeper in the dosimetric chain than simple phantom misalignment. Conclusion The dosimetric offset from the AAA algorithm is not uniform across sites using the Eclipse TPS. As there is a correlation between planned MUs and dosimetric offset a prediction of the result can be made before the audit is measured. This indicates scope to provide a level of audit where only planning is required yet useful information on the quality of the plan may be derived. Acknowledgements The Australian Clinical Dosimetry Service is a joint initiative between the Department of Health and the Australian Radiation Protection and Nuclear Safety Agency. Reference 1. Dunn L et. al., National dosimetric audit network finds discrepancies in AAA lung inhomogeneity corrections, Phys Med. 2015 Jul;31(5):435–41.

O071 How to fail an ACDS audit M. Shaw, J. Lye, A. Alves, S. Keehan, F. Gibbons, I. Williams Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction The Australian Clinical Dosimetry Service (ACDS) has conducted over 250 dosimetry audits across Australian radiotherapy facilities, testing various aspects of the treatment pathway. Of the onsite audits, 25 have resulted in Out of Tolerance outcomes.

123

Method Analysis of the Out of Tolerance outcomes shows that 14 audit outcomes resulted from dosimetric inaccuracies, 10 from procedural errors, and 1 from both dosimetric and procedural errors. For the 11 audits where a procedural error occurred, the error was rectified after the initial measurement and follow up measurements returned a Pass result. This demonstrates the Out of Tolerance result could have been avoided (Table 1). Results Detailed examples of errors will be discussed, including instances where: • • •

Override of linac jaws by physicist Inhomogeneities unknowingly switched off in TPS Error in pre-treatment image matching

All examples observed in ACDS on-site audits are errors which may occur in patient treatment. Post-audit analysis of the errors identified the most commons reasons for the error occurring were: • •

Audit procedures deviated from standard clinical practice A lack of communication within the multi-disciplinary team

Conclusion The ACDS has observed common procedural errors which have led to an audit Out of Tolerance, which could also occur in patient treatment. Whilst the ACDS does not explicitly audit a facility’s procedures, audits should be conducted according to standard clinical protocols.

O072 IMRT and VMAT Level II audits by the ACDS F. Gibbons, J. Lye, M. Shaw, A. Alves, S. Keehan, I. Williams Australian Clincial Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction The ACDS conducts dosimetry audits across Australian Radiotherapy Centres, with levels of audit designed to test various aspects of the radiotherapy treatment chain. In response to clinical demand, a project was undertaken to integrate IMRT and VMAT dosimetry into the current Level II and Level III audits. Method IMRT and VMAT planning cases were designed for addition to the current Level II and Level III audits. Clinical plans were prepared based on the AAPM Publication TG1191 and adapted for use in the ACDS audit program. The Level II audit consists of a Solid water

Australas Phys Eng Sci Med phantom and the Octavius 1500 (PTW, Freiberg, Germany) 2D ion chamber array. The Level III audit is an end-to-end test using a humanoid thorax phantom (CIRS, Norfolk, VA) with CC13 ion chambers as the primary detectors. Upon completion of the design phase of the project, field trials of the IMRT Level II audit were performed. Results The field trials performed by the ACDS consisted of 35 Level II IMRT plan audits across 13 national facilities. The field trials utilised various equipment combinations to ensure compatibility across all vendors. Both 6MV and 10MV, flattened and non-flattened plans were evaluated. Scoring criteria utilising differing gamma criteria was established based on consensus between TROG and international IMRT/VMAT audits. The adopted gamma criterion of 3%/3 mm, 20% dose threshold was also evaluated relative to local dose difference in the target volumes and the audit results were shown to be comparable. Conclusion From July 2016, IMRT dosimetry will be offered by the ACDS in the Level II audit program. Expansion of the service to incorporate both VMAT Level II and IMRT/VMAT Level III dosimetry audits will follow shortly after. Reference 1. Ezzell GA, Burmeister JW, Dogan N, LoSasso TJ and Mechalakos JG, Mihanilidis D, Molineu A, Palta JR, Ramsey CR, Salter BJ, Shi J, Xia P, Yue NH and Xiao Y. (2009) IMRT Commissioning: Multiple institution planning and dosimetry comparisons, a report from AAPM Task Group 119. Med. Phys.; 36 (11): 5359–5373

Table 1 New modalities by Audit Level Year

Level

Modalities

New Install

Ib

Standard, FFF, Small field*

1

I

Standard, FFF*, kV**

2

II

3DCRT, IMRT, FFF, DARC*, Small field/SABR**

3

I

Standard, FFF*

4

III

3DCRT, IMRT, FFF, DARC*, Small field/SABR**

* 2017 ** 2018 with removable lungs and changeable central multi chamber insert, and single chamber measurements with supplementary film measurements. Conclusion The ACDS is developing a comprehensive suite of audit modalities aimed at ensuring patient safety across a range of clinical practice.

KS09 Adaptive radiation therapy @ NKI Jan-Jakob Sonke Department of Radiation Oncology, The Netherlands Cancer Institute, Amsterdam, The Netherlands. ([email protected])

O073 Future modalities of the ACDS audits J. Lye, M. Shaw, F. Gibbons, S. Keehan, A. Alves, I. Williams Australian Clincial Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction The ACDS has recently expanded its audits by incorporating IMRT and FFF components. In parallel developments, the dosimetry of small field, dynamic arc, and kV OSLDs have been investigated. The vision of the ACDS is to provide a comprehensive suite of audit modalities covering all common clinical practice, ultimately to ensure patient safety and to improve national dosimetry. The ACDS also aims to provide dosimetric information that can be used domestically and globally in the clinical trial setting. Method To ensure efficient delivery of the audit service, all modalities relevant to a facility’s clinical practice are measured in a single audit visit. The incorporation of new audit modalities requires a consideration of phantom design suitable for multiple modalities and limitations on facility and ACDS workload. Classification of new modalities and choice of associated cases need to take into consideration the utility for clinical trials. Results The ACDS has a planned roll-out of new modalities as shown in Table 1. Other new modalities under consideration are magnetic field dosimetry and 4D radiotherapy audits. To achieve the multiple modalities in a single audit, the Level II uses a lung and water slab phantom and a combination of ion chamber array and single chamber measurements. The Level III phantom uses a custom thorax phantom

Geometric uncertainties such as setup error, posture change, organ motion, deformations and treatment response limit the precision and accuracy of radiation therapy (RT). Consequently, the actually delivered dose typically deviates from the planned dose. To mitigate geometric and dosimetric uncertainties image guided and adaptive radiotherapy (ART) are being used. ART characterizes the patient’s specific variation through an image feedback loop and adapts the patients’ treatment plan accordingly [1]. In this presentation, various of our adaptive strategies under development and clinically implemented will be discussed. Reference 1. D. Yan et al. Computed tomography guided management of interfractional patient variation, Seminars in radiation oncology 2005;15(3):168–79

O074 The TROG 15.01 stereotactic prostate adaptive radiotherapy utilising KIM (kilovoltage intrafraction monitoring) (SPARK) trial: Report of the first patient treatment D. T. Nguyen1, R. O’Brien1, J.-H. Kim1, P. Greer2, K. Legge3, L. Wilton2, J. Booth4, P. Poulsen5, J. Martin2, P. Keall1 1

Faculty of Medicine, The University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Department of Radiation Oncology,

123

Australas Phys Eng Sci Med Calvary Mater Hospital, Australia. 3School of Mathematical and Physical Sciences, University of Newcastle, Australia. 4Northern Sydney Cancer Centre, Royal North Shore Hospital, Australia. ([email protected]). 5Department of Oncology, Aarhus University Hospital, Denmark. ([email protected])

Table 1 KIM accuracy (mean error) and precision (error standard deviation) with kV-MV triangulation as ground truth Accuracy

Precision

Left – Right Translation

0.3 mm

0.6 mm

Superior – Inferior Translation

-0.2 mm

0.3 mm

Anterior – Posterior Translation

0.2 mm

0.7 mm

Roll

0.4

2.2

Pitch

-0.4

4.7

Yaw

0.1

1.0

CTV Relative Volume (%)

Introduction Kilovoltage Intrafraction Monitoring (KIM) is an emerging technology that utilises continuous kV imaging during treatment to track the cancer target translational and rotational motions. KIM is applied within the SPARK trial (TROG 15.01) to enable gated prostate SBRT in a multi-institutional setting. The primary endpoint(s) of the study are to quantify accumulated patient dose distributions with the KIM intervention compared to dose distributions estimated without the KIM intervention. We present the results of the first patient treated on this trial. Method The patient received KIM-guided SBRT delivering 36.25 Gy to the PTV in 5 fractions. Treatment was delivered with RapidArc on a Varian TrueBeam accelerator. A gating threshold of 2mm for 5 seconds was applied to gate and shift. The accuracy of localisation with KIM was measured against MV/kV triangulation as the ground truth. The performance of KIM-guided gating to deliver the planned dose was measured using an isocentre shift dose reconstruction method. [3]. The delivered dose was reconstructed on the planning CT using the daily prostate motion trajectory and linac trajectory logfiles, and compared directly against the planned dose. Results KIM was utilised successfully in 4 of 5 fractions with 3 couch shifts due to large persistent prostate movements ([2 mm for more than 5 seconds). One fraction had erratic motion, requiring two couch shifts (total shift of 5.5 mm SI, 6.0 mm AP and 0.6 mm LR). KIM was not used in one fraction due to a segmentation error. KIM translational and rotational accuracy (mean error) and precision (standard deviation) in comparison with post treatment kV-MV triangulation is shown in Table 1. The dosimetric advantage of gating

100 80 Rectum

60 40 Bladder 20

Planned Delivered with KIM No motion correction

0 0

20

40 60 80 Relative Dose (%)

100

Fig. 2 DVH of reconstructed dose for the same fraction with KIM, without KIM and planned

in the fraction with erratic motion is shown in Fig. 1. The accumulated doses for the patient are CTV D98 = 36.3 Gy (36.1 Gy), PTV D95 = 35.8 (35.4 Gy), Rectum V50 = 28.5% (29.3%) and Bladder V50 = 24.4% (22.9%) with and without KIM gating, respectively. Conclusion The first patient was successfully treated on the SPARK trial. KIM-guidance achieved sub-millimetre accuracy and precision. KIM-guided gating improved the dose distribution delivered to the patient (Fig. 2). References 1. P R Poulsen, B Cho, and P J Keall, Real-time prostate trajectory estimation with a single imager in arc radiotherapy: A simulation study, Phys. Med. Biol. 54, 4019–4035 (2009). 2. J N Tehrani, R O’Brien, P R Poulsen and P Keall, Real-time estimation of prostate tumor rotation and translation with a kV imaging system based on an iterative closest point algorithm. Phys. Med. Biol., 2013. 58(23): p. 8517. 3. P.R. Poulsen, M.L. Schmidt, P. Keall, E.S. Worm, W. Fledelius, L. Hoffmann, A method of dose reconstruction for moving targets compatible with dynamic treatments, Medical physics, 39 (2012) 6237–6246. Acknowledgements SPARK trial is administered by TROG and is funded by a Cancer Australia grant. P Keall is funded by an NHMRC Fellowship.

O075 A simple adaptive radiotherapy approach to reducing geographic miss rates and toxicities for patients undergoing post prostatectomy radiation therapy C. Lee1, C. Lac1, A. Sims1, T. Eade2, R. David1, A. Michalski1, L. Bell2, A. Kneebone2 1

Fig. 1 Dose wash (90% to 100% dose, linear blue to red colour scale) of the dose reconstruction for the fraction with the largest motion with KIM (left) and without KIM (right) correction

123

Central Coast Cancer Centre, Gosford Hospital, Gosford, NSW. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Northern Sydney Cancer

Australas Phys Eng Sci Med Centre, Royal North Shore Hospital, St Leonards, NSW. ([email protected]), ([email protected]), ([email protected])

Fig. 1 A comparison of the mean DVHs for the 15 patients studied. Rectum and bladder DVHs are shown, using the PotD approach compared to the original plan using a standard PTV size and a bony anatomy CBCT match Introduction Post prostatectomy radiation therapy (PPRT) PTVs are subject to intra and inter-fraction motion due to the influence of bladder and rectum volumes. Traditionally, daily localisation is achieved by using the bony anatomy as a reference for IGRT. Relatively large fields are used to avoid geographic miss of the PTV, leading to toxicities of the bladder and rectum. An adaptive radiotherapy approach – ‘‘Plan of the Day’’ (PotD) – was devised, using soft tissue as the IGRT reference as well as variable PTV sizes to reduce the dose to bladder and rectum without compromising PTV coverage. Method A new daily soft tissue match IGRT protocol was devised. Two additional treatment plans were created, one with a PTV having smaller margins than the standard PTV, and the other with larger margins. The PotD approach involved performing a soft tissue match, assessing if the small PTV plan provided adequate coverage of the PTV, if not, the large PTV plan was used. A retrospective study of 15 PPRT patients was conducted where the PotD approach was applied. Results Of the cohort of patients examined, 7% of daily CBCT’s acquired showed a geographic miss when using a bony anatomy match. This was reduced to 0% when using the PotD approach. As seen in Fig. 1, the mean DVH reduced for both bladder and rectum, allowing all dose constraints to be met. Conclusion It has been demonstrated in a retrospective planning study that implementing a PotD approach to the treatment of PPRT is beneficial to the patients. The rate of geographic miss was reduced to 0% and a reduction in dose to the bladder and rectum.

O076 Dose verification in real time motion adaptive radiotherapy in a heterogeneous medium M. K. Newall1, M. Duncan1, V. Caillet2, J. T. Booth2,3, P. J. Keall4, M. Petasecca1, A. B. Rosenfeld1

localisation to modify beam shape and position during treatment [1, 2]. The aim of this study is to investigate the effects of tumour motion upon dose within low density mediums with high density in-homogeneities, such as in lung cancer, and identify effectiveness of MLC tracking in mitigating motion effects. Method The monolithic silicon detector DUO comprised of 505 diodes in intersecting orthogonal linear arrays, sensitive area 0.02 9 1 mm2, pitch 0.2 mm. The data acquisition system is synchronised with the electron gun pulses of the linac. DUO is encapsulated within a timber phantom, on a movable platform, mimicking electron density, scattering conditions and motion of lung. Two scattering environments are investigated; with 1cm diameter hemispheres (solid water) above and below detector, representing tumour, and without. An electromagnetic positioning system provides real-time information for MLC tracking. The tumour is contoured on a CT data-set, static gantry 3DCRT and IMRT treatment plans delivered. The treatments are delivered without motion, with motion and with motion and MLC tracking enabled. The measurements are repeated with EBT3 film for comparison. Results The profiles along the x and y-axis of DUO are compared with film for all cases and environments. The figure illustrates agreement between DUO and film for cases: (a) without motion, (b) motion and (c) motion and MLC tracking, in timber without tumour for 3DCRT. Measurements of full-width at half-maximum (FWHM) and penumbral width (20–80%) are both within 1.37 mm.

Conclusion Profiles measured by DUO show agreement with film with all measurements of FWHM within 5%. The results acquired by DUO are used to evaluate effects of inhomogeneity and motion upon treatments in low density mediums, and identify the effectiveness of MLC tracking in mitigating these effects. References 1. P. J. Keall, E. Colvill, R. O’Brien, J. A. Ng, P. R. Poulsen, T. Eade, A. Kneebone, and J. T. Booth, ‘‘The first clinical implementation of electromagnetic transponder-guided MLC tracking.,’’ Med. Phys., vol. 41, no. 2, p. 020702, Feb. 2014. 2. M. Petasecca, M. K. Newall, J. T. Booth, M. Duncan, a. H. Aldosari, I. Fuduli, a. a. Espinoza, C. S. Porumb, S. Guatelli, P. Metcalfe, E. Colvill, D. Cammarano, M. Carolan, B. Oborn, M. L. F. Lerch, V. Perevertaylo, P. J. Keall, and a. B. Rosenfeld, ‘‘MagicPlate-512: A 2D silicon detector array for quality assurance of stereotactic motion adaptive radiotherapy,’’ Med. Phys., vol. 42, no. 6, pp. 2992–3004, 2015.

1

Centre for Medical Radiation Physics, University of Wollongong, Wollongong, New South Wales, Australia. ([email protected],au [Presenting author]). 2Northern Sydney Cancer Centre, Royal North Shore Hospital, St. Leonards, New South Wales, Australia. 3Institute of Medical Physics, School of Physics, University of Sydney, New South Wales, Australia. 4Radiation Physics Laboratory, School of Medicine, University of Sydney, New South Wales, Australia Introduction Real-time adaptive radiotherapy aims to reduce effects of anatomy changes during radiotherapy by re-optimising delivery. MLC tracking is a motion adaptive strategy using real-time tumour

O077 Pulse by pulse timing analysis in adaptive radiotherapy: A preliminary study M. Duncan1, M. K. Newall2, V. Caillet2, J. T. Booth2,3, P. J. Keall4, A. B. Rosenfeld1, M. Petasecca1 1

Centre for Medical Radiation Physics, University of Wollongong, Wollongong, New South Wales, Australia. ([email protected]

123

Australas Phys Eng Sci Med [Presenting author]). 2Northern Sydney Cancer Centre, Royal North Shore Hospital, St. Leonards, New South Wales, Australia. 3Institute of Medical Physics, School of Physics, University of Sydney, New South Wales, Australia. 4Radiation Physics Laboratory, School of Medicine, University of Sydney, New South Wales, Australia Introduction Adaptive radiotherapy (ART) monitors and modulates the beam during treatment according to patient organ motion [1]. ART is often combined with stereotactic radiotherapy, so accurate and meaningful quality assurance requires both instantaneous dose and integral dose to be measured with high spatial resolution due to the small field irradiation conditions and steep dose gradients [2, 3]. High temporal resolution is required to evaluate the interplay effect between beam modulation and the tracking system used to mitigate the effect of organ motion [4]. Method Centre for Medical Radiation Physics developed a high resolution 2D silicon diode detector (DUO) with two perpendicular strip arrays of 256 diodes, pitch 200 m, area 1 9 0.01 mm2. Detector was placed on a movable platform to mimic patient motion and dose deposited was recorded for different delivery modalities with MLC tracking using Calypso RF array. DUO’s temporal resolution allows dose to be recorded for every linac pulse, allowing investigation of the interplay effect. Fourier analysis of the reconstructed time resolved detector response shows the frequency components of the motion, defined by several peaks. The integral area (a), height (h) and recurrence of peaks in the spectrum is related to the dose delivered due to motion. Smaller amplitude peaks correspond to more accurate target tracking; we define a ‘quality’ parameter, q, which is inversely proportional to the height and area of peaks in the Fourier spectrum 1 . q ¼ ah Results A sine wave motion with tracking shows a significant increase in the quality parameter, compared with no tracking; this reflects the effectiveness of the tracking algorithm used. Figure 1 summarises the findings for various motion patterns. Comparison of quality factors for various modalies and moons 350 300 250 200 150 100 50 0

Sin 10BPM

Sin 15BPM

Sin 20BPM

High Frequency

Large Excursion

Moon

22

21

20

14

35

Passive

41

96

42

29

96

Predicve

157

294

70

70

60

Conclusion We quantified the effectiveness of MLC tracking by defining a parameter based on the Fourier spectrum of the dose response. In future this quality factor can be used to optimize a tracking system for a specific patients’ organ motion. References 1. A. P. Shah, P. a Kupelian, T. R. Willoughby, and S. L. Meeks, ‘‘Expanding the use of real-time electromagnetic tracking in radiation oncology.,’’ J. Appl. Clin. Med. Phys., vol. 12, no. 4, p. 3590, Jan. 2011. 2. E. Pappas, T. G. Maris, F. Zacharopoulou, a. Papadakis, S. Manolopoulos, S. Green, and C. Wojnecki, ‘‘Small SRS photon field profile dosimetry performed using a PinPoint air ion chamber, a diamond detector, a novel silicon-diode array (DOSI),

123

and polymer gel dosimetry. Analysis and intercomparison,’’ Med. Phys., vol. 35, no. 10, p. 4640, 2008. 3. S. Benedict and D. Schlesinger, Stereotactic Radiosurgery and Stereotactic Body Radiation Therapy. 2014. 4. M. Petasecca, M. K. Newall, J. T. Booth, M. Duncan, a. H. Aldosari, I. Fuduli, a. a. Espinoza, C. S. Porumb, S. Guatelli, P. Metcalfe, E. Colvill, D. Cammarano, M. Carolan, B. Oborn, M. L. F. Lerch, V. Perevertaylo, P. J. Keall, and a. B. Rosenfeld, ‘‘MagicPlate-512: A 2D silicon detector array for quality assurance of stereotactic motion adaptive radiotherapy,’’ Med. Phys., vol. 42, no. 6, pp. 2992–3004, 2015.

O078 Correlating delineation uncertainty metrics with dosimetric outcome for cervical brachytherapy L. R. Bell1,2, T. P. Hellebust3, K. Bruheim4, P. Petricˇ5,6, K. Tanderup7,8, E. M. Pogson1,2, P. Metcalfe1,2, L. C. Holloway9,10 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. 2Liverpool & Macarthur Cancer Therapy Centres & Ingham Institute, Liverpool, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]) 3 Department of Medical Physics, Oslo University Hospital, Oslo, Norway. ([email protected]). 4Department of Oncology, Oslo University Hospital, Oslo, Norway. ([email protected]). 5Radiation Oncology Department, National Center for Cancer Care and Research, Doha, Qatar. 6Division of Radiotherapy, Institute of Oncology Ljubljana, Slovenia. ([email protected]). 7 Department of Oncology, Aarhus University Hospital, Denmark. 8 Institute of Clinical Medicine, Aarhus University, Denmark. ([email protected]). 9SWSCS, University of New South Wales, Australia. 10Institute of Medical Physics, University of Sydney, Australia. ([email protected]) Introduction It is important to understand the dosimetric impact of delineation uncertainty in gynaecological brachytherapy, since high dose gradients abut the delineated regions of interest. This study determines whether there are emerging correlations between the metrics for contouring variation and dosimetric outcomes. Method Six cervical brachytherapy MR images, each delineated by ten observers were utilised. Consensus contours were generated by experts to be used as the gold standard for each dataset. Brachytherapy treatment plans were generated and optimised to each observer volume. The shift in centre of mass (COMShift), maximal dimensions, volume and dice similarity coefficient (DSC) [1] in reference to the consensus volume were determined for each high-risk clinical target volume (HRCTV). The optimised plans for each HRCTV were applied to the consensus volume and the dose to 90 and 98% of the consensus HRCTV (D90, D98), as well as the equivalent uniform dose (EUD) were determined. Correlations between dosimetric and contouring metrics were determined using a Spearman’s nonparametric rank-correlation (p \ 0.05 significant). Results Preliminary results (first patient) are presented. While all plans achieved the aim of HRCTV D90 of [7.8 Gy, the coverage of the consensus volume by these plans did not meet this aim for 7/10 contours. This is despite a high mean DSC indicating high similarity of observer contours with the consensus. EUD values varied up to 1.52Gy with a mean of (8.5 ± 0.4)Gy. A statistically significant correlation was found for COM shift with D90 (p = 0.01) and with

Australas Phys Eng Sci Med Table 1 Mean contouring and dosimetric parameters

Mean SD

COMShift (cm)

Dimension

13.4

5.8

4.9

5.1

69.3

0.9

7.4

6.4

8.5

0.1

0.4

0.2

0.5

7.6

0.02

0.9

0.8

0.4

X Y Z Vol DSC D90 D98 EUD (Gy) (Gy) (Gy) (cm) (cm) (cm) (cm3)

D98 (p = 0.01). There was no significant correlation between DSC and DVH parameters (Table 1). Conclusion Preliminary results suggest that some contouring metrics (DSC in this context) may not give an accurate indication of the dosimetric impact of delineation uncertainty. For cervical brachytherapy, determining the shift in COM is more appropriate in estimating the D90 and D98 coverage.

Results The mean prostate volumes for FSCI and CIA methods for the five cases were 44.4cc ± 13.8cc and 46.6cc ± 14.7cc, respectively. Tables below are the dosimetric parameters from three treatment plan sets created. Two of the five cases were not planned due to prostate volume [50cc.

Case Frozen stepped image technique (FSI) V100% V150% V200 DHI D0.1cc-rectum (Gy)

D0.1cc-rectum) (Gy)

1

98.75

61.33

18.87

0.379

155.0

198.0

2

98.87

61.28

19.37

0.380

155.0

210.0

3

98.78

61.43

20.18

0.378

169.0

193.8

Mean

98.80

61.35

19.47

0.379

154.2

0.05

0.06

0.47

0.001

6.0

SD

Case References 1. Abreu R, Zoeteweij P, van Gemund AJC (2006) An evaluation of similarity coefficients for software fault localization in dependable computing Proc. 12th Pacific Rim Int. Symp. on Dependable Computing (PRDC’06) 39–46. doi: 10.1109/PRDC.2006.18

200.7 6.0

Continuous image acquisition technique (CIA) V100% V150% V200% DHI D0.1cc-rectum D0.1cc-rectum (Gy) (Gy)

1

97.86

59.13

18.34

0.396

175

198

2

97.04

58.52

0.19

0.379

181.7

211.7

3

98.47

60.86

20.32

0.382

183.3

202.5

Mean

97.79

59.50

19.14

0.386

169.4

202.9

0.50

0.90

0.78

0.007

3.0

5.0

Std Dev.

O079 Dosimetric characteristics of two prostate volume study techniques in LDR brachytherapy E. B. Estoesta1, S. Turner1, A. Hayden1, S. Zanjalani2, L. Martin2, W. Smith1, G. Bustillo1, H.Nguyen1 1

Crown Princess Mary Cancer Care Centre, Westmead, NSW, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Blacktown Cancer Care Centre, Blacktown, NSW, Australia. ([email protected]), (Wayne.Smith @health.nsw.gov.au) Introduction The prostate volume is a critical parameter in low dose rate (LDR) prostate brachytherapy to determine the patient’s viability for this treatment. Using transrectal ultrasound (TRUS), two methods of acquiring prostate volume were presented and compared in this study. The first technique commonly used in this centre is the FrozenStep Contoured Image (FSCI) and the second is Continuous Image Acquisition (CIA). This work compared the prostate volumes for each case obtained from these two methods and its dosimetric characteristic based from the prepared treatment plans. Method Five prospective prostate LDR brachytherapy patients underwent two methods of TRUS prostate volume study. FSCI technique acquired ‘‘frozen’’ image of each scan slice of the prostate, saved each image, then immediately contoured using the ultrasound system’s tool. Each stepped image of the prostate was contoured this way until the whole prostate volume was covered. CIA was done by continuously acquiring and saving all the raw stepped images then contour the prostate volume with the VariSeed software. A treatment plan using the Variseed software was created on each FSCI data for three cases and copied to its corresponding CIA data. The plans were compared in terms of dose coverage and dosimetric parameters such as V100%, V150%, V200%, Dose Homogeneity Index (DHI), D0.1ccRECTUM and D0.1ccURETHRA.

Conclusion Two methods of prostate volume study were presented. Continuous image acquisition technique yielded larger but more realistic prostate volumes due to better image coincidence with the prostate’s real-time motion during the volume study. This method should be independently planned to ensure better dose coverage.

O080 Dose-volume histograms and dose-surface maps after registering prostate external beam radiotherapy to a subsequent HDR brachytherapy boost: The doseresponse for gastrointestinal toxicities C. R. Moulton1, M. J. House1, V. Lye2, C. I. Tang2, M. Krawiec2, D. J. Joseph2, J. W. Denham3, M. A. Ebert1,2 1

School of Physics, University of Western Australia, Crawley, Western Australia. ([email protected] [Presenting author]). 2Radiation Oncology, Sir Charles Gairdner Hospital, Nedlands, Western Australia. 3School of Medicine and Population Health, University of Newcastle, New South Wales Introduction The dose-response of observed gastrointestinal toxicities for 118 prostate cancer patients was examined using registered dose-volume histograms (DVH) and dose-surface maps (DSM) from combined external beam radiotherapy (EBRT)/high-dose-rate brachytherapy (HDR). Method The HDR CT/dose (46 Gy/23 fractions) was deformablyregistered to the EBRT CT/dose (19.5 Gy/3 fractions) using VelocityAI. The EBRT and registered HDR TG43 dose distributions in a reference 2 Gy/fraction were summed. The rectal V1-80Gy, D1100% and thresholded DSMs (1–80 Gy) were calculated. Patients

123

Australas Phys Eng Sci Med were classed into toxicity/no toxicity groups if they had at least a certain late grade LENT-SOMA toxicity. The toxicities (thresholds) were rectal bleeding (G2), stool frequency (G2), diarrhoea (G1), urgency/tenesmus (G1), anorectal pain (G1) and completeness of evacuation (G2). The groups were compared via odds ratios (logistic regression) and median comparisons (Mann-Whitney U-tests). Results The mid-high dose range was significantly correlated with stool frequency, urgency/tenesmus and rectal bleeding (see figure). Rectal bleeding was also significantly correlated with near maximum doses (e.g. V80Gy/D1%). Doses to the posterior rectal surface within the first/last 20% of rectum length and the anterior rectal surface were significantly correlated with increased probability of rectal bleeding. The DSMs for other toxicities demonstrated a spatial correlation with dose e.g. dose to the posterior rectal surface near the inferior 60–80% of rectum length was significantly correlated with increased probability of stool frequency (see Figure). Features of thresholded DSMs at various dose levels were significantly correlated with toxicity e.g. rectal bleeding patients had greater widths, lengths, areas, compactness, circularity and fitted-ellipse lateral extent for the mid-high dose levels.

limited to 2D techniques. We present a method to perform pretreatment image verification, in the treatment bunker, by reconstructing the catheter positions in 3D for direct comparison with the treatment plan. Method A phantom was positioned on our customised brachytherapy treatment couch comprising an integrated flat panel detector, which is also used for in-vivo source tracking[1]. A ceiling-suspended x-ray system was used to capture two ‘shift’ images of the implant, with an imaging geometry illustrated in Fig. 1. Corresponding catheter paths were identified in each shift image and were back-projected to create a 3D reconstruction of the implant geometry. Registration with the treatment plan was performed, and comparison of the measured and planned catheters was performed to identify displacement. The sensitivity of catheter displacement detection was investigated by

Conclusion The DVHs and DSMs indicated that constraints on the intermediate and high-dose regions for total dose may be useful for rectal bleeding, stool frequency and urgency/tenesmus. Additionally, the features of DSMs revealed regions and spatial characteristics for each toxicity that may be useful for dose optimisation. Acknowledgements This work was funded by the NHMRC (1006447, 1077788), the University of Western Australia, an APA and an Ana Africh Scholarship.

O081 A ‘Shift Image’ 3D reconstruction technique for identification of catheter displacement in HDR prostate brachytherapy

Fig. 1 An illustration of the imaging geometry to acquire the two ‘shift’ images for catheter position reconstruction. The X-ray source is shifted by a distance d, FID is the focus to imaging plane distance and hm represents the height of the catheter above the imaging plane

R. L. Smith1, A. Haworth2,3, V. Panettieri1, J. Millar1,2, R. D. Franich2 1

Alfred Health Radiation Oncology, The Alfred Hospital, Melbourne, 3004, VIC, Australia. 2School of Science, RMIT University, Melbourne, 3000, VIC, Australia. ([email protected] [Presenting author]). 3Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne, 3002, VIC, Australia. ([email protected]). Alfred Health Radiation Oncology, The Alfred Hospital, Melbourne, 3004, VIC, Australia. ([email protected]), ([email protected]), ([email protected]) Introduction High-dose-rate (HDR) prostate brachytherapy can be performed using CT imaging for treatment planning. These images are a ‘snap-shot’ in time of the implant geometry and may not represent the geometry at the time of treatment, some hours later. Uncorrected catheter displacement can have a significant impact on dosimetry. Catheter displacement can be identified with pre-treatment imaging, but this can be difficult in the treatment bunker, and often is

123

Fig. 2 The average inferior displacement per catheter for the applied shifts of 2, 3, 5 and 10 mm. Error bars denote one standard deviation of catheter marker point difference in each measured catheter

Australas Phys Eng Sci Med applying known inferior displacements of 2, 3, 5 and 10 mm to the implanted catheters. Results The ‘shift’ image reconstruction technique provided a geometrically correct reconstruction of the implant volume. Assessment of the catheter displacement throughout the implant volume was possible, and not just at the catheter tip as typically performed with 2D verification approaches. The measured inferior catheter displacements are shown in Fig. 2, where all measured catheters reflect the applied displacements, within the determined measurement uncertainty. Conclusion We have demonstrated a 3D catheter reconstruction technique can be applied to HDR prostate brachytherapy to perform catheter displacement identification. References 1. R.L. Smith, A. Haworth, V. Panettieri, J.L. Millar and R.D. Franich, ‘‘A method for verification of treatment delivery in HDR prostate brachytherapy using a flat panel detector for both imaging and source tracking,’’ Med Phys. 2016 May;43(5):2435.

Fig. 1 Isodose lines from a dose grid generated with MaxiCalc (a), from OCB (b), and a comparison between the two, with the MaxiCalc data lines overlaid on the OCB data (c)

References

O082 MaxiCalc: A real-time dose calculation engine for dosimetric treatment verification in HDR brachytherapy

1. R.L. Smith, A. Haworth, V. Panettieri, J.L. Millar and R.D. Franich, ‘‘A method for verification of treatment delivery in HDR prostate brachytherapy using a flat panel detector for both imaging and source tracking,’’ Med Phys. 2016 May;43(5):2435.

M. Hanlon1, R. L. Smith1,2, R. D. Franich1 1

School of Science, RMIT University, Melbourne, VIC, Australia. ([email protected] [Presenting author]). 2Alfred Health Radiation Oncology, The Alfred Hospital, Melbourne, VIC, Australia. ([email protected]), ([email protected])

O083 Clinical response of a fast read-out, spectroscopic detector for eye plaque brachytherapy quality assurance

Introduction Treatment verification in HDR brachytherapy can be achieved by in vivo source-tracking with our flat panel detector (FPD) system [1]. Dosimetric assessment requires a fast dose calculation engine applied to the measured dwell positions. Here we present a TG-43 based dose calculation tool to calculate 3-D dose distributions for an arbitrary set of dwell positions. We validate it by benchmarking against a commercial treatment planning system (TPS). Method The dose calculation engine, dubbed MaxiCalc, accepts the measured dwell positions and times from our source-tracking system. Dose is calculated for the same volume, dose grid size and resolution as the plan to be compared. Validation was performed against Oncentra Brachy (OCB v4.3) by comparing 27 corresponding dose points, as per OCB commissioning, and by 3D dose grid comparisons. This was repeated for single source and multi-dwell plans. Results Across the 27 reference points, MaxiCalc differed from OCB Q by a mean of 0.08% ( = 0.07%) and maximum difference of 0.41%. These differences are similar to those between OCB and the TG-43 published values. Discrepancies between the two methods arise from interpolation differences due to coarseness of the calculation tables. The 3D dose grid generated can be exported in DICOM format for analysis in other software packages. For a 3D dose grid (1 mm voxels) around a single source, there is a maximum difference between MaxiCalc and the TPS of 1% for doses up to 200% of the prescription dose. Figure 1 shows that for a plan consisting of 13 dwell positions in a coronal plane, isodose lines for the MaxiCalc dose grid are coincident with those from the TPS. Conclusion MaxiCalc is a fast, accurate tool for calculating dose for arbitrary dwell positions and times. It is a valuable tool for real-time calculation of 3D dose in a HDR treatment verification system.

A. M. Kejda1,2, D. L. Cutajar1, M. R. Weaver1, M. Petasecca1, T. Al-Salmani1, P. White3, C. Pagulayan3, A. B. Rosenfeld1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. 2Blacktown Cancer and Haematology Centre, Blacktown Hospital, Australia. ([email protected] [Presenting Author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3Nelune Cancer Care Centre, Prince of Wales Hospital, Australia. ([email protected]), ([email protected]), ([email protected]) Introduction A fast read-out, spectroscopic dosimetry system has been developed at the CMRP, UOW, for use in eye plaque brachytherapy QA. It allows for dosimetric verification immediately prior to application of the plaque, whilst maintaining sterility. Previous studies have demonstrated the system is suitable for Eye plaque brachytherapy QA with low activity I-125 seeds [1]. The aim of this study was to test the system with a typical treatment dose rate, as well as to assess the sensitivity of the system to introduced plaque packing errors. Method The QA system was tested using a 15 mm ROPES plaque loaded with 10 seeds, measuring the depth-dose profile through the central axis of the eye plaque. Instantaneous dose rate readings were measured from 1.5 to 12 mm above the plaque in 0.5 mm steps. Each acquisition taking 10 s. This profile was compared to the dose rate calculated by the local TPS. Intentional plaque configuration errors

123

Australas Phys Eng Sci Med were introduced by moving seeds from their prescribed position, swapping them from an inner- to an outer-position in the plaque and vice-versa. The resulting depth dose profile was compared to TPS data. Results Results taken using the system were found to be in good agreement with the TPS data, with a maximum of 2.1% between the two. The introduction of seed placement errors resulted in a maximum deviation from TPS data by 7.3%. Conclusion A quality assurance system for the verification of I-125 based eye plaques, capable of fast measurements and maintaining a sterile plaque environment, has been successfully tested in a clinical activity range. Its agreement to the TPS data for the unmodified plaque, and its deviation when errors were introduced, confirm this system as a viable QA tool for eye plaque brachytherapy using I-125 seeds just prior to insertion into the patient. References 1. T.Jarema, et al, Dose verification of eye plaque brachytherapy using spectroscopic dosimetry, APESM, accepted for publication, 2016

O084 Comparison of TG-43 and AcurosTM BV for HDR surface moulds E. L. Boman, T. W. S. Satherley, N. Schleich, D. Paterson, R. J. W. Louwe, L Greig

Fig. 1 Side (a, c) and Top (b, d) views of the phantom geometries used in this study, including a geometry with 9 catheters (a–b), and a spherical geometry with 11 catheters (c–d). The various layers of each phantom represent: 0–3 cm water equivalent bolus (0 HU) on top of the canthers; (1) 0.5 cm water equivalent mould material and (2) 0.5 cm skin; (3) a layer of bone (1500 HU) or water depending on the simulation; and (4) a layer of water. Images also show the target volume (PTV, red) used in optimisation and a reference line (cyan). Dwell positions are separated by 0.5 cm using catheters with 1 cm distance from each other. The black solid and dotted lines represent the central catheter (single applicator with uniform dwell times) and the central dwell position (single source loading), respectively. For mesh loading, all indicated dwell positions either have uniform or optimised dwell times

1

Blood & Cancer Centre, Wellington Hospital, Wellington, NZ. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected])

Introduction High dose rate (HDR) brachytherapy using custom made surface moulds is widely applied to treat skin malignancies [1].The American Brachytherapy Society working group [2] suggests that no bolus is required above these surface applicators and that the TG-43 model is adequate to calculate the dose for surface applicators [3, 4]. However, TG-43 does not take into account the actual scatter conditions assuming full scatter in water in every direction [5]. In contrast, dose calculation models based on the actual medium such as AcurosTM BV (Varian Medical Systems, Palo Alto, CA) are proven to calculate the dose more accurately [6, 7]. In addition, Raina et al. [8] reported that bolus is required for surface moulds. The aim of the current study is to compare the doses from TG-43 model and AcurosTM BV for surface moulds. Method Dose distributions for flat and curved surface mould geometries with various bolus thicknesses (0–3 cm) as displayed in Fig. 1 were calculated using TG-43 and AcurosTM BV (Aria v11.0, Varian Medical Systems) [5] using a 1 mm calculation grid. Various dwell loadings were applied including a single source, a single applicator or mesh with all applicators with uniform or optimised dwell times (Fig. 1). Results Simulations without a bolus for both flat and curved geometries resulted in dose differences of 5–8% on average between AcurosTM and TG-43 (Fig. 2). For geometries with 1 cm bolus, the difference decreased to 3–5%. For curved geometries and optimised cases, the differences were slightly larger compared to flat geometry or uniform mesh loading, respectively. Conclusion This planning study indicated that bolus is required for HDR treatment of skin lesions if TG-43 based dose modelling is used.

123

Fig. 2 Dose differences (%) between AcurosTM BV and TG-43 calculations for different phantom geometries as detailed in Fig 1, and various source loading patterns. Negative differencies indicate AcurosTM BV doses being smaller

Australas Phys Eng Sci Med Further investigations including accurate dosimetry will be conducted to confirm these results. References 1. Lee CD (2014), Recent developments and best practice in brachytherapy treatment planning. Br J Radiol 87 2. Ouhib Z, Kasper M, Calatayud JP, Rodriguez S, Bhatnagar A, Pai S, Strasswimmer J (2015) Aspects of dosimetry and clinical practice of skin brachytherapy: The American Brachytherapy Society working group report. Brachytherapy 14:840–858 3. Vijande J, Ballester F, Ouhib F, Granero D, Pujades-Claumarchirant MC, Perez-Calatayud J (2012) Dosimetry comparisons between TG-43 and Monte Carlo calculations using the Freiburg flap for skin high-dose-rate brachytherapy. Brachytherapy 11:528–535 4. Granero D, Perez-Calatayud J, Vijande J, Ballester F, Rivard MJ (2014) Limitations of the TG-43 formalism for skin hgh-doserate brachytherapy dose calculations. Med Phys. 41(2) 5. Varian Medical Systems (2015) BrachyVision Algorithms Reference Guide. Varian Medical Systems Inc., Palo Alto, CA, USA 6. Papagiannis P, Pantelis E, Karaiskos P (2014) Current state of the art brachytherapy treatment planning dosimetry algorithms. Br J Radiol 87 7. Beaulieu L, Carlsson Tedgren A, Carrier JF, Davis SD, Mourtada F, Rivard MJ, Thomson RM, Verhaegen F, Wareing TA, Williamson JF (2012) Report of the Task Group 186 on modelbased dose calculation methods in brachytherapy beyond the TG43 formalism: Current status and recommendations for clinical implementation. Med Phys 39(10):6208–6236 8. Raina S, Avadhani JS, Oh M, Malhotra HK, Jaggernauth W, Kuettel MR, Podgorsak MB (2005) Quantifying IOHRD brachytreapy underdosage resulting from an incomplete scatter environment. Int J Radiation Oncology Biol Phys 61(5):1582–1586

O085 The introduction and validation of a model based dose calculation algorithm for brachytherapy in an Australian Centre L. J. Hamlett1, M. A. Powers1, M Roche1, D. A. Wood2, Dr O. L. Fourie1 Medical Physics Department, The Townsville Hospital, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Brachytherapy Advanced Practitioner, Australia. ([email protected]), ([email protected])

Fig. 1 Visual dose difference comparision of TG-43. TG-186 ACE for a 16 titanium catheter prostate plan

Fig. 2 DVH comparision for TG-43 V TG-186 prostrate type plan

catheter plan was also calculated assuming plastic catheters. Each plan was recalculated using ACE standard and high accuracy and compared to the TG-43 reference plans. Plans were compared visually using the ‘Analyse’ module in Oncentra, using point doses and for the prostate type plans via dose volume histogram (DVH). Results Level 1 results were similar to those published by Ma et al [3], i.e. dose differences were seen on the single dwell plan in regions of low dose positioned on the longitudinal source axis. Level 2 plan results using the WGR data showed small differences between ACE TG-186 calculations and the WGR data. The prostate plan with plastic catheters showed good agreement between TG-43 and ACE. With titanium catheters dose variation was visible (Fig. 1) and affected the DVH i.e Urethra D10% reduced from 114.98 to 112.08% (Fig. 2). The TG-186 accuracy modes showed negligible difference for the clinical plans, calculation times range from 10 mins for standard to 120 mins for high. Conclusion ACE calculation algorithm produces water equivalent reference phantom dose distributions similar to TG-43. The prostate plans suggest that titanium needles could potentially introduce a dose distribution variation which could affect plan quality. References

Introduction This work covers the preliminary findings of validating Elekta’s Oncentra commercially available Advanced Collapsed Cone Engine (ACE). Method Following AAPM report TG-186[1], reference plans were calculated using TG-43 calculation for both level 1 and level 2 conditions. For level 1 three plans; single dwell; two dwell; and an eight dwell circular arrangement were produced in a water equivalent cube phantom. Level 2 used test plans with non water equivalent phantom materials such as air, available from the AAPM/ESTRO/ABG working group registry (WGR) [2]. Finally complex ‘HDR prostate type’ CT plans using titanium catheters with various catheter/dwell numbers (12/137, 16/205 and 20/205) were investigated. The 16

1. Beaulieu L, Carlsson Tedgren A, Carrier JF, Davis S D, Mourtada F, Rivard M J, Thomson R M, Verhaegen F, Wareing T A, Williamson J F (2012) Report of the Task Group 186 on model-based dose calculation methods in brachytherapy beyond the TG-43 formalism: Current status and recommendations for clinical implementation. Med. Phys. 39 (10) 6208–36 2. AAPM/ESTRO/ABG Brachytherapy Working Group 3. Ma Y, Lacroix F, Lavallee MC, Beaulieu L Validation of the Oncentra Brachy Advanced Collapsed Cone Engine for a commercial Ir192 source using heterogeneous geometries. Brachytherapy 14 (2015) 939–952

123

Australas Phys Eng Sci Med

KS10 Breast imaging: Virtual clinical trials Hilde Bosmans, Lesley Cockmartin, Chantal Van Ongeval, Nicholas Marshall KU Leuven, Belgium. ([email protected] [Presenting author]) New breast imaging techniques for breast cancer screening face important challenges: the number of missed cancers has to be further decreased; the fraction of small cancers at detection should be increased. In diagnostic work-up, fast acquisition and more accurate data would be highly welcomed. In order to study the impact of new technical factors on these endpoints, clinical trials are the gold standard. With many new techniques entering the radiological practices - most of them with a large parameter space—a large number of trials could be launched. Clinical trials are however expensive and time consuming. In addition to this, a clinical trial that is representative for one country, may have to be repeated in another country as the conditions may be very different. A typical example is the recall rate, which is sometimes very low or on the contrary very high. A lowering of the recall rate may be a positive end point in a high recall rate country, but not proven and not aimed for in low recall rate countries. A study for 50–65 years old women may have a different result from a larger population, starting at age 40 years. Virtual clinical trials are being set up to solve some of the questions that would normally be addressed with a clinical trial. In its ultimate implementation, a virtual clinical trial uses computer models of the breast that are representative for the population to be represented, well characterized virtual lesions are simulated into these models, projections are calculated from these models, these projections could be reconstructed into a tomosynthesis data set and model observers would then score the images in terms of detectability or any other characteristic. The scientific community is working at several components of such an imaging chain. During the presentation we will share our experience with (partially) virtual clinical trials: their basic idea, the tools, the implementation and practical results.

O086 New technologies in digital mammography Jennifer Diffey Hunter New England Imaging, John Hunter Hospital, New Lambton, NSW. ([email protected]) Introduction Digital mammography offers several advantages compared to screen-film mammography. In particular, images have improved dynamic range, better low contrast resolution and are acquired at a lower radiation dose. Additionally, higher sensitivity in dense breasts and increased detection rates of high risk lesions have been reported. Despite these improvements, digital mammography is not infallible as a screening tool. There are some important outstanding limitations to overcome, most notably lower cancer detection rates in dense breasts and overdiagnosis. Method Rapid advances in digital mammography detectors have paved the way for new technologies which may address some of these issues. Examples include tomosynthesis, contrast-enhanced spectral mammography (CESM), stereoscopic digital mammography and dedicated breast computed tomography (DBCT). There are also advanced software applications which enable us to gain useful information from existing mammographic images, such as quantitative volumetric breast

123

density which can be used to inform follow-up or intervention strategies in women with high breast density. Breast density has been shown to be a valuable inclusion in risk prediction models, which could be used to tailor screening intervals to individual risk. Results The principles of each technique will be explained. Associated benefits and limitations will be discussed including dose implications, whether the techniques are suitable for screening and / or assessment and how they address the limitations of overdiagnosis and the difficulty of detecting disease in dense breasts. Conclusion New digital mammography technologies continue to show great promise, having already demonstrated improvements in sensitivity and specificity (especially in dense breasts), reduced recall rates and increased confidence in diagnosis. Tomosynthesis is rapidly being rolled out throughout Australia and several sites are using contrast-enhanced spectral mammography. It is important that medical physicists are aware of these developments as new quality control tests and dosimetry models will potentially be required.

O087 Analysis of patient breast dose from a mammographic biopsy unit I. D. McLean1, L. J. Ryan2 1

Medical Physics and Radiation Engineering (MPRE), Canberra Hospital, Australia. ([email protected] [Presenting author]). 2 MPRE, Canberra Hospital, Australia. ([email protected]) Introduction The asymptomatic nature of breast screening examinations has been a driving force for the optimisation of routine breast imaging. However mammography of patients referred for biopsy assessment, classed as symptomatic, has received little attention. This work reports on a patient dose evaluation from one biopsy unit involved in breast screening. Method Radiographic and patient/clinical data was collected from 224 biopsy procedures conducted with a Siemens Mammotest unit from 2013 to early 2016. Radiographic data including the kV, mAs and number of exposures was coupled with equipment QA data to determine the mean glandular dose after consideration of breast thickness and breast size. Glandular tissue was assumed to be 50%. Additional procedural information, including the biopsy result, was also collected. Importantly biopsy mammography typically includes partial irradiation of the breast which was considered as recommended by the NHSBSP1. Results On average breast thickness was 5.6 cm and was exposed at 28.9 kV for 170.8 mAs with 8.5 exposures giving a mean glandular dose to the exposed area of 23.2 mGy. When considering that typically the breast field was 30% of the area of a single breast, the dose to the breast organ was 3.5 mGy. The procedure outcomes were positive 34% of the time, with 4% precancerous and hence needing ongoing monitoring and 3% abandoned. Conclusion The above study shows the importance of considering the irradiated proportion of the breast in biopsy dosimetry and serves as a platform for future comparison to new biopsy technologies now under active consideration. Acknowledgements I would like to acknowledge the invaluable assistance of the radiographic staff of BreastScreen ACT for the careful collection of data essential for this study. References 1. NHSBSP, Commissioning and routine testing of small field digital mammography systems Report No. NHSBSP Equipment Report 0705, 2007.

Australas Phys Eng Sci Med

O088 Update: ACPSEM position paper recommendation for digital mammography quality assurance program - tomosynthesis addendum L. Cartwright1, P. J. Barnes2 Medical Physics Specialist, Westmead Hospital, NSW. ([email protected] [Presenting author]). 2Medical Physics, MIA Radiology, VIC ([email protected]) Introduction In March 2015 the Australian tomosythesis working party was formed and charged with updating the existing mammography quality control (QC) recommendations to include Digital Breast Tomosynthesis (DBT). These include QC recommendations for radiographers (imaging technologists), and annual equipment testing. The recommendations are now near completion, including an update to the standard 2D QC tests. Method In preparing the new QC recommendations, a review of DBT QC guidance from Europe, America and each DBT system manufacturer was undertaken. Comprehensive QC tests and dosimetry on each commercially available DBT system have been performed to evaluate the validity of the proposed tests; giving careful consideration to the resources and expertise available in the Australasian environment. Results Recommended DBT QC tests for radiographers include: Artefact Evaluation, Image Quality Evaluation and Detector Calibration. Recommended DBT QC tests for annual equipment testing include: Missing Tissue, Image Quality Evaluation, AEC/SDNR, HVL, MGD, and Artefact Evaluation. Conclusion Testing recommendations are provided on routine QC tests for tomosynthesis systems. These also include an update to the current 2D tests. The new recommendations cover annual QC requirements and testing guidance after major equipment modifications, for example following tube changes. The existing testing Excel spreadsheet has been updated to assist users with the new protocols.

O089 A monolithic silicon detector array for small field dosimetry in stereotactic radiotherapy: DUO K. Al shukaili1, Stephanie Corde1,2, Marco Pettasecca1, M. Lerch1, A. B. Rosenfeld1 1

Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. ([email protected] [Presenting Author]). 2 Prince of Wales Hospital, Randwick, NSW, Australia. ([email protected]), ([email protected]), ([email protected]), (mailto:[email protected]) Introduction Stereotactic radiosurgery (SRS) is a technique that uses small, highly collimated photon beams, which requires high geometric precision and dosimetric accuracy1–3. The silicon diode arrays are commonly implemented in the quality assurance for SRS, as they have a number of advantages including: real time operation (compared to the film) and high spatial resolution and small size (compared to ionizing chambers)4–5 . This work aims to characterize the monolithic silicon diode array named ‘‘DUO’’ designed for QA in stereotactic radio surgery (SRS). Method DUO is a silicon monolithic detector manufactured on a p-type substrate, which was designed by CMRP at UOW as in Fig. 1. The pixels are arranged in two linear arrays orthogonal each other

Fig. 1 a DUO detector mounted on a thin PCB, b schematic of the DUO packaging

with 256 individually readout pixels for each arm. The pixel pitch is 200 lm and an overall detector area is 52 9 52 mm2. DUO is wire bonded on an almost tissue equivalent printed circuit board 0.5 mm thick. Characterization of DUO was performed, including beam profiles, percent depth dose and output factor for SRS cone collimator beams on Elkta LINAC for cone size down to 5 mm. The measurements were repeated by using EBT3 film and Stereotactic field diode (SFD), to compare with the DUO detector. Results The profiles of cone SRS for cone diameters from 5 to 50 mm, shows agreement in the FWHM and (20–80)% penumbra with EBT3 and SFD within 1.5%. The output factor agrees within 1% when compared to EBT3 film and SFD detectors for cone sizes down to 5 mm. The measured depth dose response agreed to within 2%, when compared to EBT3 films and SFD detector for depths beyond the build up region. Conclusion The response of the DUO detector has been evaluated and demonstrated suitability for use in SRS QA as a sub-millimetre spatial resolution real time dosimetry. References 1. Alfonso, R., Andreo, P., Capote, R., Huq, M. S., Kilby, W., Kja¨ll, P., & Ullrich, W. (2008). A new formalism for reference dosimetry of small and nonstandard fields. Medical physics, 35(11), 5179–5186. 2. Das, I. J., Ding, G. X., & Ahnesjo¨, A. (2008). Small fields: nonequilibrium radiation dosimetry. Medical physics, 35(1), 206–215 3. Aspradakis, M. M., Byrne, J. P., Palmans, H., Duane, S., Conway, J., Warrington, A. P., & Rosser, K. (2010). IPEM report 103: Small field MV photon dosimetry. 4. Scott, A. J., Kumar, S., Nahum, A. E., & Fenwick, J. D. (2012). Characterizing the influence of detector density on dosimeter response in non-equilibrium small photon fields. Physics in medicine and biology, 57(14), 4461. 5. Wong, J. H. D., Knittel, T., Downes, S., Carolan, M., Lerch, M. L. F., Petasecca, M., & Rosenfeld, A. B. (2011). The use of a silicon strip detector dose magnifying glass in stereotactic radiotherapy QA and dosimetry. Medical physics, 38(3), 1226–1238.

O090 Small field detector correction factors: effect of the flattening filter for Elekta and Varian linacs M Tyler1, P Liu2,3, C. Lee4, D. R. McKenzie2 1

Shoalhaven Cancer Care Centre, Nowra, NSW. ([email protected] [Presenting author]). 2School of Physics, University of Sydney, NSW, Australia. 3Chris O’Brien

123

Australas Phys Eng Sci Med 1

Royal Brisbane & Women’s Hospital, Brisbane, Australia. Queensland University of Technology, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 3Genesis Cancer Care Queensland, Brisbane, Australia. ([email protected])

2

Fig. 1 Ratio of detector response in an FFF field to that of an FF field as a function of nominal field width for ion chambers and solid state detectors Lifehouse, Sydney, NSW, Australia. ([email protected]). 4 Central Coast Cancer Centre, Gosford Hospital, Gosford, NSW, Australia. ([email protected]), ([email protected]), ([email protected]) Introduction The use of flattening filter free (FFF) beams for stereotactic radiosurgery (SRS) and stereotactic ablative radiation therapy (SABR) treatments has led to detectors, designed for dosimetry of flattened radiation beams (FF), being used in FFF beams. This study assessed the effects of the flattening filter on measured output factors for Elekta and Varian linear accelerators in nominal field sizes as small as 5 mm. Method Relative output factors were measured using FF and FFF modes for nominal field widths of 5 to 100 mm for 6 MV photon beams generated by an Elekta Axesse (Elekta, Crawley, UK) and a Varian iX (Varian, Palo Alto, CA) linear accelerator. Measurements were made with a range of detectors (diodes, ionisation chambers, radiochromic film and a diamond detector) and referenced to scintillation detector measurements to obtain a relative detector response (RDR). The ratio of RDR for FFF to FF beams as a function of nominal field size was used to assess the effects of the flattening filter on output factors. Results The removal of the flattening filter affects the output factors measured with solid state detectors (diodes and diamond), with deviations up to ±1.6% between FFF and FF modes. Ionisation chambers showed minimal change with deviations less than ±0.9% across all field sizes measured. Increased deviations for solid state detectors was found for the Varian iX at the smallest fields, which can be attributed to the larger difference in energy between the FFF and FF modes (Fig. 1). Conclusion Small field correction factors derived for each detector studied are interchangeable for a linear accelerator for 6MV FFF and FF modes, but with an additional uncertainty of up to ±1.6%. Acknowledgements The authors wish to acknowledge Anna Ralston and Kirbie Sloan for assistance with scientific discussion and data collection, and Peter Douglas (Nucletron Australia) for the loan of the microDiamond used in this study.

Introduction Collimation system reproducibility is an important parameter to be considered in the clinical introduction of stereotactic radiosurgery. Charles et al. [1] reported that a 1 mm variation in field size can result in central axis dose errors up to 20% (for a 5 mm field size). Large-area parallel-plate ionization chambers have been proposed for use in measuring output factors in small fields [2]. This study investigates the use of the more common Roos chamber to measure small field output. Method Dose-area product measurements were performed using a PTW Roos chamber in a water-equivalent phantom for small fields produced on a Varian 21iX accelerator. Four field sizes were investigated: 1 9 1 cm2 (jaws), 1 9 1 cm2 (M120 MLC), and 0.6 9 0.6 cm2 and 1.2 9 1.2 cm2 (BrainLab m3 mMLC). For each field size, measurements were additionally performed with introduced position errors of ±1 and ±2 mm. All measurements were repeated with collimators retracted and repositioned, to evaluate reproducibility. Field area was independently measured using EBT3 film. Chamber position sensitivity was characterised for the 0.6 9 0.6 cm2 and 1 9 1 cm2 fields, with 0.5–3 mm shifts in lateral and longitudinal couch directions. Results Output measurements, shown relative to 10 9 10 cm2 field dose, can be seen in Figs. 1 and 2. A linear relationship between

Fig. 1 Modified output factors, normalised to 10 9 10 cm2 output, for very small fields

O091 Measuring very small field output factors with a Roos chamber S. B. Crowe1,2, T. L. S. Tang2, S. R. Sylvander1, T. Kairn2, 3

123

Fig. 2 Modified output factors, normalised to 10 9 10 cm2 output, for measured field sizes

Australas Phys Eng Sci Med output and field size can be observed where the field is contained within the diameter of the chamber active volume. Variations of nominal jaw field edge of 1 mm resulted in output variations of up to 11% (3.3% relative to 10 9 10 cm2 field). 1 mm shifts in chamber position resulted in mean output variations of 0.6 and 1.2% for 0.6 9 0.6 cm2 and 1 9 1 cm2 fields. Conclusion The Roos chamber can be used for constancy measurements, and may be useful for verifying output after adjustments to beam steering in MLC-based systems.

Conclusion The comparison was a successful way to compare different methods of relative dosimetry in small fields. The level of agreement between the results suggests that some clinics are underestimating their uncertainties. Acknowledgements We would like to thank the many facilities that participated in the comparison and the people who provided suggestions and comments on the comparison protocol. In particular we would like to thank Madelaine Tyler and Simon Downes.

References

KS11 Protein nanocages as a display platform with high spatial control

1. Charles, P. H. et al. (2014). A practical and theoretical definition of very small field size for radiotherapy output factor measurements. Medical Physics 41(4): 041707. 2. Djouguela, A. et al. (2006). The dose-area product, a new parameter for the dosimetry of narrow photon beams. Zeitschrift fu¨r Medizinische Physik 16(3): 217–227.

O092 Results of a multi-centre small-field dosimetry comparison C. P. Oliver1, V. Takau1, G. Ramanathan1, D. J. Butler1, I. Williams2 1

Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Medical Radiation Services Branch, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]) Introduction Small fields are routinely used in modern radiotherapy, and ablative treatment approaches are utilising increasing greater doses per fraction but there is as yet no internationally accepted calibration/dosimetry protocol. Instead, clinics adopt published dosimetry methods and satisfy themselves of the accuracy by comparing different methods. This situation represents an increasing national risk which ARPANSA is attempting to address by developing small-field dosimetry methods. A set of cones were purchased in 2015, making it possible to compare these clinical methods more comprehensively. Method Teams of physicists from 15 radiotherapy facilities travelled to ARPANSA and measured the output factor for an Elekta 5 mm cone on ARPANSA’s Elekta Synergy Linac. The output factor was defined as the ratio of the central-axis absorbed dose to water under the cone, to the same quantity under the 10 cm 9 10 cm MLC-defined field, at a depth of 5 cm and source-surface distance of 95 cm. Facilities reported the output factor and its uncertainty. They also provided raw ratios and details of correction factors used, whether their method was also used to provide clinical dosimetry for 5 mm fields, and if there were any shortcuts taken as a result of the severe time restrictions (2.5 hours per group). Results A total of 30 unique measurements of the output factor were made using 7 different methods. The average of all the results (excluding an expected outlier from an ionisation chamber) was 0.616 with a standard deviation of 0.022 (3.6%). The largest deviations were ±8% from the average. The average of the standard uncertainties reported by clinics was 1.7%. The lowest estimate was 0.1% and the highest 2%.

Sierin Lim School of Chemical and Biomedical Engineering, NTU-Northwestern Institute for Nanomedicine, Nanyang Technological University, Singapore. ([email protected] [Presenting author]) Introduction Protein nanocages can be engineered to tailor their function as carriers for therapeutic and diagnostic agents. They are formed by the self-assembly of multiple subunits forming hollow spherical cage structures of nanometer size. Due to their proteinaceous nature, the protein cages allow facile modifications on their internal and external surfaces, as well as the subunit interfaces. In this presentation, we will focus on the external surface modifications of two protein cage representatives that are ferritin from Archaeoglobus fulgidus and E2 protein from Geobacillus stearothermophilus as a display platform with high spatial control for enhanced localization to specific cells and for potential vaccine applications. Method Display of targeting ligands on protein nanocages is achieved by two methods: (1) genetic modification and (2) chemical conjugation. Enhanced localization was tested by incubating the modified protein nanocages with cancer or skin cells and examined by microscopy and flow cytometry. Results We have displayed various moieties on the protein nanocages which include prostate-specific antibody and EGFR scFv through chemical conjugations, and virus epitopes (e.g. Chikungunya, respiratory syncytial virus) and small targeting peptides through genetic engineering. The moieties are displayed with high precision and the number of displayed moieties varies depending on the size of the displayed molecules. Antibody and antibody fragments ([20kDa) pose steric hindrance compared to short peptides and epitopes (\5 kDa). The protein nanocages retain their self-assembly properties upon display of the moieties. Protein nanocages displaying targeting moieties show enhanced localization on prostate and skin cells compared to unmodified controls. Conclusion Protein nanocage is a universal platform to display multiple moieties. An advantage of the protein cage over other nanoparticles are the spatial control of the displayed moieties owing to the precise locations of each amino acid on the surface of the protein nanocages. Enhanced localization on specific cells can be achieved by changing the displayed targeting moieties. References 1. Bu¨cheler J, Howard C**, de Bakker CJ, Goodall S, Jones ML, Win T, Peng T, Tan CH, Chopra A, Mahler S, Lim S* (2015) Development of a Protein Nanoparticle Platform for Targeting EGFR Expressing Cancer Cells, Journal of Chemical Technology & Biotechnology, 80(7):1230–1236. 2. Walsh EG*, Mills DR, Lim S, Sana B, Brilliant KE, Park WKC (2013) MRI Contrast Demonstration of Antigen-specific

123

Australas Phys Eng Sci Med Targeting with an Iron-based Ferritin Construct, Journal of Nanoparticle Research 15:1409–1418.

O093 On the interactions of a gold nanoparticle (GNP) cluster in a photon radiation beam 1,2

3,4

2

C. Kirkby , N Suchowerska , B. Koger , D. R. McKenzie3 1

Department of Oncology, Jack Ady Cancer Centre: Lethbridge. Department of Physics and Astronomy, University of Calgary. ([email protected]). 3School of Physics, University of Sydney, NSW, Australia. 4Chris O’Brien Lifehouse, Sydney, NSW, Australia. ([email protected]), ([email protected]), ([email protected] [Presenting author]) 2

Introduction Dose enhancement using GNPs is a current hot-topic in radiotherapy1,2. The strong interaction of the beam with gold, results in an expected dose enhancement. A significant obstacle to the clinical implementation of GNPs is that Monte Carlo predicted dose enhancements4 are not reflected in cell survival5. An assumption commonly made in simulations is that GNPs are sufficiently dispersed to avoid dosimetric interactions1,3. However, imaging of GNPs in cells confirms a range of heterogeneous distributions. The aim of this study is to determine whether clustering of GNPs is the cause of the discrepancy between theory and observation. Method Simulations were performed using the PENELOPE Monte Carlo (MC) platform (ver. 2011). GNPs were modelled as spheres of gold of radius rp and were irradiated with a uniform, parallel, monoenergetic beam of photons of energy E0. A second identical particle GNPb, was placed at a distance s from GNPa (centre to centre, Fig. 1). Clusters of 18 and 36 GNPs were also simulated. Results For a single GNPa, the shape of the curve representing dose to water as a function of radial distance does not depend on energy in the range 20keV to 10MeV or the size of the particle. When a single neighbouring GNPb is introduced, the dose enhancement caused by GNPa is not greatly affected. However, when a tight cluster of 36

GNPs is introduced, the fraction of energy it removes reduces the dose enhancement by 20%. If the GNP cluster is not tightly packed, the dose enhancement returns or may even increase by up to 4%. Conclusion This study has shown that tight, but not disperse clusters of GNPs can reduce radiation dose enhancements, resolving the apparent discrepancy reported in the literature. References 1. R. I. Berbeco, H. Korideck, W. Ngwa, R. Kumar, J. Patel, S. Sridhar, S. Johnson, B. D. Price, A. Kimmelman, and G. M. Makrigiorgos, ‘‘DNA Damage Enhancement from Gold Nanoparticles for Clinical MV Photon Beams,’’ Radiat Res (2012). 2. J. Schuemann, R. Berbeco, D. Chithrani, S. Cho, R. Kumar, S. McMahon, S. Sridhar, and S. Krishnan, ‘‘Roadmap to clinical use of gold nanoparticles for radiosensitization,’’ International Journal of Radiation Oncology* Biology* Physics (2015). 3. E. Lechtman, N. Chattopadhyay, Z. Cai, S. Mashouf, R. Reilly, and J. P. Pignol, ‘‘Implications on clinical scenario of gold nanoparticle radiosensitization in regards to photon energy, nanoparticle size, concentration and location,’’ Phys Med Biol 56, 4631–47 (2011). 4. B. L. Jones, S. Krishnan, and S. H. Cho, ‘‘Estimation of microscopic dose enhancement factor around gold nanoparticles by Monte Carlo calculations,’’ Med Phys 37, 3809–16 (2010) 5. S. Jain, J. A. Coulter, A. R. Hounsell, K. T. Butterworth, S. J. McMahon, W. B. Hyland, M. F. Muir, G. R. Dickson, K. M. Prise, F. J. Currell, J. M. O’Sullivan, and D. G. Hirst, ‘‘Cellspecific radiosensitization by gold nanoparticles at megavoltage radiation energies,’’ Int J Radiat Oncol Biol Phys 79, 531–9 (2011).

O094 An overview of Ta2O5 ceramic nanostructured particles in dose enhancement radiotherapy R. S. Brown1, Dr. M Tehei1,3, S Oktaria1,3, Dr. K. Konstantinov3,4, Prof. A Rosenfeld1,3, Dr. S Corde1, 5, Assoc. Prof. M. Lerch1,3 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. 2St George Cancer Care Centre, St George Hospital, Kogarah, Australia. 3Illawarra Health and Medical Research Institute, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 4Institute for Superconducting and Electronic Materials, University of Wollongong, Australia. ([email protected]), ([email protected]). 5Radiation Oncology Department, Prince of Wales Hospital, Randwick, Australia. ([email protected]), ([email protected])

Fig. 1 Example geometry showing two interacting GNPs in a scoring volume. The beam is incident on GNPa only

123

Introduction Dose enhancement radiotherapy (DERT) involves the introduction of high-Z atoms into close proximity to the tumour, providing localised enhancement following exposure to the local radiation field through the production of charged particles and reactive oxygen species (ROS) [1]. Although gold nanoparticles (Au NPs) have dominated successful research in DERT [2], ceramic NPs are emerging as potential alternatives. In this study, tantalum pentoxide nanostructured particles (Ta2O5 NSPs) are investigated for their application as novel non-toxic radiosensitisers in DERT [3].

Australas Phys Eng Sci Med Optimisation of key therapeutic parameters is explored, with the aim of maximising treatment efficacy on radioresistant 9L gliosarcoma cells. Methods

Acknowledgements The authors acknowledge the financial support from the National Health and Medical Research Council (APP1084994). We acknowledge the work done by Dr. David Wexler in generating HRTEM images.

X-ray Diffraction (XRD) and High-Resolution Transmission Electron Microscopy (HRTEM) Flow cytometry (internalisation assessment) and cytotoxicity assessments Cellular irradiation and sensitisation analyses Confocal microscopy

O095 Optimizing the treatment efficiency of microbeam radiation therapy with high-Z nano-structured ceramic particles

Results Variations in production methodology allowed Ta2O5 NSPs to be produced with smaller crystallite sizes and less tendency for aggregation than the pilot batch. Ta2O5 NSPs were internalised, up to 300% at 500 lg/ml, and displayed minimal toxicity of 30% at this maximum concentration to 9L cells.

E. Engels1, M. Lerch1,2, S. Guatelli1,2, S. McKinnon1, N. Li1, K. Konstantinov2,3, A. Rosenfeld1,2, M. Tehei1,2, S. Corde1,4 1

Centre for Medical Radiation Physics (CMRP), University of Wollongong, NSW, Australia. ([email protected] [Presenting author]). 2Illawarra Health and Medical Research Institute (IHMRI), University of Wollongong, NSW, Australia. ([email protected]), ([email protected]), ([email protected]). ([email protected]). 3Institute for Superconducting and Electronic Materials, University of Wollongong, Australia. ([email protected]), ([email protected]), ([email protected]). 4Radiation Oncology Department, Prince of Wales Hospital, Randwick, NSW, Australia. ([email protected])

Beam energy-dependent sensitisation was observed, however effective sensitisation was achieved with 10 MV X-ray photon fields, evidenced by an increase in the alpha parameter, an indicator of cellular radiosensitisation. The highest enhancement recorded was with irradiation of 10 MV photons at 500 lg/ml, giving an SER of 1.46. Interestingly, no significant change in sensitisation was observed with increasing NSP concentration. Conclusions This study highlights Ta2O5 NSPs as effective radiosensitisers and viable non-toxic alternatives to NPs, such as Au, currently used in DERT. Several factors influencing radiosensitisation have been explored, and optimised to maximise therapeutic effect with ceramic Ta2O5 NSPs on radioresistant 9L cells. Significant sensitisation was observed at a beam energy of 10 MV, with a maximum SER of 1.46. References 1. Corde S, Joubert A, Adam JF, Charvet AM, Le Bas JF, Esteve F, Elleaume H, Balosso J (2004) Synchrotron radiation-based experimental determination of the optimal energy for cell radiotoxicity enhancement following photoelectric effect on stable iodinated compounds. British Journal of Cancer 91:544–551 2. Jain S, Hirst DG, O’Sullivan JM (2012) Gold nanoparticles as novel agents for cancer therapy. The British Journal of Radiology 85:101–113 3. Brown R, Tehei M, Oktaria S, Briggs A, Stewart C, Konstantinov K, Rosenfeld A, Corde S, Lerch M (2013) High-Z nanostructured ceramics in radiotherapy: First evidence of Ta2O5-induced dose enhancement on radioresistant cancer cells in an MV photon field. Particle and Particle Systems Characterisation 31:500–505

Introduction Microbeam radiation therapy (MRT) implements spatially-fractionated kilovoltage x-rays for deep seated tumour treatment [1,2] with normal tissue sparing [3]. However, due to the large dose gradient from the peak (in-beam) to the valley (between microbeams), tumour treatment is not optimized. High-Z nanoparticles (NPs) enhance the dose delivered by conventional radiotherapies [4]. Nano-structured ceramic Ta2O5 NPs are non-toxic [4] and show optimal x-ray absorption in kilovoltage energies [5,6]. This research assesses the ability of these NPs to selectively raise the tumour valley dose in MRT. Method Geant4 [7,8] simulations investigated the physical dose enhancement of Ta2O5 NPs to a population of cells (Fig. 1) due to monoenergetic broadbeams and microbeams (50–200 keV). Simulation results were correlated to in-vitro experiments in hutch 2B at the Imaging and Medical Beamline (IMBL), Australian Synchrotron, using tumorous 9L gliosarcoma and healthy Madin Darby Canine Kidney cells. Ta2O5 NPs were added to cells in T12.5cm2 flasks 24 h before 90–100% confluence. Cells were irradiated using a 1.4T wiggler field to produce 50/400lm microbeams, and survival determined with clonogenic assays. Results Ta2O5 NPs improve the MRT and broad-beam selectivity towards tumour cells, due to the clustering NP distribution observed in tumour cells (Fig. 2). NPs surrounding tumour cell nuclei, rather

Fig. 1 In-vitro cell population model with NP cluster modelled in red and incident 50 lm microbeam outlined in blue

123

Australas Phys Eng Sci Med 8. Allison J, et al. (2006) Geant4 developments and applications, IEEE Trans. Nucl. Sci. 53, 270–278. Acknowledgements The authors thank the University of Wollongong Information Technology Services (ITS) for computing time on the UOW High Performance Computing Cluster, and acknowledge the financial support from the National Health and Medical Research Council (APP1084994) and Australian Synchrotron (AS161/IM/ 10479).

Fig. 2 Comparison of a MRT treatment (42 keV, 50 lm microbeams with 0.4 Gy valley dose) of tumorous 9L gliosarcoma and normal MDCK with and without Ta2O5 NPs, shown to be distributed differently in each cell population than homogeneously distributed NPs as in normal cells, was confirmed to produce greater dose enhancement experimentally and with Geant4 simulations. Modelling micro- and broad-beams showed that NP dose enhancement occurs and is energy dependent. 40 keV X-rays are optimum for Ta2O5 NPs in broad-beam cases, and microbeam energies greater than 100 keV are optimal so as to produce secondaries that raise the valley dose ([100 lm from the microbeam). Conclusion The multi-modal approach: Synchrotron Microbeam Activated Radiation Therapy (SMART) with NPs, improves the treatment tumour selectivity. The MRT (and broad beam) treatment efficiency is NP distribution and beam energy dependent, as confirmed with experimental and simulation studies. References 1. Slatkin DN, Spanne P, Dilmanian FA, Sandborg M (1992) Microbeam Radiation Therapy, Med. Phys. 19:1395–1400. 2. Crosbie JC, Anderson RL, Rothkamm K, Restall CM, Cann L, Ruwanpura S, Meachem S, Yagi N, Svalbe I, Lewis RA, Williams BR, Rogers PA (2010) Tumor cell response to synchrotron microbeam radiation therapy differs markedly from cells in normal tissues. Int. Jour. Rad. Onc. Bio. Phys. 77:886–894. 3. Laissue JA, Blattmann H, Wagner HP, Grotzer MA and Slatkin DN (2007) Prospects for microbeam radiation therapy of brain tumours in children to reduce neurological sequelae. Dev. Med. Child Neurol. 49:571–581. 4. Brown R, Tehei M, Oktaria S, Briggs A, Stewart C, Konstantinov K, Rosenfeld A, Corde S, and Lerch M (2014), High-Z Nanostructured Ceramics in Radiotherapy: First Evidence of Ta2O5-Induced Dose Enhancement on Radioresistant Cancer Cells in an MV Photon Field. Part. Part. Syst. Charact. 31: 500–505. 5. Hubbell JH (1982) Photon mass attenuation and energy-absorption coefficients from 1 keV to 20 MeV. Int. Journal of Applied Radiation and Isotopes, 33:1269–1290. 6. Seltzer SM (1993) Calculation of photon mass energy-transfer and mass energy-absorption coefficients. Radiation Research, 136: 147–170. 7. Agostinelli S et al. (2003) Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. Nucl. Instr. Meth. Phys. Res. A 506:250–303.

123

O096 Anatase titanium dioxide nanoparticles (TiO2 NPs) as a radiosensitazation agent for radiation therapy: In vitro and phantom based studies Moshi Geso1, Esho Youkhana1, Bryce Feltis2 1

Medical Radiation Discipline, School of Health and Biomedical Sciences, RMIT University, Victoria, Australia. ([email protected]), ([email protected] [Presenting author]). 2Pharmaceutical Sciences Discipline, School of Health and Biomedical Sciences, RMIT University, Victoria, Australia. ([email protected])

Introduction High atomic number (Z) materials (gold) nanoparticles ‘NP’ have been shown to enhance the radiation effects. However, very few investigations are documented about low Z materials where Compton scattering is the dominant interaction process with ionising radiation across almost the entire energy ranges.TiO2 have the ability to generate reactive oxygen species (ROS) beside the free radicals. This work investigates radiation dose enhancement caused by TiO2NPs covering entire X-ray energy ranges. TiO2NPs can be densely included in the tumour cells due to its low nontoxic. This study is based on phantom and cells in culture investigations. Methods Anatase-TiO2NPs of 30 nm average size were synthesised and dispersed in culture-medium and halocarbons (PRESAGE chemical composition). These NPs were characterised by XRD, XPS, TEM, FTIR and TGA. Two types of cell lines, Human Keratinocyte (HaCaT) and prostate (DU145) cancer cell lines are used. Also PRESAGE dosimeter/phantom is used. Kilovoltage and megavoltage x-ray beams were used separately. The PRESAGE dosimeters were scanned using UV/VIS spectrophotometer and optical CT scanner. MTS and Clonogenic assays were employed for cell viability and cytotoxicity measurements. Results TiO2NPs significantly enhances the dose by about (40%, 54%) at 80 kV x-rays as measured using both PRESAGE and cells in culture respectively. However, using PRESAGE with megavoltage beams shows insignificant dose enhancement and under same conditions about 50% dose enhancement is detected with cells. This difference can be attributed to some bio-chemical effects (generation of reactive oxygen species (ROS)), since such species don’t affect PRESAGE. Inclusion of TiO2NPs in cells-medium generate hydroxyl radicals (•OH) which is quantified in this study. Conclusion TiO2 NPs causes a significant dose enhancement with megavoltage beams in vitro study. The enhancement can be attributed to higher levels of ROS generated. This research proves potential value for more efficient beam delivery since MV beams are most commonly used in radiotherapy.

Australas Phys Eng Sci Med require the use of fluorescent probes.[1–3]. These methods show only terminal end-points and are very labour intensive at a high cost. We present the real-time imaging of cell proliferation and cellular responses under different irradiation conditions, using an IncuCyteTM (Essen Bioscience). This method enables us to monitor live cells before and after radiation for a desired period of time, which can be applied to any cell type and to both monolayer and tumour spheroid models. Method DLD-1 colorectal cancer cells were seeded in a 96 well plate, 400 cells/well, in 2% FCS supplemented advance DMEM, in a 37C and 5% CO2 incubator. Cells were subjected to different treatments with cisplatin, 5-FU, gold nanoparticles (GNPs), or a combination of cisplatin/5-FU with GNPs. After further 24 h incubation, the plate was irradiated at 2, 3 and 4 Gy on a Varian Novalix Tx MeV linac, then monitored in an IncuCyte chamber for 120 hours. Synergies between the drugs and the GNPs were investigated by comparing the results of the combination with the individual treatments. Results Cell growth inhibition was observed in the first 15–20 hours post irradiation. There was synergy between the anti-cancer drug and the GNPs significantly enhancing radiation effectiveness, especially at 2 Gy dose. 30

Control_ 2 Gy CisPt_ 2 Gy 5FU_2 Gy GNPs_2 Gy GNPs + CisPt_2 Gy GNPs + 5FU_2 Gy

Normalized confluency

25 20 15 10 5 0 0

20

40

60

80

100

120

Time (h)

O097 Real-time monitoring of the cellular response to radiation therapy: Influence of nanoparticles and anti-cancer drug as radiation enhancers Binh T. T. Pham1, Nguyen T. H. Pham1, Byung J. Kim1, Brian S. Hawkett1, David R. McKenzie2, Linda Rogers3, Joanne Toohey3, Natalka Suchowerska3

Conclusion Incucyte is a very effective and robust method to study cell responses after treatments including chemoradiation therapy. This method provides unique information to oncologists about fractionation schedules for treatment. Acknowledgements Authors would like to thank the financial support from Sirtex Medical Ltd.; technical instrumentations and scientific supports from the Radiation Oncology department, Chris O’Brien Lifehouse.

1

References

Introduction Radiation therapy chemotherapy and surgery are commonly used to treat cancer. The clonogenic assay is commonly used to study the effects of radiation and the mechanism for cell death.[1] This assay is highly dependent on the cell type and cell seeding density. Other assays to investigate the biological effects by cell apoptosis or generations of reactive oxygen species (ROS),

1. Retif P, Pinel S, Toussaint M, Frochot C, Chouikrat R, Bastogne T, et al. Nanoparticles for Radiation Therapy Enhancement: the Key Parameters. Theranostics. 2015;5:1030–45. 2. Kato Y, Yashiro M, Fuyuhiro Y, Kashiwagi S, Matsuoka J, Hirakawa T, et al. Effects of Acute and Chronic Hypoxia on the Radiosensitivity of Gastric and Esophageal Cancer Cells. Anticancer Research. 2011;31:3369–75. 3. Rogers LJ, Suchowerska N, Khan A, Polaskova V, McKenzie DR. Profiling of the secretome of human cancer cells: Preparation of supernatant for proteomic analysis. Electrophoresis. 2014;35:2626–33.

Key Centre for Polymers and Colloids, School of Chemistry, University of Sydney NSW 2006, Australia. ([email protected] [Presenting author]). 2School of Physics, University of Sydney NSW 2006, Australia. 3Chris O’Brien Lifehouse, Sydney NSW 2050, Australia

123

Australas Phys Eng Sci Med

O098 Proliferative response of head and neck cancer stem cells under X-rays P. Reid1, P. Wilson2,3, Y. Li1, A. Staudacher4,5, E. Bezak1,5 1

International Centre for Allied Health Evidence and Sansom Institute for Health Research, University of South Australia, Adelaide, Australia. ([email protected] [Presenting author]). 2 School of Engineering, University of South Australia, Adelaide, Australia. 3Department of Medical Physics, Royal Adelaide Hospital, Adelaide, Australia. 4Translational Oncology Laboratory, Centre for Cancer Biology, SA Pathology, and University of South Australia, Adelaide, Australia. 5School of Medicine, University of Adelaide, Adelaide, Australia Introduction Evidence exists of cancer stem cells (CSC), a subpopulation in cancers that can proliferate indefinitely, create all tumour cell types and self-renew their own phenotype. CSCs account for treatment resistance, failure and metastasis [1]. Identification and targeting of CSCs in head and neck cancer (HNC) is crucial to improving survival rates. Limited data exists on the amount and behaviour of HNC CSCs under irradiation. Additionally, radiotherapy stimulates CSCs, meaning the fraction of CSCs during treatment is likely to increase [2,3]. We aim to determine the contribution of CSCs to tumour growth and radioresistance following irradiation of 2 HNC cell lines. Method Two HNC cell lines (UM-SCC-1 and UM-SCC-47) were irradiated with 4 Gy dose (600 cGy/min) using a 6 MV X-ray beam from a Varian 600CD linac (Varian Medical System, Palo Alto, CA) at the Royal Adelaide Hospital. T75 cell flasks were encased in solid water (RW3) to achieve electronic equilibrium at the cell layer. Sham-irradiated flasks were used as controls. The linac was calibrated using IAEA TRS398 calibration protocol. At 24, 48 and 72 hours after irradiation, cells were examined by flow cytometry using the putative CSC markers of aldehyde dehydrogenase (ALDH) and CD44 on the viable (7-AAD-negative) cell population. Results Both UM-SCC-1 and UM-SCC-47 showed increases in proportions of CD44+/ALDH+ cells after irradiation. For UMC-SCC1 1st generation, the percentage of positive cells was maximal at 24

Fig. 1 Comparative populations of cells positive for putative CSC markers ALDH & CD44 in HNC cell lines UM-SCC-1 and UM-SCC47. The white bars represent the untreated 1st generation cells, grey bars represent 1st generation cells irradiated with 4Gy. Surviving cells from irradiated samples of each cell line were reseeded, grown and irradiated with 4Gy (2nd generation) and are represented by black bars. Cells positive for both ALDH & CD44 show significant increases against the non-irradiated controls though the character of these increases differs between cell lines

123

hours and decreased at the following time-points. UM-SCC-47 increased CD44+/ALDH+ cells at 48 and 72 hours after irradiation (Fig. 1). Conclusion X-radiation of UM-SCC-1 and UM-SCC-47 showed an increasing proportion of putative CSCs in the population. This may be a result of enhanced radioresistance of this cell population. Clonogenic assays are required to determine survival curves and to determine changes in radiation sensitivity in the cell population following subsequent exposures. References 1. Prince M, Sivanandan R, Kaczorowski A, Wolf G, Kaplan M, Dalerba P, Weissman I, Clarke M, Ailles, L (2007) Identification of a subpopulation of cells with cancer stem cell properties in head and neck squamous cell carcinoma. Proceedings of the National Academy of Sciences104: 973–978 2. Lagadec C, Vlashi E, Della Donna L, Dekmezian C, Pajonk, F (2012) Radiation-Induced Reprogramming of Breast Cancer Cells. Stem Cells 30: 833–844. 3. Kim MJ, Kim RK, Yoon CH, An S, Hwang SG, Suh Y, Park MJ, Chung HY, Kim IG, Lee, SJ (2011) Importance of PKCdelta signaling in fractionated-radiation-induced expansion of gliomainitiating cells and resistance to cancer treatment. J Cell Sci 124: 3084–3094.

KS12 Benchmarking and Optimising CT in Scotland – an exercise in patience D. G. Sutton Medical Physics Dept, Ninewells Hospital & Medical Scholl, Dundee, DD19SY, UK. ([email protected] [Presenting author]) Outline We have previously reported on a dose audit of CT examinations at different scanner sites across Scotland[1].The results called into question the validity of DRLs and suggested that they may have become dose limits and that a more refined approach to optimisation is needed [2]. This project lead directly to two collaborative initiatives from the five teaching hospitals in Scotland, with funding from the Scottish Government. The first, entirely within our control, involved the purchase of 3 ATOM CIRS paediatric phantoms (representative of 1, 5 and 10 year old patients) to be subject to head and chest CT examinations on 21 CT scanners across Scotland. The aim of the project is to benchmark current practice in paediatric CT, identify the variation in patient dose and image quality, and embark on a nationwide optimisation exercise. Significant examination findings included inconsistent selection of exam kV and poorly set up mA-modulation systems, especially with respect to the choice of minimum examination mA. Existing manufacturer defaults could be doing more harm than good. The second project relies on the fact that Scotland is uniquely placed in as much as it has a) a national PACS infrastructure supplied by the same vendor and b) every inhabitant has a unique health identifier. This makes large-scale patient studies conceptually easy to implement and manage. The initial stage of the project was centred around a nationwide dose analysis performed across a range of examination types using software from a commercial provider mining data stored on the PACS. A major aim of the study was to evaluate the utility performance and impact of installing such a system. Implementation of this project was not within our control. The presentation will

Australas Phys Eng Sci Med and intervention: initial results from the establishment of a multicentre diagnostic reference level in Queensland public hospitals. J Med Radiat Sci. 61(3):135–41. doi: 10.1002/jmrs.67.

review the outcomes of the paediatric project, and discuss the issues identified and outcomes of the CT dose mining work. Acknowledgements Medical Physicists from the 5 major Scottish centres. References

O100 Diagnostic reference levels for digital mammography; time for a new paradigm

1. Sutton DG et al CT chest abdomen pelvis doses in Scotland: has the DRL had its day? Br J Radiol. 2014 Sep;87(1041):20140157 2. Rehani MM 2015 Limitations of diagnostic reference level (DRL) and introduction of acceptable quality dose (AQD).Br J Radiol Jan;88(1045):20140344

Moayyad E Suleiman1, Jennifer Diffey2, Lucy Cartwright3, Mark F McEntee1 1

O099 Results of the 2015 ARPANSA image-guided interventional procedures DRL survey P. D. Thomas, T. Beveridge, A. Wallace Medical Imaging Section, ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction The ARPANSA National Diagnostic Reference Level Service (NDRLS) collected data on Imaged-Guided Interventional Procedures (IGIP) from October 2014 to December 2015. Data were requested for normal coronary angiogram, endoscopic retrograde cholangiopancreatography (ERCP), cerebral angiography and selective abdominal angiography. The aim was to collect data to assist in establishing diagnostic reference levels for these procedures. Method Patient data, dose metrics and procedure information were requested for a sample of 30 patients for each procedure surveyed. Patient characteristics included age, sex and weight. Dose metrics included total dose-area product (DAP) for the procedure and cumulative dose at the system reference point (Ref Dose). Other data collected included fluoroscopy time, frame rate, digital acquisition time, total number of digital frames and access point. Facility reference levels (FRLs) for both DAP (in Gy.cm2) and Ref Dose (in Gy) were calculated for all surveys with at least 20 patients. The median values of the dose indices were used as the FRLs. Results A total of 47 surveys were received. Thirty-one surveys were received for the normal coronary angiogram procedure. The 75th percentile of the distribution of FRLs was 34.9 Gy.cm2 for DAP and 0.55 Gy for Ref Dose; consistent with data reported by Erskine et al [1] and by Crowhust et al [2]. Eleven surveys were received for cerebral angiography; however these were spread across complexity levels of 1–3 vessels and 4 or more vessels and with or without rotational angiography. Five surveys were received for the ERCP procedure. No abdominal surveys were received. Conclusion Sufficient data was received to propose a DRL for a normal coronary angiogram. The data for other procedures are not sufficient to propose DRLs. Further feedback will be sought to consider which procedures are most suited to the establishment of DRLs and to refine procedure descriptions. References 1. Erskine BJ, Brady Z, Marshall EM (2014) Local diagnostic reference levels for angiographic and fluoroscopic procedures: Australian practice. Australas Phys Eng Sci Med. 37(1):75–82. doi: 10.1007/s13246-014-0244-2. 2. Crowhurst JA, Whitby M, Thiele D, Halligan T, Westerink A, Crown S, Milne J (2014) Radiation dose in coronary angiography

The University of Sydney, Faculty of Health Sciences, M205, Cumberland Campus, 75 East St, Lidcombe, NSW 2141, Australia. ([email protected] [Presenting author]), ([email protected]). 2Hunter New England Imaging John Hunter Hospital, Lookout Rd, New Lambton Heights, NSW 2305, Australia. ([email protected]). 3Westmead Hospital, Hawkesbury Road, Westmead, NSW 2145, Australia. ([email protected]), ([email protected]) Introduction Diagnostic reference levels (DRLs) provide a measure of quality control to reduce variations in dose among and within imaging centres. The International Commission of Radiation Protection defines DRLs as: ‘‘A form of investigation level, applied to an easily measured quantity, usually the absorbed dose in air, or tissue-equivalent material, at the surface of a simple phantom or a representative patient [1].’’ Although Mean Glandular Dose (MGD) has been established as the ‘‘easily measured quantity’’ [2–10] wide variation exists in how it is calculated. Furthermore the ‘‘representative patient’’ assumes a homogeneity that does not exist. The aim of this work is establish DRLs for NSW with a focus on stratifying DRLs according to breast thickness range. Method Anonymised mammograms were downloaded from the PACS of the Cancer Institute NSW, Information from 53,405 mammograms DICOM headers were extracted using third party software. The data were filtered to include only headers with complete information required for the calculation of MGD, for women with no breast implants, and breast thicknesses ranging from 20 to 110 mm. Data from 45,054 mammograms representing 61 BreastScreen mammography centres and mobile units were imported to an in house developed Microsoft Excel workbook to calculate MGD using the methods of Dance et al. [5–7].

Table 1 The 75th and 95th percentiles for mean glandular dose based on 45,054 digital mammograms from 61 BreastScreen centres in NSW/Australia Breast thickness range (mm)

75th percentile

95th percentile

25 ± 5

0.98

1.25

35 ± 5

1.15

1.58

45 ± 5 55 ± 5

1.35 1.74

1.89 2.49

65 ± 5

2.46

3.05

75 ± 5

2.36

4.41

85 ± 5

2.69

5.71

95 ± 5

2.97

7.81

105 ± 5

3.15

6.01

123

Australas Phys Eng Sci Med Results Mean breast thickness was 58mm; mean MGD was1.51 mGy (range 0.19–10.00 mGy). Table 1 shows the 75th and 95th percentiles for nine breast thickness ranges. The DRLs at 25, 45, 65 and 85 mm constitute non-overlapping ranges and effectively different sample populations. Conclusion DRLs for digital mammography based on breast thickness were proposed (75th percentiles) for the first time in Australia. DRLs varied for different ranges of breast thicknesses, which provided important insight into the status of mammography in NSW and supports the need for the stratification of DRLs according to breast thicknesses. References 1. ICRP (1996) Radiological protection and safety in medicine. ICRP Publication 73 ICRP Publication 73. Ann. ICRP 26 (2). (2):1–47 2. Boone JM (1999) Glandular breast dose for monoenergetic and high-energy X-ray beams: Monte Carlo assessment. Radiology 213 (1):23–37. doi:10.1148/radiology.213.1.r99oc3923 3. Boone JM (2002) Normalized glandular dose (DgN) coefficients for arbitrary X-ray spectra in mammography: computer-fit values of Monte Carlo derived data. Med Phys 29 (5):869–875 4. Boone JM, Fewell TR, Jennings RJ (1997) Molybdenum, rhodium, and tungsten anode spectral models using interpolating polynomials with application to mammography. Med Phys 24 (12):1863–1874 5. Dance D (1990) Monte-Carlo calculation of conversion factors for the estimation of mean glandular breast dose. Phys Med Biol 35 (9):1211 6. Dance DR, Skinner CL, Young KC, Beckett JR, Kotre CJ (2000) Additional factors for the estimation of mean glandular breast dose using the UK mammography dosimetry protocol. Phys Med Biol 45 (11):3225–3240 7. Dance DR, Young KC, van Engen RE (2009) Further factors for the estimation of mean glandular dose using the United Kingdom, European and IAEA breast dosimetry protocols. Phys Med Biol 54 (14):4361–4372. doi:10.1088/0031-9155/ 54/14/002 8. Sobol WT, Wu X (1997) Parametrization of mammography normalized average glandular dose tables. Med Phys 24 (4):547–554 9. Wu X, Barnes GT, Tucker DM (1991) Spectral dependence of glandular tissue dose in screen-film mammography. Radiology 179 (1):143–148 10. Wu X, Gingold EL, Barnes GT, Tucker DM (1994) Normalized average glandular dose in molybdenum target-rhodium filter and rhodium target-rhodium filter mammography. Radiology 193 (1):83–89. doi:10.1148/radiology.193.1.8090926

O101 More than meets the CTDI: A review of CT dose audit methods Jennifer Diffey, Brett Roworth Hunter New England Imaging, John Hunter Hospital, New Lambton, NSW. ([email protected] [Presenting author]), ([email protected])

123

Introduction Diagnostic Reference Levels (DRLs) are a valuable tool for optimisation of radiation protection. Since the publication of ARPANSA DRLs for MDCT1, we have conducted multiple dose audits for 9 CT scanners. Data were collected manually and also extracted from RIS. Merits and limitations of each approach are discussed and future strategies are proposed. Method Manual data collection was carried out in accordance with ARPANSA guidance. Data on CTDIvol and DLP were requested for 20 patients of ‘‘standard’’ size. DLP data for 150,000 CT examinations were extracted from RIS. RIS procedure codes were matched to the six ARPANSA-defined ‘‘common protocols’’1 which had an established DRL. Results Manual data collection took 1–12 months. Although median DLP at most sites was below the ARPANSA DRL, the quality of data was variable. CTDIvol for multi-phase studies was often recorded incorrectly and there was a huge spread in data, even intra-site. RIS data extraction for 150,000 CT examinations was completed overnight. Matching the RIS codes to the ARPANSA protocols required specialist clinical knowledge; for example, Neck Non-Contrast met the ARPANSA criteria for CT Neck, but Neck IV-Contrast did not. Limitations of RIS data are that only DLP is recorded, data entry is manual and RIS codes may not represent the scan performed so should be reviewed in conjunction with the request form. Although taking the median of large data samples eliminates the effect of outliers, ‘‘common protocols’’ may be too broad to achieve meaningful optimisation. For example, AbdoPelvis will encompass a range of examinations for which DLP can vary by a factor of 10. Conclusion Despite initial perceptions that dose audit is straightforward, complexities were encountered. Collaboration between medical physicists, radiographers and radiologists is essential. Prospective data collection using well-defined selection criteria may enable more meaningful optimisation than retrospective analysis of large samples using broad criteria. Acknowledgements We are very grateful to Dr Christian Abel and Keira Gordon for valuable advice on dose audit strategies. Many thanks to Steve Beautement for data extraction from RIS and to all radiographers who assisted with data collection. References 1. http://www.arpansa.gov.au/services/ndrl/adult.cfm. Accessed 17 June 2016

O102 Five-year review of the Australian National Diagnostic Reference Levels for CT P. D. Thomas, M. Sanagou, A. Wallace Medical Imaging Section, ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction The ARPANSA National Diagnostic Reference Level Service (NDRLS) commenced collection of patient dose metrics in computed tomography in 2011. Australian national diagnostic reference levels (DRLs) for CT were promulgated in 2012 based on the data collected in 2011 [1]. DRLs should be reviewed on a periodic basis to ensure that they reflect current practice. There are now five years of data from which to examine trends in patient dose for CT and to review the national DRLs.

Australas Phys Eng Sci Med Method Facilities submit dose (CTDIvol and DLP) and patient (age, sex and weight) information for up to 20 patients for each body region and CT scanner, along with protocol technique factors. Facility Reference Levels (FRLs) are determined for both dose metrics using the median values for each survey, provided at least 10 patients were included. Data are categorised by body region, number of scan phases, use of iterative reconstruction, and survey year. Information on the use of iterative reconstruction was only available after April 2013. Results Third quartiles of the FRL distributions for DLP (in mGy.cm) are shown in the table below, categorised by body region, number of scan phases, use of iterative reconstruction and survey year. Also shown are the current DRL and the third quartile for 2015 rounded to the nearest multiple of ten. The reduction in dose with iterative reconstruction reported previously [2] has been maintained in more recent data.

Third Quartile of FRLs for DLP (mGy.cm) Region

Phases

Iterative Recon

Survey Year

DRL

Rounded

O103 RBE studies in particle therapy with solid state microdosimetry L. T. Tran1, L. Chartier1, D. Prokopovich2, D. Bolst1, S. Guatelli1, A. Pogossov1, M Petasecca1, M. Lerch1, M. Reinhard2, A. Kok3, M. Povoli3, M. Jackson4, T. Kanai5, N. Matsufuji6, A. B. Rosenfeld1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]). 2ANSTO, Australia. ([email protected],au), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3 SINTEF, Norway. ([email protected]), ([email protected]). 4Prince of Wales Hospital, Australia. ([email protected]). 5Gunma University, Japan. ([email protected]). QST, Japan. ([email protected]), ([email protected])

2015 2011

2012

696

714

2013

2014

2015

AbdoPelvis

1

Unknown

AbdoPelvis

1

No

926

640

649

AbdoPelvis

1

Yes

584

531

525

Chest

1

Unknown

Chest

1

No

458

474

460

Chest

1

Yes

378

365

351

Head

1

Unknown

Head

1

No

1158

1071

992

990

Head

1

Yes

901

872

829

830

Neck

1

Unknown

Neck

1

No

577

827

557

Neck

1

Yes

468

445

411

L-Spine

1

Unknown

L-Spine

1

No

946

838

840

L-Spine

1

Yes

693

691

608

CAP

1

Unknown

CAP

1

No

970

851

967

CAP

1

Yes

666

624

702

CAP

2

Unknown

CAP

2

No

1154

1131

1023

1020

CAP

2

Yes

1010

983

935

940

458

970

582

885

980

1181

495

970

573

851

974

1115

692

700

461

650 520 450

1064

460 350 1000

478

600

856

560 410 900

602

840 610 1200

1190

970 700 1200

Conclusion The third quartile of the distribution of the FRLs has in general remained consistent with the DRL for scans undertaken without the use of iterative reconstruction. A slight reduction has been observed for some scans. Doses are markedly lower in all body regions for scans conducted using iterative reconstruction. The data suggest that separate DRLs could be proposed for scans with and without the use of iterative reconstruction. References 1. Hayton A, Wallace A, Marks P, Edmonds K, Tingey D, Johnston P (2013) Australian diagnostic reference levels for multi detector computed tomography. Australas Phys Eng Sci 36(1):19–26. 2. Thomas P, Hayton A, Beveridge T, Marks P, Wallace A (2015) Evidence of dose saving in routine CT practice using iterative reconstruction derived from a national diagnostic reference level survey. Br J Radiol. 88(1053):20150380. doi:10.1259/bjr. 20150380.

Introduction In proton and 12C ion therapy, the determination of relative biological effectiveness (RBE) is crucial as a parameter in the patient treatment planning to predict physical dose delivery. The MicroPlus probe utilizing new generation of solid state microdosimeter ‘‘Mushroom’’ with truly 3D sensitive volumes (SV) modelling biological cells was developed for RBE studies. Method Stationary and movable with lung temporary pattern MicroPlus probe was used in a water phantom for RBE studies during passive and PBS delivery at F.Burr Proton Therapy Centre, USA; the HIMAC facility using passive scattering delivery with 290MeV/u 12C, 400MeV/ u 16O and 180MeV/u 14N ions; and at Gunma University HIT facility in Japan using a 290MeV/u 12C ions high intensity pencil beam with 3.3 mm and beam intensity as high as 106 particles/spill. Results Dose average lineal energy and derived RBE10 values obtained with the MicroPlus probe matched well with those obtained with the TEPC in passive beam delivery while allowed submillimetre spatial resolution [1, 2]. RBE in proton therapy delivered with passive and scanning beam is increasing essentially in the last 8 mm of the proton range and reaching 1.8 in a distal part of the SOBP. The maximum RBE10 in 12C ion therapy PBS and passive delivery was 2.5 while for 400MeV/u 16O and 180MeV/u 14N ions was about 3.2. It was demonstrated that RBE10 derived by stationary and movable MicroPlus probe can be essentially different up to 30%. Conclusion High spatial resolution easy use MicroPlus probe utilizing 3D solid state microdosimetry detectors ‘‘Mushroom’’ provide new domain of quality assurance in particle therapy important for many clinical studies which were impossible with Tissue Equivalent Proportional Counter. Studies demonstrated that RBE in PT and RBE10 in 12C ion therapy in the proximity of the ion range are higher that accepted and should be considered during the planning as well as RBE variation with target motion. References 1. Kase Y, et al. Microdosimetric measurements and estimation of human cell survival for heavy-ion beams. Radiation Res. 166:629–38, 2006 2. L. T. Tran, et al. 3D Silicon Microdosimetry and RBE study using 12C ion of different energies. IEEE Trans. on Nucl. Sci. vol. 62, no. 6, pp 3027–3033, 2015.

123

Australas Phys Eng Sci Med

O104 Determining the dose-equivalence of synchrotron microbeam radiotherapy and conventional radiotherapy using diffuse intrinsic pontine glioma cell lines L. M. Smyth1, P. A. Rogers2, J. C. Crosbie3, J. F. Donoghue2,3 1

Epworth Radiation Oncology, Melbourne, Australia. 2Department of Obstetrics & Gynaecology, University of Melbourne, Australia. ([email protected] [Presenting author]). 2Department of Obstetrics & Gynaecology, University of Melbourne, Australia. ([email protected]). 3School of Applied Sciences, RMIT University, Melbourne, Australia. ([email protected]), ([email protected])

polyploidy and delayed G2/M arrest, which are associated with radioresistance [2, 3]. Apoptosis and cell-cycle assays showed that at equivalent doses, MRT induced more unrepaired DNA damage that is acutely detrimental to the cell-cycle and cell viability. Conclusion This is the first study to compare the response of DIPG cell lines to MRT and CRT. The acute effects of MRT and CRT were distinctly different, however JHH was intrinsically more radio-resistant to both irradiation modalities compared to SF7761. These results will inform future in vivo studies comparing the efficacy of MRT and CRT on DIPG. Acknowledgements LMS is supported by an Australian Postgraduate Award (University of Melbourne). JFD, PAR and JCC are supported by the Robert Connor Dawes Foundation and the Isabella and Marcus Foundation. This work is supported by the NHMRC [Project Grant 1061772] and Cancer Council Victoria [Grant-in-aid 2013].

References

Introduction Diffuse Intrinsic Pontine Glioma (DIPG) is a devastating paediatric brainstem tumour with extremely poor prognosis and limited treatment options. Radiotherapy is the mainstay treatment for the disease but is limited to palliative use. Synchrotron microbeam radiation therapy (MRT) has shown promising tumour-control results in pre-clinical animal studies whilst demonstrating a remarkable ability to spare normal tissue [1]. The aim of this study was to determine dose-equivalence between MRT and conventional radiotherapy (CRT) and to compare the response of two human DIPG cell lines (SF7761 and JHH) to both modalities. Method Each cell line was exposed to MRT (112 to 560Gy) or CRT (2 to 8Gy) to produce clonogenic cell-survival curves which were fit to the linear-quadratic model using non-linear regression. Equivalent CRT doses were interpolated for each MRT dose. Immunocytochemistry, apoptosis and cell-cycle assays were performed to assess the differences in cellular response between the cell lines and radiotherapy modalities. Results Equivalent CRT and MRT doses for each cell line are summarised in Table 1. SF7761 was significantly more radiosensitive than JHH to both radiation modalities (Fig. 1). JHH exhibited

1. Smyth LM, Senthi S, Crosbie JC, Rogers PA (2016) The normal tissue effects of microbeam radiotherapy: What do we know, and what do we need to know to plan a human clinical trial? Int J Radiat Biol 92:302–11. 2. Erenpreisa J, Cragg MS (2013) Three steps to the immortality of cancer cells: senescence, polyploidy and self-renewal Canc Cell Int 13:92. 3. Gogineni VR, Nalla AK, Gupta R, Dinh DH, Klopfenstein JD, Rao JS (2011) Chk2-mediated G2/M cell cycle arrest maintains radiation resistance in malignant meningioma cells Cancer Lett 313:64–75.

Table 1 Interpolated equivalent CRT doses for increasing MRT doses

1

Equivalent CRT dosesa Cell Line

112 Gy

MRT

250 Gy

MRT

560 Gy

MRT

SF7761

3.21

0.25

6.76

0.43

9.07b

9.07

JHH

2.46

0.11

6.05

0.21

9.25

0.31

a

Interpolated from linear quadratic dose-response curves fitted to the clonogenic survival data

b

Interpolation possible for only one sample

Fig. 1 Day fourteen clonogenic cell-survival curves. Data are presented as mean ± SEM, n = 3, *P \ 0.05, **P \ 0.01

123

O105 A pre-clinical sample positioning system for microbeam radiotherapy at the Australian Synchrotron D. Pelliccia1, J. Livingstone2, F. Gagliardi3, M. Barnes4, D Ha¨usermann5, J. C. Crosbie6 School of Science, RMIT University, Melbourne, Victoria, Australia; Imaging and Medical Beamline, Australian Synchrotron, Clayton, Victoria, Australia, ([email protected]). 2 Imaging and Medical Beamline, Australian Synchrotron, Clayton, Victoria, Australia, ([email protected]). 3Alfred Health Radiation Oncology, Alfred Health, Melbourne, Victoria, Australia, ([email protected]) 4School of Science, RMIT University, Melbourne, Victoria, Australia, ([email protected]). 5Imaging and Medical Beamline, Australian Synchrotron, Clayton, Victoria, Australia, ([email protected]). 6School of Science, RMIT University, Melbourne, Victoria, Australia, ([email protected] [Presenting author]) Introduction Microbeam radiation therapy (MRT), using X-rays from a synchrotron, is a novel, preclinical form of radiotherapy that shows promise of providing a major advance in cancer control if successfully translated to clinical practice (Brauer-Krisch et al, 2010; Crosbie et al, 2010). Clinical translation of MRT requires developing a protocol for a patient positioning system (PPS). Following recent developments in image-guided synchrotron MRT (Pelliccia et al, 2016a and 2016b), we present the implementation of a pre-clinical protocol at the Imaging and Medical Beamline of the Australian Synchrotron.

Australas Phys Eng Sci Med 2

Laser Physics and Photonic Devices, University of South Australia, Australia, Institute for Photonics and Advanced Sensing, University of Adelaide, Australia. ([email protected]). 3 Department of Medical Physics, Royal Adelaide Hospital, Adelaide, Australia, School of Physical Sciences, University of Adelaide. ([email protected]). 4Sansom Institute for Health Research, University of South Australia, Adelaide, Australia International Centre for Allied Health Evidence, University of South Australia, Adelaide, Australia School of Physical Sciences, University of Adelaide, Adelaide, Australia. ([email protected])

Fig. 1 a Digitally reconstructed radiography (DRR) of a plastinated mouse. b–c Details of the DRR showing the target region for the image registration. d Synchrotron image of the same region obtained after registration and positioning of the plastinated mouse

Method The synchrotron PPS will be composed of three key elements: (1) Treatment planning (2) Synchrotron imaging (3) Image registration and patient alignment. The treatment plan, available before the synchrotron session, is imported into the synchrotron control system. Imaging of the patient is done at the beam line, using either the synchrotron beam or a conventional x-ray tube unit. The images are registered with the existing treatment plan and the patient is aligned according to the registration. Verification is performed after alignment and the treatment is initiated. Results We developed the sample positioning system protocol using a small animal phantom, namely a plastinated mouse. The procedure is shown in Fig. 1. A CT scan of the plastinated mouse is imported on the beam line control system (a–c). The sample is then imaged at the beam line and the image is registered with the Digitally Reconstructed Radiography (DRR) from the CT. The registration prompts a sample alignment and image verification (d). Conclusion A preclinical sample positioning system protocol for synchrotron microbeam radiotherapy (MRT) has been realized at the Australian Synchrotron. The protocol marks a further step towards the clinical translation of synchrotron MRT. References 1. Brauer-Krisch E et al (2010) Effects of pulsed, spatially fractionated, microscopic synchrotron X-ray beams on normal and tumoral brain tissue. Mutat Res 704:160–166. 2. Crosbie, J C et al (2010) Tumor cell response to synchrotron microbeam radiation therapy differs markedly from cells in normal tissues. Int J Radiat Oncol Biol Phys 77:886–894. 3. Pelliccia D et al (2016a) Image guidance protocol for synchrotron microbeam radiation therapy. J Synchrotron Rad 23:566–573. 4. Pelliccia D et al (2016b) Phase contrast image guidance for synchrotron microbeam radiotherapy. Phys Med Biol accepted.

Introduction Use of protons from nuclear medicine cyclotrons to irradiate biological samples is a promising tool for preclinical radiobiological studies. This project aims to develop an external beam port on a 16.5 MeV GE PETtrace cyclotron, located at SAHMRI, for this purpose. During routine radioisotope production, the cyclotron operates with beam currents in lA range (up to 135 lA). However, a range of 1–10 nA is required to achieve doserates suitable for radiobiological applications. Therefore, our aims are: (1) to reduce the beam intensity, (2) to measure low beam currents with significantly higher resolution than the standard cyclotron measuring equipment, (3) to add beam scattering foils to reduce beam fluence. Methods A Havar foil was mounted on the end of the cyclotron beamline to produce an external proton beam of 1cm diameter. The cyclotron was run in manual mode to produce stable low current beams in the range 10–1000 nA. The low target currents were measured by a picoampere metre (Keithley) controlled using an in-house developed LabView interface. Additionally, Monte Carlo simulations have been performed (SRIM) to model scatter foils to broaden the beam before hitting the target. Results Stable beam currents below 100 nA have been achieved that are accurately monitored by the picoamper meter. However, the dose

Fig. 1 Left Beam shape and intensity of a non-diffused beam (SRIM data). Right Beam shape and intensity of a beam diffused by a 0.25 mm gold foil (SRIM data)

O016 Design of an external beam port on a 16.5 MeV medical cyclotron for radiobiological studies J. Asp1, S. Afshar2, A. Santos3, E. Bezak4 1

South Australian Health and Medical Research institute, Adelaide, Australia, Laser Physics and Photonic Devices, University of South Australia, Australia, ([email protected] [Presenting author]).

Fig. 2 Beam intensity comparison between the simulation data of a non-diffused beam and a beam diffused by a 0.25 mm gold foil

123

Australas Phys Eng Sci Med rates delivered by these beam currents are in the range of 500 Gy/s, due to the intensity of the beam and the small target volume irradiated by protons. SRIM Monte Carlo simulations have shown that using a 0.25 mm gold foil spreads the beam to *4 9 4 cm2, and reduces the proton energy to 7.5 MeV (Figs. 1 and 2). Conclusion We have been able to stably run the cyclotron at low current below 100 nA and monitor it using a picoampere meter. In order to achieve biologically relevant doses in the range of 1–10 Gy can be achieved by including a gold scattering foil.

O107 Microcalorimetry for synchrotron microbeam radiation therapy beams P. D. Harty1, J. E. Lye2, G. Ramanathan3, D. J. Butler4, T. E. Wright5, V. Takau6, J. C. Crosbie7, A. W. Stevenson8 1

Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected] [Presenting author]. 2ACDS, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]). 3Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]). 4Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]). 5Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]). 6Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]). 7RMIT University, Melbourne VIC. ([email protected]). 8CSIRO and IMBL, Australian Synchrotron, Clayton VIC. ([email protected]) Introduction Future clinical use of microbeam radiation therapy (MRT) beams on the Imaging and Medical Beamline (IMBL) at the Australian Synchrotron will be dependent on high quality dosimetry of the microbeams. For this purpose, a microcalorimeter has been developed at ARPANSA. To provide confidence in the performance of this microcalorimeter, dosimetry measurements have been compared to conventional graphite calorimeter dosimetry using broad beam from the IMBL. Method The microcalorimeter consists of a thin strip of graphene 25 microns thick, sandwiched between much thicker insulating layers of rolled aerogel. The pieces are clamped together with supporting slabs of graphite. A 420-micron diameter thermistor is in contact with the graphene at one end to measure the temperature rise. The thermistors in both the microcalorimeter and conventional calorimeter were calibrated against our secondary standard platinum resistance thermometer. The radiation doses are calculated from the temperature rises in the calorimeters using the specific heat capacity of graphene/graphite and the relative areas of irradiated and unirradiated parts of the graphene/graphite. Results Measurements have been made comparing the microcalorimeter to conventional calorimeter at field sizes 20 mm 9 20 mm, 10 mm 9 10 mm, and 5 mm 9 5 mm. Comparisons of dose determination in a broad beam using the microcalorimeter and a conventional graphite calorimeter show good agreement. The conventional graphite calorimeter was previously shown to be able to measure dose in IMBL beams to an accuracy of *2.3% (k = 1).1,2

123

Conclusion This work represents an important novel contribution to the ongoing pre-clinical radiotherapy experiments on IMBL. To our knowledge, it is the first time a graphene strip has been used to achieve such high spatial resolution in calorimetry measurements. Since the microcalorimeter dosimetry has been validated, it can be used in measurements of the peak to valley dose ratio on MRT beams. References 1. P. D. Harty, J. E. Lye, G. Ramanathan, D. J. Butler, C. J. Hall, A. W. Stevenson, and P. N. Johnston, Med. Phys. 41, 052101 (2014). 2. J. E. Lye, P. D. Harty, D. J. Butler, J. C. Crosbie, J. Livingstone, C. M. Poole, G. Ramanathan, T. Wright and A. W. Stevenson, Phys. Med. Biol. 61, 4201 (2016).

O108 Treatment planning for synchrotron microbeam radiotherapy L. R. J. Day1, L. M. Smyth2, M. Holm3, P. A. W Rogers4, P. E. Engstro¨m5, C. Ceberg6, C. M. Poole7, J. C. Crosbie8 1

School of Science, RMIT University, Australia. ([email protected] [Presenting author]) 2Department of Obstetrics and Gynaecology, University of Melbourne, Australia. ([email protected]). 3Department of Medical Radiation Physics, Lund University, Sweden. ([email protected]). 4 Department of Obstetrics and Gynaecology, University of Melbourne, Australia. ([email protected]). 5Department of Radiation Physics, Lund University Hospital, Sweden. ([email protected]). 6Department of Radiation Physics, Lund University Hospital, Sweden. ([email protected]). 7School of Science, RMIT University, Australia, 2MRD, Melbourne, Australia. ([email protected]). 8School of Science, RMIT University, Australia. ([email protected]) Introduction Synchrotron microbeam radiation therapy (MRT) is a novel radiotherapy modality with significant clinical potential. We have produced a simple dose calculation algorithm for MRT using the Eclipse Treatment Planning System (TPS), by Varian Medical Systems. Method The calculation engine in Eclipse was configured to directly evaluate ‘peak’ doses. Monte Carlo-simulated Peak-to-Valley Dose Ratios were used to obtain the ‘valley’ dose displayed in Eclipse. We compared dose profiles generated by Eclipse with Geant4 Monte Carlo simulations and measurements from the Imaging & Medical Beamline at The Australian Synchrotron. We also performed a plan comparison study using anonymised patient datasets, comparing kilovoltage MRT plans with clinical megavoltage treatment plans. Results The Eclipse TPS performed well in calculating ‘peak’ doses in a water phantom. Considering the simplicity of the algorithm, the ‘valley’ dose and field profiles were also produced with reasonable accuracy, albeit with some underestimation of the valley dose for larger field sizes. Compared to the clinical megavoltage treatment plans, MRT plans demonstrated adequate target coverage whilst meeting normal tissue dose constraints when target volumes were small and relatively superficial. As expected, planning goals for deep seated tumours and target regions distal to bone could not be met using MRT.

Australas Phys Eng Sci Med

Conclusion There are real advantages to using the familiar environment of Eclipse with a new radiotherapy paradigm such as MRT. Although, there are limitations to our MRT calculation engine in Eclipse and further work is required, the data generated in this work are overall encouraging and indicate that the potential for this calculation engine to be implemented in the future as part of a Phase 1 clinical trial.

O109 Effect of micro-motion on proton mini-beams James T. Eagle1, Juergen Meyer2, Steven Marsh3 1

PHD student, Department of Physics and Astronomy, University of Canterbury, NZ. ([email protected] [Presenting author]). 2 Associate Professor of Medical Physics, Radiation Oncology Department, University of Washington, USA. ([email protected]). 3 Director of the Medical Physics Programme, Department of Physics and Astronomy, University of Canterbury, NZ. ([email protected]) Introduction The aim of this research is to quantify the dosimetric impact of internal motion within the brain on spatially modulated proton minibeam radiation therapy (pMRT) for small animal research. Motion of an animal brain caused by cardiac-induced pulsations (CIP) can impact dose deposition. For synchrotron generated high dose rate X-ray microbeams this effect is evaded due to the quasi-instantaneous delivery. By comparison, pMRT potentially suffers increased spread due to lower dose rates. However, for a given dose rate it is less susceptible to beam smearing than microbeams, in the tens of microns range, due to the lower frequency of the spatial modulation, in the hundreds of microns range. Method Monte Carlo simulations in TOPAS were used to model the beam spread for a 50.5 MeV pMRT beam. Motion effects were simulated for a 50 mm thick brass collimator with 0.3 mm slit width and 1.0 mm center-to-center spacing in both a water and a semi realistic water with bone phantom. The maximum motion in a rat brain due to CIP has been reported to be 0.06 mm [1]. Motion was simulated with a peak amplitude in the range 0–0.15 mm. The peak-

Fig. 1 a) Static proton minibeam dose distribution and b) with 0.2 mm periodic peak motion

Fig. 2 Ratio of the motion blurred PVDR to the PVDR for static delivery. The bone phantom has 0.8 mm of surface bone material. The maximum expected cardiac induced brain motion is 60 microns to-valley dose ratio (PVDR) was utilized to evaluate the degradation in the quality of the microbeam. Results The impact of 0.06 mm peak motion was minimal and reduced the PVDR by about 1% at a depth of 10 mm. For 0.2 mm peak motion the PVDR was reduced by 16% at a depth of 10 mm (Figs. 1 and 2). Conclusion Cardiac-induced brain motion has minimal impact on the PVDR for the investigated collimator geometry for the pMRT beam. For more narrow beams the effect is likely to be larger. This indicates that delivery of pMRT to small animal brains should not be affected considerably by beamlines with linac compatible dose rates. References 1. Gilletti, A., & Muthuswamy, J. (2006). Brain micromotion around implants in the rodent somatosensory cortex. Journal of neural engineering, 3(3), 189.

O110 Feasibility of using linac high definition multileaf collimators to create synchrotron-like microbeam responses V. Peng2, L. J. Rogers1,2, E. Claridge Mackonis1, D. R. McKenzie2, N. Suchowerska1,2 1

Dept Radiation Oncology, Chris O’Brien Lifehouse, Sydney, Australia. 2School of Physics, The University of Sydney, Sydney, Australia. ([email protected] [Presenting Author]), ([email protected]), ([email protected]) Keywords grid therapy, microbeam radiotherapy, high definition multileaf collimators Introduction This study examines the feasibility of using high definition multileaf collimators (HDMLC) to create highly modulated dose distributions for realizing the benefits of Microbeam Radiotherapy (MRT). MRT delivers gains in normal tissue sparing, while successfully treating tumours [1–4]. A constraint of MRT is that it requires a medical synchrotron beam-line. Method The HDMLC of a Varian Novalis TxTM linac was used to generate 2.5 and 5.0 mm modulated fields (‘dots, ‘stripes’, ‘tartan’), which were characterized dosimetrically using GafChromicTM EBT3 film. The fields were evaluated using the clonogenic survival assay for normal (HUVEC) and cancer (NCI-H460, HCC-1954) cell lines.

123

Australas Phys Eng Sci Med

Fig. 1 Gafchromic film exposures showing dots, stripes and tartan fields Results Relative to an open field of the same average dose, the 2.5 mm stripes field gave a significantly higher overall survival for the normal cells, but a lower overall survival for cancer cells. Although the same trends were seen for the 5.0 mm stripes field, the difference in survival was not significant (Fig. 1). Conclusion We have confirmed the feasibility of using HDMLCs to create spatially modulated fields to improve the therapeutic ratio, relative to that achieved with a uniform field. The results of this study show how HDMLCs can be utilized to bring the benefits of MRT to the clinic, without the need for specialized microbeam equipment such as the synchrotron. Acknowledgements This work was performed as part of a Masters of Medical Physics research project at the University of Sydney.

Fig. 1 Example of drop-down menus in RABBIT template

References 1. D.N. Slatkin, P. Spanne, F.A. Dilmanian, M. Sandborg, Microbeam radiation therapy, Med Phys, 19 (1992) 1395–1400. 2. F.A. Dilmanian, Y. Qu, L.E. Feinendegen, L.A. Pena, T. Bacarian, F.A. Henn, J. Kalef-Ezra, S. Liu, Z. Zhong, J.W. McDonald, Tissue-sparing effect of x-ray microplanar beams particularly in the CNS: is a bystander effect involved?, Exp Hematol, 35 (2007) 69–77. 3. E. Brauer-Krisch, R. Serduc, E.A. Siegbahn, G. Le Duc, Y. Prezado, A. Bravin, H. Blattmann, J.A. Laissue, Effects of pulsed, spatially fractionated, microscopic synchrotron X-ray beams on normal and tumoral brain tissue, Mutat Res, 704 (2010) 160–166. 4. M.A. Grotzer, E. Schultke, E. Brauer-Krisch, J.A. Laissue, Microbeam radiation therapy: Clinical perspectives, Phys Med, 31 (2015) 564–567.

O111 The RABBIT: A novel tool incorporating TG-100 and TRL for the safe and effective implementation of medical technology A. Ralston1, J. Yuen1 1

St George Hospital Cancer Care Centre, Sydney Australia. ([email protected] [Presenting author]). ([email protected])

Introduction TG-100 [1] outlines the level of risk and quality management now expected in radiation oncology, which is far above current practice in most centres. Local health regulators and standards bodies also mandate active risk management strategies [2, 3]. The aim of the RABBIT (Risk And Benefit Balance In Technology) project was to create a versatile and user-friendly tool to aid compliance with these requirements and ensure the safe and timely implementation of new or existing medical technology. It applies to both commercial products and in-house software/hardware, from early development stages through to clinical implementation.

123

Fig. 2 RABBIT workflow Method The RABBIT methodology is optimised for use by multidisciplinary teams (MDTs). A core concept is the technology readiness level (TRL) approach of Mankins [4] which assesses the maturity of the technology and its associated process factors. The risk/ benefit assessment can use either a simple risk matrix [2], the TG-100 FMEA approach [1] or other methods.

Australas Phys Eng Sci Med

1. 2. 3. 4.

Scope: intended use, restrictions and precautions Assessment of technology readiness level Risk/benefit assessment MDT decision to terminate the project, continue the implementation process or release the technology for clinical use.

These steps can be iterative, depending on departmental goals and resources, until consensus is reached on the final project status. Results The RABBIT has been employed to release many technologies for clinical use, avoiding the challenges that often compromise projects such as unclear goals or scope, insufficient resources and inadequate risk assessments. The RABBIT had widespread acceptance from all professional groups and stimulated overall departmental efforts, such as improving technology, generating intellectual property and decreasing risks. Conclusion The RABBIT methodology is a user-friendly tool which increases safety and efficiency of technology implementation and facilitates compliance with risk and quality management requirements. References 1. Huq MS et al (2016) The report of Task Group 100 of the AAPM: Application of risk analysis methods to radiation therapy quality management. Medical Physics 43, 4209 2. NSW Health (2015) Enterprise-Wide Risk Management Policy and Framework – NSW Health PD2015_043 13-Oct-2015. http://www0.health.nsw.gov.au/policies/pd/2015/pdf/PD2015_ 043.pdf Accessed 19 June 2016. 3. Standards Australia (2009) AS/NZS ISO 31000:2009 Risk Management - Principles and Guidelines. 4. Mankins, J.C. (2009) Technology readiness and risk assessments: A new approach. Acta Astronautica. 65(9)

O112 SNC Machine – Implementing a consistent Linac QA approach across multiple sites A. P. Walsh1 1

Physics Department, ROC Springfield, QLD. ([email protected] [Presenting author])

Introduction With the expansion of Radiation Oncology Centres (ROC) and the need for physicists to provide support in unfamiliar sites, streamlining and achieving consistency in QA processes is important to maintain the safe delivery of RT. SNC Machine offers the possibility to standardise a large portion of routine Physics QA across multiple sites using its server based application accessed via citrix/web browser. ROC have purchased SNC Machine and have begun implementing across their sites. Method Current QA tests were replaced where possible using the SNC Machine equivalent tests. This involved replacing some ‘inhouse’ written programs which were site specific, and not practicable across a multisite environment. These tests included (but not limited to) light radiation congruence, picket fence tests, VMAT Ling tests and image quality tests. A treatment plan was created in the R&V system for each linac, containing a number of fields, where each field corresponds to a particular test. The plan is delivered on the linac and the MV/ kV/ CBCT image is captured and exported to SNC Machine for analysis, either by file export, or a direct DICOM query/retrieve

process. SNC provide phantoms for light radiation congruence and image quality tests. Results So far to date a consistent protocol for linac QA utilising SNC machine has been successfully rolled out in six different ROC sites, the initial implementation site has been using SNC machine for eleven months. Conclusion SNC Machine offers a robust and configurable QA system that can be implemented across multiple sites. This allowed a single procedure to be developed and followed, reducing the training burden and improving consistency of QA. Having a single database also allows a large number of linacs across different sites to have QA results trended and compared on a like for like basis in a central location.

O113 Experience with an in-house developed automated linac QA program Jack Chonjnowski1, Dr. Jonathan Skyes2,3, Shan Yau3, Prof David Thwaites2 1

Mid North Coast Cancer Care Institute, Coffs Harbour, Australia. Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected] [Presenting author]). 3Sydney West Cancer Network, Australia

2

Introduction With advances in linear accelerator (linac) technology and the development of modern treatment techniques, it is becoming more difficult and tedious to perform adequate high quality performance checks of linacs. The recommended QA tests and frequencies are becoming outdated (e.g. TG-142). To overcome those issues, the linac QA program has been entirely reorganised to make use of the EPID and in-house developed software. Method In this approach, most of the regular linac QA is done by executing specially designed plans in clinical mode and collecting acquired EPID images on different linacs. Those images are then further analysed and a report is generated with detailed results. The development, implementation, application and use of such an inhouse developed automated linac QA program presents practical problems, for example how to replace assessment of coincidence of radiation and mechanical isocentres or spokeshots by other methods utilising the EPID. Results It takes less than 3 hours to complete linac monthly QA, delivery is independent on the user skills and experience and results are calculated and presented in less than a minute. The results of the new program, introduced 3 years ago at the Crown Princess Mary Cancer Care Centre (Westmead, Australia) are highly reproducible with submillimetre accuracy as in the example of the MLC alignment assessment in Fig. 1.

MLC alignment [mm]

The RABBIT utilises an embedded Word template with customisable drop-down selections (Fig. 1) in four steps (Fig. 2):

a)

Crossleaf

Inleaf

0.15 0.1 0.05 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Monthly QA

b)

Fig. 1 (a) Visual assessment of the MLC alignment test (b) New method of quantitative monitoring of MLC alignment results from automated linac QA

123

Australas Phys Eng Sci Med Conclusion The automated linac QA program is now in routine use in the department and its implementation has significantly improved the efficiency and effectiveness of monitoring and detecting out of tolerance machine performance parameters.

O114 Investigation of a gantry-mounted wire array for quality assurance of dynamic wedge and IMRT plan transfer between linear accelerators M. Lange1, T. Kron1 Dept. of Physical Sciences, Peter MacCallum Cancer Centre, Melbourne, Australia. ([email protected]). ([email protected] [Presenting author]) Introduction We have explored additional uses for a transmission multiwire ionization chamber (PTW David). In particular we have addressed the transfer of photon IMRT plans between dosimetrically equivalent linacs and the quality control (QC) of dynamic wedges. Method The detector features multiple wires monitoring the windows between the leaf pairs of a multi-leaf collimator. The photon fluence in each window is measured and compared to a reference delivery of a radiotherapy treatment plan. We have repeatedly measured a 7-field prostate IMRT plan both at the same and an equivalent linac. Because of the simple yet precise localization in the accessory tray and high measurement precision we have also explored the use of the transmission chamber for QC of the Varian enhanced dynamic wedge. Four EDW fields were compared on the same and between different linacs over the course of several days, compensating for zero effect and change in linac output. Results For the IMRT plan, we have found that while the mean deviation over all leaf pairs increased by less than 0.5% by changing the linac, particularly the outermost leaf pairs of an IMRT field differed by as much as 5% between two linacs (compared to 1.5% on identical linacs). Variations in EDW delivery were below 0.2% using a single linac. Between two linacs, a similar agreement was found over most of the field, but a persistent difference of around 4% was observed on the wires closest to the stop position of the travelling jaw, which can be explained by a slight offset in the Y jaw positions between the machines. Conclusion The high measurement precision of the David transmission detector, together with the self-aligning mount make it a useful tool both for QA of a single machine, as well as for intercomparison studies within a group of equivalent linacs. Acknowledgement We acknowledge support of this work from PTW and Nucletron Australia by supplying the detector system for the duration of the measurements.

O115 A simple and robust model for in vivo dosimetry with a water equivalent EPID S. Deshpande1,2, A. Xing1, S. Blake1,3, P. Metcalfe1,2, L. Holloway1,2,3,4, P. Vial1,2 1

Liverpool and Macarthur Cancer Therapy Centre and Ingham Institute for Applied Medical Research, Sydney Australia. ([email protected]). 2Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia ([email protected]), ([email protected]), ([email protected] [Presenting author]). 3Institute

123

Fig. 1 Representation of phantom and virtual WE-EPID setup in TPS of Medical Physics, school of Physics, University of Sydney, NSW, Australia. ([email protected]). 4South West Sydney Clinical School, School of Medicine, University of NSW, Australia. ([email protected]) Introduction This work demonstrates how a modified water-equivalent electronic portal imaging device (WE-EPID) and a conventional treatment planning system (TPS) can facilitate accurate in vivo dosimetry. Method A standard EPID was modified into a WE-EPID as described in previous studies [1,2]. In this study we use a conventional TPS (Pinnacle, v9.10, Philips Medical Systems, Fitchburg, WI) to calculate the dose to a WE-EPID under clinical conditions. To facilitate dose calculation at the EPID distance, which is outside the normal CT field of view (FOV), the CT images were padded to an extended FOV with pixels assigned the density of air (Fig. 1). The TPS model accuracy for calculating dose at the EPID plane was evaluated by comparison to dose measurement in transit geometry with different conditions of field size, off-axis position and thickness of object in beam. Reference dose measurements were conducted with both a 2D ion-chamber array (ICA) and the WE-EPID. Clinical fields (open, IMRT and VMAT) with homogenous and inhomogeneous phantoms in beam were evaluated using 3%/3 mm gamma criteria. All experiments were conducted with same source to detector distance of 150 cm using 6 MV photon beam only. Results In all open fields the measured dose response agreed with TPS to within 1.5%. Gamma evaluation showed good agreement ([94%) for both integrated (IMRT and VMAT) and control-point doses (VMAT) for all clinical fields for both WE-EPID and ICA when compared with calculated TPS dose. Conclusion The TPS dose in water model accurately calculates dose at EPID plane. This work presents a new approach for implementing in vivo dosimetry that has all the advantages of EPID dosimetry but without the additional processes and uncertainties required to model non-water equivalent EPID dose response. References 1. P. Vial, P.B. Greer, L. Oliver, C. Baldock, ‘‘Initial evaluation of a commercial EPID modified to a novel direct-detection configuration for radiotherapy dosimetry,’’ Medical physics 35, 4362–4374 (2008). 2. S.J. Blake, A.L. McNamara, S. Deshpande, L. Holloway, P.B. Greer, Z. Kuncic, P. Vial, ‘‘Characterization of a novel EPID

Australas Phys Eng Sci Med designed for simultaneous imaging and dose verification in radiotherapy,’’ Med. Phys. 40, 091902(091901–091911) (2013). Acknowledgements This work was supported from Cancer Institute NSW Australia (Research Equipment Grant 10/REG/1-20) and Cancer Council NSW (Grant ID RG 1-06).

MO25 Monte Carlo Model of the Intrabeam kV System using BEAMnrc M. A. Barry1, R. Jones1, M. Fay1, D. J. Butler2, J. Lehmann1,3,4 1

Radiation Oncology Department, Calvary Mater Newcastle, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2 Australian Radiation Protection and Nuclear Safety Agency, Yallambie, Australia. ([email protected]). 3The University of Newcastle, Newcastle, Australia. 4The University of Sydney, Sydney, Australia. ([email protected]) Introduction IntrabeamTM is a low energy (50 kV) X-ray system designed for intracavity treatments. The unit at Calvary Mater Newcastle is being commissioned for radiobiology research, especially cell and small animal irradiations. To assist in the commissioning process, a Monte Carlo model of the system was developed using the EGSnrc platform. Method The BEAMnrc model is based on manufacturer documentation and published data [1] and uses the FLATFILT component module and the NRC swept beam source. Initial simulations (2e9 histories/simulation) were carried out matching the modified, open applicator (Fig. 1). At 5 cm from the tip of the Intrabeam and perpendicular to its axis, film in air was simulated (4e7 histories/simulation) in DOSXYZnrc. Physical dimensions matched that of EBT3 film but the media was set as water. Measurements with the same set-up were also carried out. To convert EBT3 film readings to (relative) dose, low energy film calibrations were performed at both ARPANSA and with the Intrabeam system itself at Newcastle. Results Initial simulations (1–12) were carried out for different swept beam angles covering the full extent of the tube tip, 1 being the largest angle, decreasing to zero in 12 equal steps. The radius of the parallel beam was set as 0.005 mm. Figure 2 shows the dose profiles for the closest matching DOSXYZnrc simulations (4, 5 and 6) compared to film results. Conclusion Data from different versions of the initial model has been collected and compared to experimental film data. Further work (starting with increasing parallel beam radii) will be carried out to fine tune the model, which will then be used to aid in the design of

Fig. 1 BEAMnrc Intrabeam model geometry and photo of Intrabeam system

Fig. 2 Dose profiles for measured and simulated data collimators for the Intrabeam system and to optimise the experimental setup. References 1. S. Clausen, F. Schneider, L. Jahnke, J. Fleckenstein, J. Hesser, G. Glatting, and F. Wenz, (2012) ‘‘A Monte Carlo based source model for dose calculation of endovaginal TARGIT brachytherapy with INTRABEAM and a cylindrical applicator.,’’ Z. Med. Phys., 22(3):197–204.

MO26 Characterisation of MRT field at the Australian Synchrotron using a novel silicon strip detector A. Dipuglia1, M. Petasecca1, M. Cameron1, J. Davis1, V. Perevertaylo2, A. B. Rosenfeld1, M. Lerch1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]) ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2 SPA-BIT, Kiev, Ukraine. ([email protected]) Introduction Microbeam Radiation Therapy (MRT) uses synchrotron generated X-rays with an extremely high dose rate *20 kGy/sec. The X-rays are physically collimated into a planar, parallel array of microbeams. The beams are produced by a Multi-Slit Collimator (MSC) resulting in 50 lm FWHM microbeams with spacing of 400 lm between them. Using such beams leads to the dose volume effect resulting in a non-cancerous tissue sparing effect, which is not observed to the same extent in tumour tissue.[1] Method A novel n-type silicon detector produced by The Centre for Medical Radiation Physics (CMRP) at UOW – the PNP – was design to satisfy the requirements for MRT dosimetry. Characterisation of the MRT field was performed in hutch 2b of the Imaging and Medical Beamline (IMBL) at the Australian Synchrotron (AS). The MRT field consisted of vertical microbeams with 50 lm FWHM peaks and a pitch of 400 lm. The profiles were obtained with the PNP SSD in a RMI457 solid water phantom at 2 cm depth after calibration with a pinpoint ionisation chamber. Results The profile shown in Fig. 1 shows the beam profile obtained at the IMBL, AS. The average FWHM of the microbeams was calculated to be (56.0 ± 0.3) lm with a Peak to Valley Dose Ratio

123

Australas Phys Eng Sci Med ([email protected]). 2Discipline of Biomedical Sciences, School of Health and Biomedical Sciences, RMIT University, Melbourne, Australia. ([email protected]). 3State Key Laboratory of Electroanalytical Chemistry Changchun Institute of Applied Chemistry Chinese Academy of Sciences (CIAC), Changchun, China. ([email protected]). 4The Division of Health Sciences, School of Pharmacy and Medical Sciences, University of South Australia (Unisa), Adelaide, Australia. ([email protected])

Fig. 1 (a) Measured microbeam profile of the intrinsic MRT radiation field. (b) Zoomed microbeam profile displaying three peaks near the centre of the MRT field (PVDR) of (178 ± 3). Due to the divergence of the beam the FWHM is expected to be larger than the manufactured MSC width. Peak profile asymmetry may be related to substrate dose enhancement [2] Conclusion The CMRP PNP detector can successfully resolve the microbeams and display the beam profile used at the AS on IMBL in hutch 2b. Comparison with other detectors and geant4 simulations will be performed to verify the results.

Nano-materials with high atomic number atoms have been demonstrated to enhance the effective radiation dose and thus potentially could improve radiotherapy effect. Recent studies conducted by our group have shown that bismuth based nanoparticles (NPs) ‘‘Bi2S3’’ demonstrate a higher radiation dose enhancement when compared to gold nanoparticles in both phantom1 and cell studies2. In this study we used Bi nanoparticles coated with starch and Bi2S3 coated with PVP. These NPs are of low toxicity and are one of the least expensive heavy metal-based nanoparticles. The aims of this study were to synthesise Bi2S3 and Bi NPs, and examine their cytotoxicity to human lung adenocarcinoma epithelial cells (A549). The dose enhancing effects of NPs on A549 cells were examined at both KV and MV energies. The preliminary results revealed that the NPs show increased radio-sensitisation of cells, displaying dose enhancement with KV X-ray energies and to a lesser degree for the MV energies. We also observed that Bi NPs generated a greater dose enhancement effect than Bi2S3 NPs in irradiated A549 cells. The maximum Dose Enhancement Factor (DEF) was obtained at lower energy KV range when cells treated with BiNPs (1.5) compared to the DEF of 1.2 when cells treated with Bi2S3 NPs. Less radiation dose enhancement was observed when using high energy MV beam with higher DEF value of BiNPs treatment (1.26) as compared to 1.06 DEF value with Bi2S3

References 1. Zeman, W. Curtis, H. and Baker, C. (1961) ‘‘Histopathologic effect of high-energy-particle microbeams on the visual cortex of the mouse brain’’ Radiat. Res. 15, 496–514. 2. Rosenfeld, A. Siegbahn E. Brauer-Krisch, E. Holmes-Siedle, A. Lerch, M. (2005) ‘‘Edge-on face-to-face MOSFET for synchrotron microbeam dosimetry: MC modelling’’ IEEE Transaction on Nuclear Science, 52(6)1, 2562–2569. Acknowledgements CMRP thanks the AS-IMBL team (A. Stevenson, J. Livingston, C. Hall, and D. Hausermann) for their assistance and acknowledge support of the NH&MRC (APP1093256) and the Australian Synchrotron (AS153/IM/10045).

MO27 Comparison of Bismuth-based nanoparticles as radiosensitization agents for Radiotherapy M. Algethami1, M. Geso1, F. Bryce2, L. Lu3, A. Blencowe4 1

Discipline of Medical Radiations, School of Health and Biomedical Sciences, RMIT University, Melbourne, Australia. ([email protected] [Presenting author]),

123

Fig. 1 (a) Relevant interactions of incident x-rays with cells treated with Bi NPs. (b) Effect of KV X-rays on the survival of A549 cells treated with Bi2S3 and Bi NPs at different concentration

Australas Phys Eng Sci Med NPs. The greater dose enhancement was achieved at KV, due the effect of the photoelectric effect which is the dominant process of interaction of X-ray. The cytotoxic effect of Bi NPs on enhancing the X-ray dose was higher due to the higher amount of elemental Bismuth present in Bi NPs compared to Bi2S3 NPs. Therefore, Bismuth NPs can be considered as valuable dose enhancing agents when used in clinical applications (Fig. 1). References 1. Alqathami M, Blencowe A, Yeo U, et al. Enhancement of radiation effects by bismuth oxide nanoparticles for kilovoltage x-ray beams: a dosimetric study using a novel multi-compartment 3D radiochromic dosimeter. Paper presented at: Journal of Physics: Conference Series 2013. 2. Algethami M, Geso M, Piva T, Blencowe A, Lu L. Radiation Dose Enhancement Using Bi2S3 Nanoparticles in Cultured Mouse PC3 Prostate and B16 Melanoma Cells. NanoWorld J. 2015;1(3):97–102.

MO28 Recycling LCD monitors as radiation warning lights M. Sheedy1, A. T. Cousins1, P. Tolson1, C. Morison1 1

Medical Physics & Bioengineering, Canterbury District Health Board. ([email protected]), ([email protected] [Presenting author])

Introduction New Zealand Radiation Safety Regulations (1) and Codes of Safe Practice (2) state the requirements for radiation warning lights and signs at the entrance to linear accelerator bunkers. The existing lights consisted of a green lamp indicating an active controlled area and a red lamp indicating an active radiation beam. These lamps had become obsolete and replacement bulbs were becoming increasingly difficult to find. Additionally it was not thought clear what the individual lamps meant. Method The Christchurch Hospital Information Services Group was replacing LCD monitors. The discarded 19’’ LCD monitors were used to replace the green and red lamps. They are driven by a single Arduino Uno board to keep cost to a minimum. The LCD monitors displaying the radiation beam on status needed to become active in time with the beam. The horizontal and vertical synchronization lines were kept active to enable rapid switching of the red background. The signs were connected to each Elekta LINAC through the customer interface board. This enabled switching of the ‘‘Radiation Controlled Area’’ LCD using the controlled area lamp contacts and the ‘‘Beam On’’ LCD using the radiation on lamp contacts. Results The LCD Monitor based radiation warning signs give a clear indication of the controlled area and radiation beam activity. The signs can be seen anywhere from within the treatment waiting area. Conclusion Recycling LCD monitors has enabled a cheap but effective solution. The LINAC controlled area is now well defined and compliant with the relevant legislation minimising the risk of unauthorised entry and treatment delays.

MO29 Rectal dose-response relationships from pooled prostate radiotherapy outcomes data: Utility of equivalent uniform dose A. Cicchetti1, T. Rancati1, M. A. Ebert2,3, C. Fiorino4, A. Kennedy2, D. J. Joseph2, J. W. Denham5, V. Vavassori6, G. Fellin7, R. Valdagni8 1

Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy. ([email protected]), ([email protected]). 2Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia. ([email protected]), ([email protected]). 3School of Physics, University of Western Australia, Western Australia. ([email protected] [Presenting author]). 4San Raffaele Scientific Institute, Milan, Italy. ([email protected]). 5School of Medicine and Public Health, University of Newcastle, Callaghan, NSW. ([email protected]). 6Cliniche HumanitasGavazzeni, Bergamo, Italy.([email protected]). 7 Ospedale Santa Chiara, Trento, Italy. ([email protected]). 8 Universita` di Milano Radiation Oncology and Prostate Cancer Program Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy. ([email protected]) Introduction Collection of digital radiotherapy planning data has facilitated the pooling of datasets for outcomes analysis with higher power than that of the individual studies. Two large prospective trial datasets were combined to develop models predictive of late rectal bleeding in prostate cancer patients following radiotherapy. Method Data were pooled from the Airopros 0102 (Fellin et al 2014) and TROG 03.04 RADAR (Denham et al, 2012) trials of 3DCRT at 66–80 Gy. G3 late rectal bleeding was prospectively scored using the SOMA/LENT questionnaire, with a minimum follow-up of 3 years. Rectal dose-volume histograms were reduced to equivalent uniform dose (EUD) calculated with previously-derived volume parameter n = [0.06; 0.05; 0.018] and combined with clinical and treatment features in multivariable logistic (MVL) regression. The results of multivariate analyses were used to develop a nomogram to predict long-term toxicity. Results Data were available for a combined 1337 participants. G3 LRB was scored in 95 pts (7.1%). EUD calculated with n = 0.06 was

References 1. Radiation Protection Regulations 1982. Retrieved from http:// www.legislation.govt.nz/regulation/public/1982/0072/latest/ DLM81126.html 2. Ministry of Health. Code Of Safe Practice For The Use Of Irradiating Apparatus In Medical Therapy, Version 1.4, Office of Radiation Safety, 2010.

Fig. 1 Grade 3 late rectal bleeding as a function of EUD (n = 0.06) and clinical features. Lines represent model predictions and symbols observations

123

Australas Phys Eng Sci Med the best dosimetric predictor. A 4-variable MVL model was fitted including EUD (odds ratio (OR) = 1.07 p = 0.16), irradiation of seminal vesicles (OR = 4.75 p \ 0.001), previous abdominal surgery (OR = 2.30 p = 0.02) and cardiovascular disease (OR = 1.42 p = 0.18). Predicted toxicity curves together with observed toxicity rates are presented in Fig. 1. The AUC of the model was 0.63. Inclusion of acute toxicity (OR = 2.34 p \ 0.001) slightly improved AUC (0.65), thus confirming a possible role of consequential injury. Conclusion The predictive power of EUD at low n emphasises the importance of high doses on late rectal bleeding. Several clinical dose-modifying factors were identified. This study confirmed the benefit of pooling of data sources for the improvement and validation of predictive models. References 1. Fellin G, Rancati T, Fiorino C et al (2014) Long term rectal function after high-dose prostate cancer radiotherapy: results from a prospective cohort study. Radioth Oncol, 110:272–7 2. Denham JW, Wilcox C, Lamb D et al (2012) Rectal and urinary dysfunction in the TROG 03.04 RADAR trial for locally advanced prostate cancer. Radioth Oncol, 105:184–92

Fig. 1 Water tank with and without anthropomorphic phantom to provide patient scatter

Acknowledgements The study was funded by: AIRC IG16087, Fondazione Monzino, NHMRC (300705, 455521, 1006447).

MO30 Out-of-field dose to critical structures during VMAT T. Baradaran1, D. L. Cutajar1, Y. Qi1, A. Ralston2 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2St George Cancer Care Centre, St George Hospital, Australia. ([email protected]) Introduction A re-evaluation of the eye lens radio-sensitivity by ICRP has reduced the threshold dose for lens opacities from 8 Gy to 0.5 Gy [1]. The most radio-sensitive region of the eye lens is at a depth of 2–4 mm [2]. The adoption of IMRT/VMAT has led to an increase in monitor units delivered, resulting in higher out-of-field doses. The aim of this study was to provide an estimation of out-offield doses to the eye lens during 6MV VMAT. Method The out-of-field dose distribution was profiled for a Varian iX 6MV photon field at depths of 5 mm, 15 mm and 37.5 mm, using an ionisation chamber within a motorised water tank (Fig. 1), expanding upon previous studies [3,4]. Profiles were obtained with and without an anthropomorphic pelvis phantom to provide patient scatter. The dose to the eye lens from clinical VMAT prostate treatments was also measured on a full anthropomorphic phantom with a cylindrical chamber placed at a depth of 3 mm. Results The out-of-field dose profiles (Fig. 2) show the dose is higher at shallower depths than dmax, probably due to electrons. There is a dose peak at 20 cm from isocentre resulting in a four-fold dose increase. Eye lens doses measured on the full anthropomorphic phantom for three VMAT plans delivering 70 Gy to the prostate ranged from 6 to 10 cGy. Conclusion As expected, doses to the eye lens for a clinical VMAT prostate treatment were measured to be below the recommended

123

Fig. 2 Out-of-field dose profiles of head leakage/scatter (no patient scatter) for a 10 cm 9 10 cm field

thresholds (10 cGy compared to 50 cGy). However, this work has shown that doses may be higher than predicted by most planning systems as they do not accurately model photon or electron leakage/ scatter from the linac head. References 1. ICRP, 2011. Statement on Tissue Reactions, ref 4825-3093-1464 2. Behrens R, Dietze G and Zankl M, ‘‘Dose conversion coefficients for electron exposure of the human eye lens’’, Phys. Med. Biol. 54, 2006, 4069–4087 3. Kry SF, White Allen R, Salehpour M, Mohan R. ‘‘A Monte Carlo model for calculating out-of-field dose from a Varian 6 MV beam.’’ Int. J. Med. Phys. Res. Prac 33, 2006, 4405–4413 4. Ruben, Jeremy D., Craig M. Lancaster, Phillip Jones, and Ryan L. Smith. ‘‘A comparison of out-of-field dose and its constituent components for intensity-modulated radiation therapy versus conformal radiation therapy: implications for carcinogenesis.’’, Int J. Rad. Oncol. Biol. Phys. 81(5), 2011, 1458–1464

Australas Phys Eng Sci Med References

MO31 Magnetic fringe field influence on linac beam profiles J. Begg1,2,3, B. Beeksma1,2, G. Goozee1,4, B. Dong1,2, P. Keall5,2, G. Liney1,2, L. Holloway1,2,3,4,6,7 1

Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Australia. 2Ingham Institute for Applied Medical Research, Australia. ([email protected]) ([email protected]), ([email protected]). 3 South Western Sydney Clinical School, University of New South Wales, Australia. ([email protected] [Presenting author]). 4Institute for Applied Medical Research, Sydney, Australia. ([email protected]). 5Sydney Medical School, University of Sydney, Sydney, Australia. ([email protected]). 6Centre for Medical Radiation Physics, UOW, Australia. 7Sydney Medical School, University of Sydney, Australia. ([email protected])

A 1 0 -1

2

B 1 0

-12 -8 -4 0 4 8 12 Off axis distance (cm)

2

C 1 0

-1

-1

MO32 The impact of incorporating radiomic features in a NSCLC two-year survival model for radiotherapy decision support M. Field1, M. S. Barakat1, D. Thwaites2, A. Ghose1, D. Stirling1, A. Dekker3, M. Carolan4, S. Vinod5,6, L. Holloway7 1

Faculty of Engineering and Information Sciences, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected]). 3Department of Radiation Oncology (MAASTRO), GROW, School for Oncology and Developmental Biology, Maastricht University Medical Centre, Maastricht, The Netherlands. ([email protected]). 4Illawarra and Shoalhaven Cancer Care Centres and Centre of Medical Radiation Physics, University of Wollongong. ([email protected]). 5Liverpool Cancer Therapy Centre, Liverpool Hospital, Australia. 6SWSCS, University of New South Wales Australia. ([email protected]). 7 Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute for Applied Medical Research, Institute of Medical Physics, University of Sydney, Centre of Medical Radiation Physics, University of Wollongong and SWSCS, University of New South Wales Australia. ([email protected])

-2

-2

-2

Diff rel to Gantry 0(%)

2

Diff rel to Gantry 0(%)

Diff rel to Gantry 0(%)

Introduction The presence of MRIs in radiotherapy clinics in Australia is increasing due to dedicated planning scanners[1] and the development of MRI guided radiotherapy systems[2–5]. Magnetic fringe field from MRIs will potentially influence nearby clinical linacs[6]. This work investigates the potential influence of the magnetic field from the Australian MRI-Linac on a nearby Elekta Synergy linac. Method The Australian MRI-linac bunker was designed with passive shielding to reduce the magnetic fringe field within the nearby bunker to acceptable limits. Magnetic field strength was measured pre and post ramp-up of the magnet with a Vector/Magnitude Gaussmeter Model VGM(AlphaLab,Inc), at isocentre and 1m from isocentre in the G-T, A-B and Up-Down directions. Magnetic field strengths were compared to environmental guidelines(\1 G at any point over the linac, \0.2 G change over 0.5 m)[7]. A MatriXX (IBA-Dosimetry, Germany) ion chamber array was mounted on the linac to acquire 20 9 20 cm2 fields at isocentre distance, 5.3 cm depth, for gantry angles of 0, 90, 180 and 270 pre and post magnet ramp-up. Profile differences were calculated relative to gantry 0. Results Changes in magnetic field were within specifications. The largest change in magnitude was -0.74 G (y-axis:+0.14 G to -0.6 G) at the point closest to the MRI-linac. Crossplane profile differences acquired pre, post ramp-up and post ramp-up following standard control mechanisms are in Fig. 1. Profiles differences were within AAPM TG-142[8] tolerances after application of standard control mechanisms. Conclusion Changes in the magnetic field were within environmental specifications, however differences between profiles measured at gantry 90, 180 and 270 degrees relative to gantry 0 were observed. Differences were successfully corrected using Elekta standard control mechanisms.

1. Batumalai V (2016) Personal Communication. 2. Keall PJ, Barton M, Crozier S (2014) The Australian Magnetic Resonance Imaging–Linac Program. Seminars in Radiation Oncology 24 (3):203–206. doi:10.1016/j.semradonc.2014.02.015 3. Lagendijk JJW, Raaymakers BW, van Vulpen M The Magnetic Resonance Imaging–Linac System. Seminars in Radiation Oncology 24 (3):207–209. doi:10.1016/j.semradonc.2014.02. 009 4. Fallone BG The Rotating Biplanar Linac–Magnetic Resonance Imaging System. Seminars in Radiation Oncology 24 (3):200–202. doi:10.1016/j.semradonc.2014.02.011 5. Mutic S, Dempsey JF The ViewRay System: Magnetic Resonance–Guided and Controlled Radiotherapy. Seminars in Radiation Oncology 24 (3):196–199. doi:10.1016/j.semradonc. 2014.02.008 6. Raaymakers B, Lagendijk J, Overweg J, Kok J, Raaijmakers A, Kerkhof E, Van Der Put R, Meijsing I, Crijns S, Benedosso F (2009) Integrating a 1.5 T MRI scanner with a 6 MV accelerator: proof of concept. Physics in medicine and biology 54 (12):N229 7. Elekta Site Planning Guidelines. 8. Klein EE, Hanley J, Bayouth J, Yin FF, Simon W, Dresser S, Serago C, Aguirre F, Ma L, Arjomandy B (2009) Task Group 142 report: quality assurance of medical accelerators. Medical physics 36:4197

-12 -8 -4 0 4 8 12 Off axis distance (cm)

-12

-8 -4 0 4 8 Off axis distance (cm)

12

Fig. 1 Crossplane profile differences relative to gantry 0 acquired at gantry 90 (solid line), 180 (dashed line) and 270 (dotted) prior to magnet ramp-up (A), post magnet ramp-up (B) and post magnet ramp up after application of standard control mechanisms (C)

Introduction The characterisation of contoured CT tumour regions with a range of high-throughput shape and texture descriptors, termed radiomics, is expected improve the stratification of survival outcomes [1]. The aim of this study is to quantify the impact of including radiomic features on clinical decision support systems (DSS) for predicting two year survival from non-small cell lung cancer.

123

Australas Phys Eng Sci Med Method Data was extracted from Liverpool Cancer Therapy Centre from oncology and image information systems in a standardized and de-identified format. Clinical variables included lung function test, ECOG, age, gender and tumour volume (GTV and nodes). Radiomic features were calculated on DICOM CT and structure set with an open-source toolbox [2], resulting in 42 tumour descriptors. Logistic regression was applied as a classifier in each of the datasets after employing different feature selection approaches, such as ranking with Spearman correlation, univariate classification and supervised principal component analysis, and we reported the average 10-fold cross validation AUC of the classifier. Results A group of 255 non-metastatic, radically treated ([40 Gy prescription dose) patients were included in the dataset. In comparing feature ranking approaches with dimensionality reduction and the eventual inclusion or exclusion of standard clinical variables the analysis indicates that marginal improvement is gained in predictive accuracy. A DSS trained with clinical variables alone achieved 0.67 while including the radiomic feature with highest Spearman’s correlation coefficient with respect to outcome or including up to 10 principal components reached an AUC of 0.69 on hold-out datasets. Conclusion The inclusion of radiomics features within clinical DSSs can improve outcome prediction accuracy if feature selection is applied. Negligible differences were found between the investigated approaches to feature selection. Further improvements in classification may be facilitated by diversifying the data set with a multicenter distributed learning network and by additional feature extraction methods. References 1. Aerts et al. ‘‘Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach’’, Nature Communications, 2014 2. Vallieres et al. ‘‘A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in softtissue sarcomas of extremities’’, Physics in Medicine and Biology, 2015.

MO33 Characterisation of the MV imaging performance of a novel water-equivalent EPID Z. Cheng1, S. J. Blake1,2, S. Atakaramians3, M. Lu4, S. Meikle5, P. Vial1,2,6, Z. Kuncic1 1

Institute of Medical Physics, School of Physics, University of Sydney, Australia. ([email protected] [Presenting author]). 2 Ingham Institute for Applied Medical Research, Australia. 3Institute of Photonics and Optical Science, School of Physics, University of Sydney, Australia. 4Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centres, Australia. 5PerkinElmer Medical Imaging, United States of America. 6Faculty of Health Sciences & Brain and Mind Centre, University of Sydney, Australia Introduction Our group has developed a novel water-equivalent EPID prototype that utilises an array of plastic scintillating fibres [1]. Monte Carlo (MC) simulations of this novel EPID were used with physical measurements to characterise the prototype’s imaging performance at megavoltage (MV) energies relevant for radiotherapy.

123

Fig. 1 Measured and simulated normalized MTF for the standard and prototype EPID configurations

Method The prototype EPID is comprised of an array of 3 cm long plastic scintillating fibres situated directly on the a-Si photodiode array of a PerkinElmer flat-panel x-ray imager, replacing the standard copper/phosphor screen. Using clinical 6 MV photon beams, imaging performance metrics including noise power spectrum (NPS), modulation transfer function (MTF) and detective quantum efficiency (DQE) were evaluated relative to those of a standard EPID using MC simulations and experimental measurements. Results The spatial frequency f50 at which the standard EPID’s MTF equals 50% was measured as 0.26 mm-1 and MC-calculated as 0.21 mm-1, whereas f50 for the prototype EPID was 0.12 mm-1 (measured) and 0.16 mm-1 (simulated) (see Fig. 1). Intial estimates of the MC-calculated DQE(0) for the standard and prototype EPID configurations were *1% and 6.8%, respectively, which are consistent with previously published findings [2]. The prototype configuration’s DQE(0) was found to be sensitive to variations in optical transport parameters, which suggests that optimizing select optical properties may enhance its DQE. Conclusion A water-equivlaent EPID prototype has been developed. Its imaging perfomance has been characterised in terms of MTF, NPS and DQE relative to that of a standard phosphor-based EPID. The prototype exhibits a reduced MTF compared to the standard EPID, yet offers greater DQE(0). Further optimisation of geometrical and optical parameters of the scintillator is required to match or improve upon the imaging performance of standard EPIDs. Acknowledgements The authors acknowledge funding from the Australian Research Council (ARC) Linkage Project grant LP150101212, Perkin-Elmer Pty Ltd, Cancer Institute NSW (Research Equipment Grant 10/REG/1-20) and Cancer Council NSW (Grant ID RG 11-06). ZC also thanks The University of Sydney and the Australian Government for scholarship support in the form of an Australian Postgraduate Award. SA acknowledges a support of ARC funding DE140100614.

Australas Phys Eng Sci Med References 1. Blake et al. (2013) Characterization of a novel EPID designed for simultaneous imaging and dose verification in radiotherapy. Med Phys 40(9) 091902 2. Teymurazyan et al. (2012) Monte Carlo simulation of a novel water-equivalent electronic portal imaging device using plastic scintillating fibers. Med Phys 39(3) 1518–29

MO34 A feasibility study of machine performance check for daily beam output of truebeam linear accelerator Woo Sang Ahn1, Wonsik Choi1, Seong Soo Shin1 1

Department of Radiation Oncology, Gangneung Asan Hospital, University of Ulsan College of Medicine, Republic of Korea. ([email protected] [Presenting author])

Introduction Machine Performance Check (MPC) is an integrated self-check software application designed for reliable and fast testing for daily beam output using electronic portal imaging device (EPID) before routine treatment starts. The purpose of this study is to evaluate the feasibility of Varian’s Machine Performance Check to track daily beam output as compared with an independent device. Method Before measuring the daily beam outputs, baseline values for MPC and QA BeamChecker Plus (Standard Imaging, Madison, WI) were taken for 6, 10, 15 MV flattening filtered (FF) photon beams, and 6 MV and 10 MV flattening filter-free (FFF) photon beams. Daily beam outputs were measured for 3 months. Percent differences from the baseline value subsequently recorded. Corrections between the data of the MPC and QA BeamChecker Plus were analysed. A very strong correlation was considered as a Pearson’s correlation coefficient (r) [ 0.8; strong for 0.6 \ r \ 0.8; moderate for 0.4 \ r \ 0.6; poor for r \ 0.4. Results All daily beam outputs for MPC and QA BeamChecker Plus were obtained within ±0.5% for most measurements with a maximum deviation of ±1.5%. For 15 MV, the strongest positive correlation (r = 0.81, p \ 0.001) existed between two devices. The correlation between two devices was strong for 10 MV (r = 0.64, p \ 0.001), 6 MV FFF (r = 0.73, p \ 0.001), and 10 MV FFF (r = 0.79, p \ 0.001), and was moderate for 6 MV (r = 0.56, p \ 0.001) (Fig. 1). Conclusion Daily beam output measurements of EPID-based MPC strongly corrected with the widely-used QA BeamChecker Plus for other FF and FFF photon beams excluding the 6 MV. MPC is a useful tool for checking the reliable and fast daily beam output without the set-up of the device for measurement. Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2015R1C1A1A02036358).

MO35 Is Machine Performance Check (MPC) a useful tool in testing of linac radiation isocentre? F. Q. Chen1, T. Moodie1, E. Estoesta1, R. Gajewski1, J. Chojnowski2 1

Crown Princess Mary Cancer Care Centre, Westmead, NSW, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2Mid North Coast Cancer Institute, Coffs Harbour, NSW, Australia. ([email protected]) Introduction New radiotherapy techniques such as VMAT, SRS and SBRT require high accuracy and precision linear accelerators to deliver the radiation dose. Modern linacs are now equipped with software and accessories to test its performance on daily basis. Varian TrueBeam linacs come equipped with the Machine Performance Check (MPC) software package for testing its mechanical, geometric, radiation and dosimetry characteristics. These parameters must be within recommended thresholds in order to deliver accurate, precise and safe radiation treatment. MPC quantitatively determines the radiation isocentre sphere size, which is dependent on the radiation source position in respect to the axis of collimator rotation. This work was to prove that MPC can be relied upon when testing and adjusting radiation source position of linear accelerators. Method Data on radiation isocentre size using MPC on a Varian TrueBeam linac were collected daily and compared with the monthly data from the in-house software LINACQA. At one stage, the MPC radiation source size measured was consistently close to 0.5 mm maximum tolerance value. LINACQA results showed the radiation source position to be 0.3 mm off the central axis. The results of the LINACQA software were verified with the half-beam block test to determine the source position. Both tests using MPC and LINACQA were performed before and after the adjustment to verify the size of the radiation isocentre sphere. Results Before the adjustment the radiation isocentre size determined using MPC and LINACQA were 0.47 mm ± 0.02 mm and 0.52 mm ± 0.03 mm, respectively. After the source position adjustment, the radiation isocentre sizes determined by MPC and LINACQA were 0.36 mm and 0.33 mm, respectively which showed that MPC test is sensitive to the changes of radiation source position. Conclusion MPC is a useful QA feature of TrueBeam linacs and is a reliable tool for testing the size of radiation isocentre. References 1. Klein EE, et al. (2009) Task Group 142 report: quality assurance of medical accelerators. Med Phys. 36(9): 4197–4212. 2. Clivio A, et al. (2015) Evaluation of the machine performance check application for TrueBeam linac. Radiation Oncology 10:97.

MO36 Real-time verification and error detection for MLC tracking deliveries using an electronic portal imaging device Fig. 1 Daily beam outputs measured with MPC and QA BeamChecker Plus for photon beams

B. J. Zwan1,2, E. Colvill3,4, J. Booth3,5, D. J. O’Connor2, P. Keall4, P. B. Greer2,6

123

Australas Phys Eng Sci Med 1

Department of Radiation Oncology, Central Coast Cancer Centre, Gosford, Australia. 2School of Maths and Physical Sciences, University of Newcastle, Newcastle, Australia. ([email protected] [Presenting author]), ([email protected]). 3Northern Sydney Cancer Centre, Royal North Shore Hospital, Sydney, Australia. 4Radiation Physics Laboratory, University of Sydney, Sydney, Australia. ([email protected]), ([email protected]). 5School of Physics, University of Sydney, Sydney, Australia. ([email protected]). 6Department of Radiation Oncology, Calvary Mater Hospital, Newcastle, Australia. ([email protected])

Introduction The added complexity of real-time MLC tracking increases the likelihood of undetected MLC delivery errors. In this work we develop and test a system for real-time delivery verification and error detection for MLC tracking radiotherapy using an electronic portal imaging device (EPID). Method The delivery verification system relies on acquisition and real-time analysis of transmission EPID image frames acquired at 8.4 fps. In-house software was developed to extract the MLC positions from each image frame. Three comparison metrics were used to verify the MLC positions in real-time: (1) field size, (2) field location and, (3) field shape. The delivery verification system was tested for 8 VMAT MLC tracking deliveries (4 prostate and 4 lung) where real patient target motion was reproduced using a Hexamotion motion stage and a Calypso system. Sensitivity and detection delay was quantified for various types of MLC and system errors. Results For both the prostate and lung test deliveries the MLCdefined field size was measured with an accuracy of 1.3 cm2 (1 SD). The field location was measured with an accuracy of 0.6 mm and 0.8 mm (1 SD) for lung and prostate respectively. Field location errors (i.e. tracking in wrong direction) with a magnitude of 3 mm were detected within 0.4 s of occurrence in the X direction and 0.8 s in the Y direction. Systematic MLC gap errors were detected as small as 3 mm. The method was not found to be sensitive to random MLC errors and individual MLC calibration errors up to 5 mm (Fig. 1). Conclusion EPID imaging may be used for independent real-time verification of MLC trajectories during MLC tracking deliveries. Thresholds have been determined for error detection and the EPIDbased system has been shown to be sensitive to a range of delivery errors.

Fig. 1 Examples of real-time verification for (a) prostate motion, (b) lung motion, (c) field shape and (d) field size

123

MO37 Correlation between inter-observer contour variation and simulated treatment outcome within the prostate D. Roach1, Dr. M. G. Jameson2, Dr. J. A. Dowling3, Prof. M. Ebert4, Prof. P. Greer5, S. C. Watt6, Assoc. Prof. L. Holloway7 1

Medical Physics, Liverpool and Macarthur Cancer Therapy Centres, SWS Clinical School, University of New South Wales and the Ingham Institute for Applied Medical Research, Australia. ([email protected] [Presenting author]). 2Medical Physics, Liverpool and Macarthur Cancer Therapy Centres, Ingham Institute for Applied Medical Research, Australia. ([email protected]). 3Australian e-Health Research Centre, CSIRO, Royal Brisbane Hospital, Australia. ([email protected]). 4Radiation Oncology, Sir Charles Gairdner Hospital, School of Physics, University of Western Australia, Australia. ([email protected]). 5Newcastle Calvary Mater Hospital, School of Mathematical and Physical Sciences, University of Newcastle, Australia. ([email protected]). 6Liverpool and Macarthur Cancer Therapy Centres, Australia. ([email protected]). 7 Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute, Faculty of Medicine, SWS Clinical School, University of New South Wales, Institute of Medical Physics, University of Sydney, Centre for Medical Radiation Physics, University of Wollongong. ([email protected]) Introduction Uncertainties and inter-observer variability in contouring remains a significant challenge for quality assurance during radiotherapy treatment [1], however a lack of consensus exists within the literature regarding how to report these discrepancies [2, 3]. Previous studies found that metrics commonly used to parameterise contour variation in lung patients did not show strong correlation with simulated treatment outcome [4, 5]. This study’s goal is to establish the metrics with the strongest correlation to simulated treatment outcome for patients with localised prostate cancer. Method Data was available for 42 patients with localised prostate cancer, with each patient having undergone CT and MRI scanning

Fig. 2 Observed variation in CTV contouring

Australas Phys Eng Sci Med

MO38 Commissioning a novel optical tracking system based on Microsoft Kinect for respiratory monitoring during radiation therapy D. T. Nguyen1, J. Barber2, K. Makhija1, R. O’Brien1, Fiona Hegi1, P. Keall1 1

Faculty of Medicine, The University of Sydney, Australia. ([email protected] [Presenting author]). 2Blacktown Cancer & Haematology Centre, New South Wales, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected])

Fig. 2 Observed correlations in 10 patient cohort (p \ 0.05) prior to radiotherapy treatment. Three expert observers independently contoured the CTV, bladder, and rectum (Fig. 1). Gold standard volumes for these structures were generated using the STAPLE algorithm [6]. VMAT treatment plans (78 Gy to PTV) were generated for the observer and gold standard volumes. Geometric and radiobiological metrics covering those found in the literature for prostate cancer were calculated for these treatment plans, with correlations assessed using Spearman’s rank correlation coefficient. Results Preliminary results investigating 10 patients indicate correlations within the CTV, PTV, and rectum with p \ 0.05 (Fig. 2). No significant correlations have been found for the bladder. Further analysis of the remaining patient data will allow for more significant correlations between metrics to be evaluated. Conclusion Results indicate that volume similarity and differences in centre of mass correlate strongest with variations in simulated treatment outcome for prostate cancer. This information could help future automated registration and atlas methods, allowing them to be guided on metrics based on clinical significance. Further work will also determine the effect of stereotactic radiotherapy treatment on metric assessment. Acknowledgements NHMRC project grant (1077788).

Introduction We report the accuracy of a respiratory monitoring sensor based on optical depth map acquired with a Microsoft Kinect (Microsoft Corp., WA, USA). The depth sensor provides a surrogate optical signal to monitor patient’s respiratory signal during radiation therapy without introducing artefacts. Additionally, the Kinect sensor is cheaper than the commercial optical-based system, RPM system (Varian Medical System, CA, USA). In order to commission the Kinect sensor for respiratory monitoring, we report the outcomes of experiments in which the signals acquired from Kinect-sensor were directly compared with RPM signals acquired at the same time. Method Figure 1 shows the experimental setup of the Kinect sensor with the Quasar phantom. Following recommendations in AAPM taskforce report 147 [1], the following parameters were tested for both systems simultaneously:

Kinect sensor

References 1. Weiss, E. and C.F. Hess, The impact of gross tumor volume (GTV) and clinical target volume (CTV) definition on the total accuracy in radiotherapy. Strahlentherapie und Onkologie, 2003. 179(1): p. 21–30. 2. Fotina, I., et al., Critical discussion of evaluation parameters for inter-observer variability in target definition for radiation therapy. Strahlentherapie und Onkologie, 2012. 188(2): p. 160–167. 3. Jameson, M.G., et al., A review of methods of analysis in contouring studies for radiation oncology. Journal of medical imaging and radiation oncology, 2010. 54(5): p. 401–410. 4. Jameson, M.G., et al., Correlation of contouring variation with modeled outcome for conformal non-small cell lung cancer radiotherapy. Radiotherapy and Oncology, 2014. 112(3): p. 332–336. 5. Vinod, S., et al., Dosimetric implications of the addition of 18 fluorodeoxyglucose-positron emission tomography in CT-based radiotherapy planning for non-small-cell lung cancer. Journal of medical imaging and radiation oncology, 2010. 54(2): p. 152–160. 6. Warfield, S.K., K.H. Zou, and W.M. Wells, Simultaneous truth and performance level estimation (STAPLE): an algorithm for the validation of image segmentation. Medical Imaging, IEEE Transactions on, 2004. 23(7): p. 903–921.

Couch Mount

Quasar phantom IR reflecve box for RPM

Fig. 1 Experimental setup

123

Australas Phys Eng Sci Med Table 1 Amplitude of signals acquired by Kinect and RPM Lighting Period condition (seconds)

Amplitude RPM amplitude (mm) mean ± sd* (mm)

Kinect amplitude mean ± sd* (mm)

Mean Difference (mm)

Full light 5

10

9.91 ± 0.01

9.71 ± 0.68

0.20

Dark

5

10

9.92 ± 0.01

9.49 ± 0.09

0.43

Mid light 5

10

9.92 ± 0.01

9.48 ± 0.06

0.43

Full light 5

2

1.65 ± 0.04

1.72 ± 0.20

0.08

Full light 5

5

4.63 ± 0.01

4.28 ± 0.06

0.34

Full light 3

5

5.18 ± 0.01

4.82 ± 0.11

0.36

Full light 8

5

5.21 ± 0.01

5.11 ± 0.32

0.10

Full light Varying from 3 to 8

5

5.22 ± 0.03

5.09 ± 0.21

0.13

*sd: standard deviation. Mean and standard deviation were calculated based on amplitudes of each cycle in the acquired signal, typically at least 20 cycles for each test condition

• •

Regular motion: varies frequency (20, 12 and 7.5 bpm) and amplitude (2 mm, 5 mm, 10 mm) of respiratory motion. Irregular motion: Abrupt stops and starts of sinusoidal motion, changes in speed of the tracked object.

Additionally, as both systems rely on optical signals, the room lighting conditions can interfere with the system’s performance. Results Results of the tests are summarised Table 1. With both regular sinusoidal motion and irregular motion, the mean amplitude difference between signals acquired by the Kinect and RPM systems was below 0.5 mm in every test. The RPM system performed consistently in every tested lighting condition while the Kinect system worked best in moderately lit rooms with standard deviation decreasing from 0.68 in full light room to 0.06 in mid-light. Conclusion The amplitudes of breathing-like traces acquired at the same time by the Kinect-based system and the RPM system were comparable, with less than 0.5 mm difference between the two systems while measuring both regular and irregular breathing. With proven accuracy, a Kinect system for respiratory monitoring during radiation therapy was implemented and currently in-use in the first lung Kilovoltage Intrafraction Monitoring at Nepean Cancer Care Centre.

Introduction Radiation shielding design for diagnostic radiology depends upon several factors including distance, workload, occupancy, kVp, direction of scatter and dose constraints. Calculations are typically performed in accordance with NCRP 1471 or the BIR Report.2 We evaluated the effects of different methodologies and assumptions on the lead thickness requirements for secondary radiation in a radiographic room. Since lead sheet is available in standard thicknesses, the practical significance of our findings is discussed. Method Lead thickness requirements based on workload were calculated using NCRP and BIR methodologies for an adjacent area at distance 2 m, occupancy 1 and dose constraint 1 mSv per year. Workload expressed in mAmin (NCRP) was converted to Kerma Air Product (BIR) using distances and field sizes from NCRP 147 and assuming a tube output of 60 uGy/mAs at 1 m at 80 kVp. Effects of dose constraint, distance and occupancy on lead thickness requirements were examined using the BIR method. Results Figure 1 shows the relationship between lead thickness and workload. Variations between methodologies are attributed to the choice of scatter angle. The BIR approach excludes leakage but calculates maximum scatter, based on angle and the inverse square law. Interestingly, if using the BIR or NCRP general guidance on workand 240–320 mA min week-1 load (75 mA min week-1 respectively), the required shielding based on available lead thickness would be unaffected by the method used. However, this may not be the case when using local workload data. Figure 2 shows that occupancy, distance and dose constraint have a significant effect on required lead thickness. Because dose constraints vary with state legislation, a workload of 100 mA min week-1 could require three different lead thicknesses!

References 1. T Willoughby, J Lehmann, A Bencomo, S K Jani, L Santanam, A Sethi, T D Solberg, W A Tome´, Timothy J. Waldron, Quality assurance for nonradiographic radiotherapy localization and positioning systems: Report of Task Group 147, Med. Phys. 39 (4), April 2012. Acknowledgements The authors acknowledge the support of Nepean Cancer Care Centre (Kingswood, New South Wales, Australia) for this work. P Keall is funded by an NHMRC fellowship.

MO39 Evaluation of the factors affecting radiation shielding for diagnostic radiology N. Mirjalili1, N. Hille2, J. L. Diffey2 1

Biomedical Technology Services, The Prince Charles Hospital, Brisbane, QLD. ([email protected] [Presenting author]). 2Hunter New England Imaging, John Hunter Hospital, New Lambton, NSW. ([email protected]), ([email protected])

123

Conclusion Dose constraint, distance, workload and occupancy have a far greater effect on required lead thickness than the methodology used. Data assumptions for shielding calculations should therefore be as accurate as possible. Acknowledgements We would like to thank Lucy Cartwright for developing and sharing a spreadsheet for NCRP 147 shielding calculations.

Australas Phys Eng Sci Med References 1. NCRP Report No. 147. Structural Shielding Design for Medical X-ray Imaging Facilities. Recommendations of the National Council on Radiation Protection and Measurements. 2004 2. Radiology Shielding for Diagnostic Radiology, 2nd Edition. Report of a BIR working party. 2012

MO40 Experimental verification of the magnetic field of a new apparatus for performing experiments in MRI-Linac dosimetry

Fig. 2 Left measured magnetic field strength at central axis of magnet, Centre modeled magnetic field strength, Right comparison of measured and modeled field

T. J. Causer1, B. M. Oborn2, M. Gargett3, A. B. Rosenfeld1, P. E. Metcalfe1 1

Centre of Medical Radiation Physics (CMRP), University of Wollongong, NSW 2522, Australia. ([email protected]). 2 Illawarra Cancer Care Centre (ICCC), Wollongong, NSW 2522, Australia. ([email protected]). 3Royal North Shore Hospital, Sydney, NSW 2065, Australia. ([email protected]), ([email protected]), ([email protected] [Presenting author]) Introduction MRI-guided radiotherapy is an emerging field of radiotherapy and has led to the development of treatment units which combine MRI scanners with radiation therapy sources. To perform experiments related to MRI-guided radiotherapy, without the cost of such units, a variable magnetic field strength and portable specialised magnetic field apparatus has been created using strong permanent magnets. We present the performance of the devices’ magnetic field with both measurement and simulation results. Method An automated 2 Dimensional scanning system has been designed and constructed using the combination of ThorlabsTM LTS150 stages, InventablesTM X-carve components, 3D printed mounts and a MAGSYSTM HGM09s gaussmeter shown in Fig. 1. The measurement probe is suspended via an acrylic rod at the central plane between the magnets focusing cones. The field between and surrounding the focusing cones is scanned in a raster fashion at 2 mm intervals. Results The measurement of the magnetic field matches the modelled prediction to within +/-2.1% across the plane of measurement shown in Fig. 2. In an arrangement designed for detector studies, the magnetic field between the focusing cones (3 9 3 9 3 cm3) was measured as 1.20 ± 0.02 T.

Fig. 1 Magnetic field apparatus, modeled magnetic field magnitude shown around the iron yoke and focusing cones. As indicated, a radiation beam can be incident from either parallel (A) or perpendicular (B) to the magnetic field direction

Conclusion The portable specialised magnetic field apparatus offers a high degree of flexibility both for magnetic field strengths and radiation beam orientations. The automated scanning system provides means of accurate experimental verification of the apparatus, this is important for the planned detector and cell studies.

MO41 Usage of energy weighted interpolation method to determine Cs137 calibration factor for survey meter calibration M. Ibrahim1, T. Ravichander1 1

Department of Medical Physics &Radiation Engineering, The Canberra Hospital, ACT, Australia. ([email protected] [Presenting author])

Introduction All the survey meters used in radiation protection measurements must be calibrated every two years to comply with the regulatory requirements. The Medical Physics Department at Canberra Hospital has a Cs137 calibration source to calibrate survey meters. It is important to use a detector so that the calibration can be traceable back to a Primary Standard Laboratory (PSL). We present usage of energy weighted interpolation method to determine Cs137 calibration factor for the reference chamber. A local reference Farmer ionization chamber (NE 2571) traceable to ARPANSA was used as a reference and compared to survey meter response. Method The weighted energy interpolation method used to derive Nk for Ir192 Brachytherapy source [1] was used to determine Airkerma calibration factor (Nk). Nk of the Farmer chamber for Cs137 radiation was derived from Nk of Co60 and 250 KVp x-ray. Airkerma measurements at closer distances to the source were determined using the Farmer chamber. Dose equivalent rates were derived for the corresponding distances [2]. Dose equivalent rate at farther distances from the source were measured directly using a calibrated survey meter (Victoreen Fluke 451). Dose equivalent rate Vs 1/distance2 was plotted in single graph for both set of measured values. Plot was fitted to a straight line. Results The fit (Fig. 1) shows that dose equivalent rate measured by both equipments are agreed to a unique trend with the change of distances. The two independent methods are agreeing with each other thus confirming the accuracy of the measurements. Conclusion Energy weighed interpolation method can be used to derive Nk for a Cs137 source. Usage of ARPANSA calibrated farmer chamber confirms the tractability of calibration to a Primary Standard Laboratory.

123

Australas Phys Eng Sci Med

Fig. 1 Dose equivalent rate vs. 1/d2 References 1. Calibration of photon and beta ray sources used in brachytherapy. International Atomic Energy Agency Vienna, 2002 iaea-tecdoc-1274. 2. Calibration of radiation protection monitoring instruments. International Atomic Energy Agency Vienna, 2000 Safety Reports Series no. 16

MO42 Evaluation of a PET radiation shielding plan with MrVoxel E. McKay1 1

Dept. Nuclear Medicine, St. George Hospital, Australia. ([email protected] [Presenting author])

Introduction Evaluation of shielding plans for PET facilities is complicated by the need to consider the potential additive effect of separate sources, as well as irradiation of areas above, below and outside the facility. This work aims to facilitate rapid shielding plan evaluation by automating the calculation of distance and attenuation between source/target pairs, to the point where it is possible to predict dose rates at the centimeter scale if so desired. Method In 2010, Smith described software for calculating dose rates in a plane around shielded, isotropic radiation sources. This concept has been re-implemented using Java software developed by the author (MrVoxel, 2003). The software imports a floorplan of one or more levels in PNG format. Distance calibration is applied, then source, target and shielding locations are marked on each level. Each source is assigned a dose constant (lSv/week at 1 m) and each target is assigned a dose constraint (lSv/week). Each shield is assigned a thickness and material type, used to parameterise an attenuation function (BIR 2012). The software produces 2 outputs: a colour overlay on the floorplan indicating weekly dose at the pixel level and a collection of tables, one set for each source, detailing distance to each target and the thickness of each material that must be penetrated. These tables are exported to a spreadsheet where the dose to each target from each source is calculated from these inputs. Results The software was used to iteratively evaluate the shielding design for a new PET suite to be constructed at St. George Hospital in 2016. Shielding plans developed using the software were assessed and approved by 2 separate CREs at different stages of the design process.

123

Conclusion Software for evaluating radiation shielding plans for multiple isotropic sources has been developed and used to optimise the shielding design for a new PET suite. References 1. BIR 2012: Sutton D G, Martin C J, Williams J R, Peet D J, ‘Radiation Shielding for Diagnostic Radiology, 2nd edition’, British Institute of Radiology, 2012. 2. Smith 2010: Smith J, Bartlett M, ‘‘Nuclear Medicine Shielding Program’’’, 40th Annual Meeting of the Australia and New Zealand Society of Nuclear Medicine, Auckland, New Zealand, 2010. 3. MrVoxel 2003: McKay E, ‘‘A Software Tool for Specifying Voxel Models for Dosimetry Estimation’’, Cancer Biotherapy & Radiopharmaceuticals 18:3, 2003.

MO43 Solid-state detector spatial response measured with 0.1 mm resolution D. J. Butler1, J. Lehmann1, T. Beveridge2, C. P. Oliver1, T. E. Wright1, A. W. Stevenson3, J. Livingstone3, J. Crosbie4 1

Radiotherapy Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected] [Presenting author]), ([email protected]). 2Medical Imaging Section, ARPANSA, 619 Lower Plenty Rd Yallambie VIC. ([email protected]), ([email protected]), ([email protected]). 3IMBL, Australian Synchrotron, Clayton VIC. ([email protected]), ([email protected]). 4RMIT University, Melbourne. ([email protected]) Introduction The Australian Synchrotron produces a beam of kilovoltage photons bright enough to collimate to smaller than 0.1 mm and still produce a measurable signal in common radiotherapy detectors. This enables the response of detectors to be measured with high spatial resolution. Such response maps can be used to understand detector behaviour, and previously we have reported the response of common ionisation chambers [1]. In this work we apply the technique to several solid-state detectors, with the goal of confirming the

Australas Phys Eng Sci Med ([email protected]). 6Sydney Medical School, University of Sydney, Australia. ([email protected]), ([email protected]), ([email protected])

Fig. 1 Spatial response scans for two PTW diodes – the ‘‘Diode P’’ and ‘‘Diode SRS’’. The active area is clearly seen as the blue spot in these false-colour maps

diameter of the active area, a parameter important to determine volume averaging corrections. Method Four solid-state radiotherapy detectors (PTW 60017 Diode E, PTW 60016 Diode P, PTW 60018 Diode SRS and PTW 60019 microDiamond) were scanned in the end-on orientation using a beam of average energy 95 keV and diameter 0.1 mm. The current from the detector was measured as a function of position to create a response map. Results The results for two diodes are shown in Fig. 1. The scans are 5 mm 9 5 mm and consist of 51 horizontal lines each containing 140 points. Both detectors shown in Fig. 1 have the same nominal active diameter (1.2 mm), although the SRS diode has ten times the volume. The scans indicate the active area agrees well with this value. Higher resolution scans were performed and additional results will be available shortly, including for the microDiamond and Diode E detectors. Conclusion We have scanned four detectors and determined their spatial response for a kV beam. We have shown that the measured active area agrees well with the value in the specifications.

Introduction Longitudinal magnetic fields have a focussing effect on electrons and may narrow the penumbra[1–3]; this can tighten lateral spread of secondary electrons in air cavities, including lung tissue[3]. The MRI-linac consisted of a 1.5 T Sonata MRI and Varian industrial linatron (nominal 4MV and 6MV energies). Profiles on Gafchromic EBT3 film were used to investigate differences between penumbra in solid water and solid lung, and under 0 T and 1.5 T conditions. Method RMI solid water and solid lung was positioned in the bore centre, perpendicular to the beam’s horizontal central axis (Fig. 1). Film was positioned at solid water depths corresponding to entrance/ exit positions of solid lung of various thicknesses. An equivalent dose was given to each film. MLC-defined field sizes were set to 3 9 3 cm2 and 10 9 10 cm2 at the linatron’s extended SSD of 276.9 cm, corresponding to 1 9 1 cm2 and 3.2 9 3.2 cm2 at 100 cm SSD. From normalised profiles, 80–20% penumbral width was determined. Results In 0 T conditions, solid lung profiles had a larger penumbra than solid water in 74% of cases by a maximum of 1.8 mm. Under 1.5 T conditions, penumbra narrowing (Fig. 2) was apparent for both

References 1. High spatial resolution dosimetric response maps for radiotherapy ionization chambers measured using kilovoltage synchrotron radiation, D J Butler, A W Stevenson, T E Wright, P D Harty, J Lehmann, J Livingstone and J C Crosbie, Phys. Med. Biol. 60 (2015) 8625–8641 Fig. 1 Experimental set-up, (a) Solid water, (b) Solid lung

MO44 Comparison of solid water and solid lung GafchromicÒ film profiles in 0 T and 1.5 T longitudinal magnetic field of the Australian MRI-Linac S. J. Alnaghy1,2, J. Begg2,3,4, T. Causer1,2, T. Alharthi1,4, L. Glaubes4, A. George2,3, L. Holloway1,2,3,4,5,6, A. B. Rosenfeld1, P. Metcalfe1,2 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. 2Ingham Institute for Applied Medical Research, Sydney, Australia. ([email protected] [Presenting author]). 3 Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Sydney, Australia. 4South Western Sydney Clinical School, University of New South Wales, Australia. ([email protected]), ([email protected]). 4 Institute of Medical Physics, University of Sydney, Australia. ([email protected]), ([email protected]),

Fig. 2 Entrance and exit profiles

123

Australas Phys Eng Sci Med field sizes, all depths of solid water, and all solid lung thicknesses; penumbral differences between 0 T and 1.5 T ranged from 1.0 to 4.4 mm. Conclusion The solid lung profiles were consistently more affected by the magnetic field, although only by a maximum of 1.7 mm in penumbral tightening. Future work will repeat measurements in a 1 T longitudinal magnetic field from an Agilent magnet; the effect of the magnetic field on penumbra is expected to be significant but slightly less than at 1.5 T. References 1. Oborn BM, Metcalfe PE, Butson MJ, Rosenfeld AB, Keall PJ (2012) Electron contamination modeling and skin dose in 6 MV longitudinal field MRIgRT: Impact of the MRI and MRI fringe field. Med Phys 39(2):874–90. doi: 10.1118/1.3676181 2. Bielajew AF (1993) The effect of strong longitudinal magnetic fields on dose deposition from electron and photon beams. Med Phys 20(4):1171–1179. doi: 10.1118/1.597149 3. Naqvi SA, Li XA, Ramahi SW, Chu JC, Ye SJ (2001) Reducing loss in lateral charged-particle equilibrium due to air cavities present in x-ray irradiated media by using longitudinal magnetic fields. Med Phys 28(4):603–611. doi: 10.1118/1.1357816

MO45 Beam perturbation characteristic of monolithic Magic Plate 512 (MP512) K. Utitsarn1, Z. A. Alrowaili1, N. Stansook1, M. Lerch1, M. Carolan1,2, M. Petasecca1, A. Rosenfeld1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), 2Illawarra Cancer Care Centre, Wollongong Hospital, Wollongong, Australia. ([email protected]), ([email protected]), ([email protected]) Introduction Studies show the increase in surface dose when transmission-type detectors are used for in-field measurements [1–3]. The purpose of this study is to investigate the influence of a monolithic 2D diode array MP512 on the surface dose when use in transmission mode as a function of different field sizes and distances from phantom. Method The influence of the MP512 on surface dose when operated in transmission mode was evaluated by a Markus ionization chamber (SSD 100 cm). The MP512 placed on a movable stand holder was positioned at various distances (4.5 cm to 18 cm) from the solid water phantom surface (dsw). 6 MV photons were delivered in a range of irradiation field sizes (IFSs) from 5 9 5 cm2 to 10 9 10 cm2. Surface dose measurements were taken with and without the MP512 in position. Additional measurements were taken with the MP512 faced up and faced down when in position, to investigate the influence of the MP512 printed circuit board (PCB) on the surface dose. Results For all distances measured from the phantom surface and all IFSs the results indicate that the MP512 led to surface increasing in arrange 4.5% \ DDs \ 25% when in place. For dsw 18 cm and above the surface dose increasing is \5% and not measurable for fields 10 9 10 cm2 and 5 9 5 cm2 respectively. Skin dose perturbation effect is a combination of absorption of electron generated in air and Compton scattered from MP. The effect of the MP512 on the surface dose change in position face up compared to face down is negligible (\1%) for all IFSs.

123

Conclusion We have demonstrated that the monolithic MP512 positioned in air between the Linac and the patient produces minimal perturbation of the surface dose while providing high spatial resolution (2–4 mm depends on dsw) real time dosimetry on a patient. The MP512 is an excellent candidate for real-time dose reconstruction of delivered small size fields. Reference 1. Alrowaili ZA, Lerch MLF, Petasecca M, Carolan MG, Metcalfe PE, Rosenfeld AB (2016) Beam perturbation characteristics of a 2D transmission silicon diode array, magic plate. Journal of Applied Clinical Medical Physics 17 (2):85–98 2. Nakaguchi Y, Maruyama M, Araki F, Saiga S (2012) Dose verification of IMRT by use of a COMPASS transmission detector. Radiological Physics and Technology 5 (1):63–70. 3. Venkataraman S, Malkoske KE, Jensen M, Nakonechny KD, Asuni G, McCurdy BMC (2009) The influence of a novel transmission detector on 6 MV x-ray beam characteristics. Physics In Medicine And Biology 54 (10):3173–3183.

MO46 Angular dependence of a 2D monolithic array dosimeter and correction required for small field dosimetry N. Stansook1, M. Petasecca1, K. Utitsarn1, M. K. Newall1, M. Duncan1, K. Nitschke2, M. Carolan3, P. Metcalfe1, A. Rosenfeld1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected], [email protected]), ([email protected]), ([email protected]). 2Illawarra Heath Medical Research Institute, Wollongong, Australia. ([email protected]). 3Illawarra Cancer Care Centre – Wollongong Hospital, Wollongong, Australia. ([email protected]), ([email protected]), ([email protected])

Introduction Angular dependence of detectors is one of the factors which affects the accuracy of dosimetry quality assurance for modulated arc radiotherapy. The purposes of this study are to determine the angular dependence of a 2D monolithic silicon array dosimeter MagicPlate-512 (MP512) and provide a method for angular response correction for stereotactic radiotherapy. Method The MP512 is a monolithic silicon detector of 52 9 52 mm2 with 512 pixels and a spatial resolution of 2 mm. The detector was placed inside a cylindrical phantom (Fig. 1) and aligned at isocenter. The response of MP512 was measured at 15 angle increments from 0 to 180. Field sizes were various from 1 9 1 cm2 to 10 9 10 cm2 for 6 MV and 10 MV. The correction factors were calculated from the ratio of MP512 response to EBT3 film doses as a function of incident angle (h) and were normalized at zero incident angle. Beam profiles of the corrected MP512 were compared with EBT3 for verification of the effectiveness of the method adopted. Results MP512 shows an uncorrected angular response of -15.5% and -18.5% at 90 for 10 MV and 6 MV photon beams, respectively. The study demonstrates that the variation of angular response is not affected by the field size. In contrast, the angular response is sensitive to the energy (Fig. 2) and an energy specific angular correction factor should be adopted. The beam profiles of corrected MP512 and EBT3 film shows an agreement within ±2% for all field sizes and angles.

Australas Phys Eng Sci Med mathematical concept of percolation was investigated as a tool for accounting for ad hoc spatial dose distributions. Method Application of percolation to response modelling was undertaken numerically. Functional sub units (FSUs) were represented by grid elements and were allocated a probability of response according to a simulated local dose delivery and a linear-quadratic response model. Cluster formation was examined by comparing the response with a ‘percolation lattice’ comprised of an equivalent grid with uniformly distributed probabilities. A weighting function, g(s), was hypothesised that leads P to complication incidence when a complication measure, C ¼ s s  ms  gðsÞ, (with ms the multiplicity of clusters of size s) exceeds a threshold, giving a measure of NTCP: P s  ms  gðsÞ ð1Þ NTCP ¼ s smax  gðsmax Þ Fig. 1 Schematic diagram of the angular response measurement

Fig. 2 The relative response of MP512 as a function of angle, (a) an open field size of 10 9 10 cm2 for 6 MV and 10 MV photons, (b) the various field size for 6 MV photons

Conclusion The major cause of intrinsic angular response of MP512 is its packaging and size of the silicon crystal. The correction factor obtained for a 10 9 10 cm2 field can be utilised for other field size and proves the suitability of MP512 to be used as a high spatial resolution dosimeter for arc modulated stereotactic radiotherapy. References 1. A H. Aldosari, M. Petasecca, A. Espinoza, et al. (2014) A two dimensional silicon detectors array for quality assurance in stereotactic radiotherapy: MagicPlate-512. Med. Phys 41(9): 091707-10.

MO47 Percolation as the basis for modelling critical elements in radiation response N. Gale1, M. House2, M. A. Ebert2,3

Serial (simulated as fibres) and parallel FSU arrangements were simulated with several hypothetical dose distributions intended to test the predictions of the percolation model. Results Simulations yielded functional forms between dose and response that mimic generic clinically-observed patterns of response. Responses were compared to a conventional dose-volume based logistic function combined with a volume-reduction measure. Unlike the logistic function, the percolation responses were characterised by a critical dose above which NTCP increased. A quadratic weighting function greatly increased NTCP for dose hotspots relative to a linear function. Conclusion The percolation model was capable of qualitatively replicating dose-response relationships observed clinically, whilst accommodating spatial relationships in FSU response by examining the formation of clusters of all sizes. The resulting response relationships are non-degenerate. This form of response model opens many opportunities to account for, investigate and capitalise on ad hoc spatial dose patterns and can accommodate spatial correlation with anatomic, functional and molecular distributions.

MO48 Finite Element Model of a gridded electron gun: Characterisation in magnetic fields and implications for MRI-Linac therapy Brendan Whelan1, Dragos Constantin2, Lois Holloway3, Brad Oborn4, Magdalena Bazalova-Carter5, Rebecca Fahrig2, Paul Keall1 1

Radiation Physics Laboratory, University of Sydney, Sydney, Australia. ([email protected]) [Presenting author]. 2 Stanford University, CA, USA. ([email protected]). 3 Ingham Institute for Applied Medical Research, Liverpool, Australia. ([email protected]). 4Illawarra Cancer Care Centre, Illawarra Hospital, Wollongong, Australia. ([email protected]). 5 XCITE Lab, University of Victoria, Victoria, Canada. ([email protected]), ([email protected]), ([email protected])

1

School of Mathematics and Statistics, University of Western Australia, Western Australia. ([email protected]). 2 School of Physics, University of Western Australia, Western Australia. ([email protected]). 3Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia. ([email protected] [Presenting author]) Introduction Contemporary radiotherapy permits considerable flexibility in sculpting radiation dose to anatomy. Models which relate highly non-uniform distributions to outcome, taking account of the spatial dependence of response, are lacking. In this study, the

Introduction MRI-Linac therapy holds the promise of greatly improved cancer treatment outcomes by combining the exquisite soft tissue contrast, high temporal resolution, and functional imaging capabilities of MRI with the established therapeutic gains of radiotherapy. A key challenge is ensuring the linac functions within the MRI scanner fringe fields. The most sensitive component of the linac is the electron gun. In this work, we develop a finite element (FEM) model of a gridded medical electron gun, and characterise its performance in magnetic fields. Method Geometry of a Varian gridded electron gun was measured using 3D laser scanning and digital calliper measurements. High voltage (HV),

123

Australas Phys Eng Sci Med grid voltage (GV) and emission current were measured from 6 dosematched TrueBeam linacs for the 6X, 10X, and 15X beam modes. Based on these measurements, an FEM model was developed using Opera/ SCALA (Vector Fields, UK). The mean operating conditions were used to simulate each mode. The model was solved in constant magnetic fields between 0–200 G (in-line orientation) and 0–35 G (perpendicular) and injection current was monitored.

Fig. 1 RMG 4DCT can improve lung cancer imaging

Results The measured HV, GV, and current were 15 ± 0.03 kV, 10 ± 0.8 kV and 11k ± 0.03 kV; 93 ± 7 V, 41 ± 3 V and 70 ± 6 V; 327 ± 27 mA, 129 ± 10 mA and 214 ± 19 mA for the 6X, 10X, and 15X modes respectively. The error in the simulated current of each mode was 3%, 6%, and 3%. For each mode, 50% beam loss occurred at 114, 96, and 97 G for in line fields, and 21, 13, and 14 G for perpendicular. Sensitivity to external magnetic fields is a function of both geometry and HV setting.

Conclusion An accurate FEM model of a clinically used gridded electron gun was developed (to the best of our knowledge the first such model presented in the literature) and its sensitivity to external magnetic fields characterised. The results hold important implications for the linac operation within an MRI environment, as needed for MRI-Linac therapy.

O116 In silico performance characterisation of respiratory motion guided four dimensional computed tomography (RMG 4DCT)

Method Our RMG 4DCT software prototype used an in-house C# library to playback Varian RPM logs and perform real-time phase estimation. Both RMG and clinical 4DCT scans were simulated for [100 hours of breathing motion (523 RPM traces) for 20 lung cancer patients. RMG 4DCT used prospective regularity gating windows based on a 60-second prescan training session, whereas clinical 4DCT used no gating. We applied paired t-tests to determine significance in terms of: (i) breathing irregularities during imaging (i.e. root mean square error of the RPM signal), (ii) the time spent imaging (a surrogate for imaging dose) and (iii) total scan time. Scan parameters were chosen for clinical relevance (10 phase bins, 20 couch positions, 0.25 s gantry rotation time). Results On average, RMG 4DCT lead to a 13% reduction in breathing irregularities compared to clinical 4DCT (p \ 0.001). Imaging time was also significantly reduced (31%), since the prospective gating eliminates oversampling of 4DCT slices. RMG 4DCT exhibited a 33% increase in total scan time, but this was small in absolute terms (1–2 min). Conclusion RMG 4DCT represents the ‘‘next generation’’ of 4DCT, with the potential to reduce image artifacts and imaging dose at a small cost in scan time. Future work will investigate the real-world benefits using an experimental implementation. References 1. Yamamoto, T., et al., Retrospective Analysis of Artifacts in FourDimensional CT Images of 50 Abdominal and Thoracic Radiotherapy Patients. Int J Radiat Oncol Biol Phys; 72:1250–1258, 2008. 2. Keall, P.J., et al., Respiratory regularity gated 4D CT acquisition: concepts and proof of principle. Australas Phys Eng Sci Med;30:211–220, 2007. 3. Bernatowicz, K., et al., Quantifying the impact of respiratorygated 4D CT acquisition on thoracic image quality: A digital phantom study. Med Phys;42:324–334, 2015. Acknowledgements This work was supported by a Cancer Institute NSW Early Career Fellowship and an NHMRC Australia Fellowship.

J. Kipritidis1, S. Martin1, R. T. O’Brien1, P. J. Keall1 1

Sydney Medical School, University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected])

Introduction Currently 90% of four-dimensional CT (4DCT) scans suffer image errors1 that limit the accuracy of tumour motion management in lung cancer radiotherapy. The main culprit is irregular breathing, which is unaccounted for by clinical scanners2,3. To solve this, we have developed respiratory motion guided (RMG) 4DCT, which pauses imaging during irregular breathing. This work performs the first investigations of RMG 4DCT using an advanced software prototype, testing the hypothesis that it can reduce breathing irregularities during imaging compared to clinical 4DCT (Fig. 1).

123

O117 Potential of continuous positive airway pressure for reduced organ at risk dose in external beam radiotherapy N. Hardcastle1, V. Caillet1,2, C. Crasta1, K. Szymura1, B. Harris3, C. Haddad1, P. Keall2, J. Booth1 1

Northern Sydney Cancer Centre, Royal North Shore Hospital, NSW, Australia. ([email protected] [Presenting author]). 2Radiation Physics Laboratory, University of Sydney, NSW, Australia. ([email protected]), ([email protected]), ([email protected]). 3Department of Respiratory

Australas Phys Eng Sci Med Medicine, Royal North Shore Hospital, NSW, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction Respiratory motion can result in either increased or decreased organ-at-risk doses in radiotherapy; tissues surrounding a moving tumour receive a higher dose but in deep inspiration breath hold (DIBH), the heart potentially receives a lower dose. Continuous Positive Airway Pressure (CPAP) maintains a continuous full lung over which the patient respires. The effect of CPAP on thoracic target and heart motion has been investigated for a small cohort of patients. Method A sub-study of the LIGHT SABR trial investigates CPAP on implanted electromagnetic beacon motion. For each patient, CPAP at 15 cm H20 is applied at simulation and treatment sessions. A low dose 4DCT is acquired with CPAP; the tumour as well as heart position and lung volume are compared with that in the treatment planning 4DCT. The beacon motion with CPAP was measured at the end of each treatment session and compared with that from no CPAP. Patient tolerance was recorded. Results Three patients (estimated 6 more in 2016) have undergone CPAP. CPAP was tolerated by all three patients at simulation, and 2/3 patients at treatment. The third patient did not participate in CPAP at treatment due to deteriorating lung function. Exhale lung volume

increased by 16%, 43% and 36% for patients 1, 2 and 3 respectively (Fig. 1). The minimum distance from the heart to chest wall at the position of 4th rib reduced by 1.9 cm, 3.4 cm and 2.0 cm for patients 1, 2, and 3 respectively. CPAP reduced beacon motion at treatment for both patients by up to 50%; this was not predicted by simulation 4DCT (Fig. 2). Conclusion CPAP shows potential for reduction of tumour motion due to respiration as well as for moving the heart inferiorly, similar to DIBH. CPAP is well tolerated and warrants further investigation as a motion management tool.

O118 Commissioning and quality assurance of the first clinical implementation of a kilovoltage image guided MLC tracking for VMAT D. T. Nguyen1, J. T. Booth2, R O’Brien1, V. Caillet2, M. Gargett2, P. Poulsen3, P. Keall1 1

Faculty of Medicine, The University of Sydney, Australia. ([email protected]) [Presenting author]. 2Northern Sydney Cancer Centre, Royal North Shore Hospital, Australia. ([email protected]), ([email protected]). 3 Department of Oncology, Aarhus University Hospital, Denmark. ([email protected]), ([email protected]) Introduction Kilovoltage Intrafraction Monitoring (KIM) and MultiLeaf Collimator (MLC) tracking are emerging technologies that ensure radiotherapy finds the target and hits the target. Each technology has been clinically implemented separately for prostate

Table 1 Results of KIM-MLC tests A Commissioning and QA tests criteria and results for KIM-MLC adaptation system

Fig. 1 Patient 2 free breathing and CPAP exhale phase CTs showing increased lung volume, lower heart position and smaller tumour motion envelope (green contour)

Test

Criteria

Results

Coordinate system

KIM-MLC coordinate matches linac

Pass

Dynamic Pass Latency

\1 mm error in each direction

localisation (Table 1B) \1 s

0.32 second. Pass

Interruption/ Generate beam-hold if any Pass. All 3 conditions Interlock condition matched resulted in beam-hold Dosimetry

[98% pixels passing gamma criteria 2%/ 2 mm

Pass

B. Dynamic localisation results of KIM-MLC system Mean LR Stable motion

Fig. 2 Reduction in beacon centre-of-mass position with CPAP at treatment (dots = mean, lines = range, bar = st. dev.). FB = free breathing

-0.08

Standard Deviation SI

AP

LR

SI

AP

0.06

0.1

0.25

0.29

0.23

Continuous motion 0.8 -0.02

0.41 0.38

0.2

0.53

Persistent motion

0.3

0.13

0.44

0.05

0.01

0.23

Transient excursion 0.04 -0.011 0.06 0.55

0.44

0.68

Erratic excursion

0.15

0.89

0.45 -0.15 -0.25 0.47

123

Australas Phys Eng Sci Med radiotherapy [1,2]. Both technologies leverage standard linac hardware to enable highly accurate real-time adaptive radiotherapy. KIMguided MLC tracking will be first implemented within the SPARK trial (TROG 15.01). Commissioning and TROG credentialing present novel challenges with neither KIM nor MLC tracking listed on the ARTG. We present commissioning of the first KIM-guided MLC adaptation for VMAT treatments. Method Commissioning and QA tests have previously been developed for KIM gating and MLC tracking from Failure Mode and Effects Analysis (FMEA) [3]. Process mapping and iterative software development of KIM-guided MLC tracking was performed with a revisited FMEA to produce a specific set commissioning tests and safety interlocks. Commissioning tests included: Coordinate system test, dynamic motion accuracy test [4], latency test, interruption/Interlock test and film dosimetry. A new beam hold criteria was implemented to hold the beam in the presence of segmentation errors that undermine the accuracy of KIM. Results The KIM-MLC tracking correctly determined the coordinate system at static couch positions and accurately determined prostate motion simulated with anthropomorphic phantom using the HexaMotion (Table 1). The system latency measured 320 ms. Each of the beam-hold triggering conditions successfully initiated beamhold interlocks to the linac. Figure 1 shows film dosimetry

comparison between a static delivery and MLC tracking adapted to a persistent motion of 8 mm and a combined motion trajectory featuring a range of clinically measured prostate motions. Both dynamic adaptive deliveries passed gamma tests with a 2%/2 mm criterion with 98.3% and 99.7% pixels for persistent motion trajectory and combined dynamic trajectory, respectively. Without KIM-MLC adaptation, the same film dosimetry tests showed 84.3% and 97.1% pixels passing the same gamma criterion. Conclusion An iterative FMEA and software development process was used to safely implement a user-written software that interfaces directly into the linac for real-time adaptive radiotherapy. Five commissioning and QA tests are applied to support clinical implementation of the world first KIM-guided MLC adaptation system. Acknowledgements SPARK is administered by TROG and is funded by a Cancer Australia grant. P Keall is supported by an NHMRC fellowship. References 1. PJ Keall, J Ng, R O’Brien, E Colvill, CY Huang, P R Poulsen, W Fledelius, P Juneja, E Simpson, L Bell, F Alfieri, T Eade, A Kneebone , J T Booth, The first clinical treatment with kilovoltage intrafraction monitoring (KIM): a real-time image guidance method, Med Phys. 42(1):354–8, 2015. 2. P J Keall, E Colvill, R O’Brien, J A Ng, P R Poulsen, Eade, A Kneebone, J T Booth, The first clinical implementation of electromagnetic transponder-guided MLC tracking, Med. Phys. 41(2). 3. A Sawant, S Dieterich, M Svatos, and P Keall, Failure mode and effect analysis-based quality assurance for dynamic MLC tracking systems, Med. Phys. 37, 6466–6479. 4. J A Ng, J T Booth, R O’Brien, E Colvill, C Y Huang, P R Poulsen, PJ Keall, Quality assurance for the clinical implementation of kilovoltage intrafraction monitoring for prostate cancer VMAT, Med Phys. 2014 Nov; 41(11).

O119 Quantifying the accuracy and precision of realtime 6degree-of-freedom kilovoltage intrafraction monitoring (KIM) of tumour motion J.-H. Kim1, D. T. Nguyen1, T. Fuangrod2, C.-Y. Huang1, V. Caillet1, R. O’Brien1, P. Poulsen3, J. Booth4, P. Keall1 1

Faculty of Medicine, The University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]). 2Department of Radiation Oncology, Calvary Mater Newcastle Hospital, Newcastle, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3Department of Oncology, Aarhus University Hospital, Denmark. ([email protected]) 4Department of Medical Physics, Royal North Shore Hospital, Australia. ([email protected]), ([email protected])

Fig. 1 Dose distribution delivered when phantom was static (A), moved with persistent motion with KIM-MLC adaptation (B1) and without adaptation (B2) and moved with combinatory dynamic motion with adaptation (C1) and without adaptation (C2). Motion input traces are shown in C3: persistent motion and C3: combined dynamic motion

123

Introduction Tracking tumour centroid translation may not be sufficient and rotation can impact significantly on delivered dose particularly depending on tumour shape and isocentre placement. As a prelude to 6degrees-of-freedom (6DoF) tumour motion adaptive radiotherapy, real-time monitoring of 6DoF intrafraction tumour motion has recently been clinically implemented in the TROG 15.01

Australas Phys Eng Sci Med SPARK trial and pilot [1]. We evaluate the accuracy and precision of the 6DoF motion implementation. Method Real-time 6DoF tumour motion measurements with KIM was enabled by implementing an iterative closest point algorithm [2,3]. The 6DoF KIM motion measurements were compared against the ground truth motion derived from kV/MV triangulation for a range of lung and prostate tumour motion trajectories [3] and various static poses. Simultaneous kV and MV-images of a moving custom-built phantom with three gold markers were acquired during the delivery of an arc field (-140 to 140) of 6MV x-rays (Trilogy, Varian). The motion was applied with a modified 5DoF motion platform (HexaMotionTM) and a rotating treatment couch. KIM provided real-time measurements of the phantom motion, and the corresponding motion was retrospectively derived from kV/MV triangulated markers’ positions using a closed-form least squares method [4]. The accuracy and precision of 6DoF KIM were calculated as the mean and s.d. differences for each DoF, respectively. Results Figure 1 compares time-series 6DoF lung tumour motion trajectories estimated by KIM in real-time and by kV/MV triangulation with transient rotation motion of up to +6 about SI-axis. Table 1 shows the mean (SD) differences in each DoF, all of which were within 1.0 and 1.0 mm, except for rSI of prostate trace (s.d. of 1.3). Conclusion The accuracy and precision of real-time tumour motion measurements using 6DoF KIM were evaluated for a series of realistic tumour motion trajectories. The results demonstrate that KIM is able to provide 6DoF motion within sub-degree and sub-millimetre accuracy.

Table 1 The mean (s.d.) differences between KIM and kV/MV triangulation estimated six DoF motion, calculated for various motion traces in each DoF. Over 1000 images were used for each of the experiments

Rotations ()

rSI

Static poses

Dynamic trace

±2 about each axis

Prostate

Lung

0.01 (0.24)

0.46 (1.28)

0.01 (0.92)

rLR -0.04 (0.38) -0.28 (0.98) -0.05 (0.67) rAP -0.06 (0.16) 0.01 (0.32) Translations (mm) tLR 0.07 (0.37) tSI

-0.14 (0.31)

-0.07 (0.62) -0.03 (0.68)

-0.04 (0.11) -0.04 (0.15) -0.01 (0.14)

tAP -0.05 (0.39) 0.06 (0.99)

-0.11 (0.74)

References 1. D T Nguyen, J-H Kim, R O’Brien, C-Y Huang, J Booth, P Greer, K Legge, P Poulsen, J Martin, P Keall. The first clinical implementation of a real-time six degrees-of-freedom tracking system during radiation therapy. in AAPM 2016. 2016. Washington DC, USA. 2. J N Tehrani, R O’Brien, P R Poulsen and P Keall, Real-time estimation of prostate tumor rotation and translation with a kV imaging system based on an iterative closest point algorithm. Phys. Med. Biol., 2013. 58(23): p. 8517. 3. C-Y Huang, et al., Six degrees-of-freedom prostate and lung tumor motion measurements using kilovoltage intrafraction monitoring. Int. J. Radiation Oncol. Biol. Phys., 2014. 91(2): p. 368–375. 4. Horn, B.K.P., Closed-form solution of absolute orientation using unit quaternions. J. Optical Soc. Amer., 1987. 4: p. 629–642.

O120 Validation of a hybrid VMAT treatment technique for use in breast radiotherapy J. Luo1, J. Foo1, S. White1, C. Nguyen1 Nepean Cancer Care Centre, NSW, Australia. [email protected] [Presenting author]. ([email protected]), ([email protected]), ([email protected])

Fig. 1 An example of 6DoF lung motion trajectories acquired using KIM (solid lines) and kV/MV triangulation (symbols): (upper) rotations; (lower) translations about SI, LR and AP-axes

Introduction At Nepean Cancer Care Centre, hybrid IMRT has been used to treat breast cancer since 2000. To further improve the plan quality and decrease the treatment time, hybrid VMAT Breast plans were created in Philips Pinnacle3 TPS for an Elekta Synergy linear accelerator. The main aim of this project was to evaluate various dosimetric criteria for hybrid VMAT plans, with the goal of adopting this technique for clinical treatment and further reducing heart doses. Method 10 breast patients were planned using two different methods, hybrid VMAT and hybrid IMRT. The hybrid technique employs a combination of conformal and VMAT/IMRT fields. Both IMRT and VMAT plans were evaluated and compared based on isodose, homogeneity index, conformity index, max/min dose and

123

Australas Phys Eng Sci Med

Fig. 1 A side by side comparison of two plans performed on the same patient DVH for the target volume. Normal tissue index, max/mean doses and DVH for the organs at risk (lung, heart, and normal tissue) were also evaluated. Results The statistics for all 10 patients were analysed. Initial results show that the hybrid VMAT technique provides better PTV isodose coverage, better uniformity and less hot spots present in the organs at risk. Conformity Index (Paddick (2000)) improved from 0.69 (IMRT) to 0.78 (VMAT). Normal Tissue Index (YIM (2015)) decreased from 0.07 to 0.03, indicating less dose to normal healthy tissues. Mean heart doses decreased by 6%, with significant reduction in maximum dose of 6 Gy. Notably, mean dose increases of approximately 1 Gy were observed to both the contralateral breast and lung. Conclusion Our Hybrid VMAT Breast technique compares favourably to the IMRT equivalent, producing higher quality plans with better conformity and reduced doses to the organs at risk, including the heart. The VMAT technique does result in plans with an increased low dose wash, in particular to the contralateral breast and lung. This must be taken into consideration when adopting this technique (Fig. 1).

treatment planning system. On treatment, these parameters can deviate (within tolerances set) from the idealised parameters. In this study the dosimetric impact of this was assessed using the actual treatment parameters retrieved from the record and verify system for each delivered fraction. In-house software written in MatLab was used to convert these parameters into a format compatible with the treatment planning system allowing the dose distribution for a delivered fraction to be calculated. Method Dose distributions for each individual fraction of a patient’s treatment were calculated for 3 patients (1 pelvis and 2 head and neck). Dose planes were collected for each fraction and compared to equivalent planes relating to the originally planned dose distribution. Using OmniProTM software, a 1%/1 mm gamma analysis was performed between the planned and delivered planes producing 2-D error plots. Each plot was converted such that each pixel either passed or failed. Error maps for each fraction were then summed, the resulting error map showing the frequency of failure for each pixel. Dose distributions for each fraction were summed over the entire treatment. The summed distribution was subtracted from the original planned distribution thus producing a difference distribution. DVH parameters were collected for both the summed and the original clinical planned distributions. Results A representative 2D error maps (head and neck patient) is shown in Figure 1a adjacent to the corresponding dose difference plane, 1b. Table 1 outlines the colour coding used in these plots and these were consistent for all patients.

References 1. PADDICK I, ‘‘A simple scoring ratio to index the conformity of radiosurgical treatment plans’’ J Neurosurg (Suppl 3) 2000:93(12):219–222 2. YIM J, et al, ‘‘Intensity modulated radiotherapy and 3D conformal radiotherapy for whole breast irradiation: a comparative dosimetric study and introduction of a novel qualitative index for plan evaluation, the normal tissue index’’ J Med Radiat Sci., 2015:62(3):184–191

O121 Assessing delivery errors in VMAT treatments using record and verify log files. M. A. Barry1, J. Boyd1, P H Charles1,2, P. O’Connor3, A. Fielding4 Colour

2D error map Occurrence range

Dose difference plot isodose contour

Blue

30

2

1

Radiation Oncology Department, Calvary Mater Newcastle, Australia. ([email protected]) [Presenting author], ([email protected]). 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected]). 3Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. (Patrick.O’[email protected]). 4Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected]) Introduction For VMAT radiotherapy treatments, delivery parameters are set based on a required dose distribution as calculated in the

123

Red

26 and 29

1.5

Yellow

17 and 25

1

Pink

7 and 16

0.5

Conclusion Results showed the most frequently occurring errors (systematic errors) corresponded to dose differences in the entire

Australas Phys Eng Sci Med treatment that were greater than ±1.5 Gy. This was reflected in the DVH parameters however all were within clinical tolerances set.

into comprehensive data mining strategies to better inform future research and patient care.

IS08 Science and engineering in medicine: The vision (from the perspective of radiation oncology)

O122 An efficient EBT3 film pre-treatment QA software program based on linear dose response

V. A. Ahern1

M. Cai1, Y. Wang2, P. Metcalfe3

1

1

Sydney West Radiation Oncology Network (Crown Princess Mary Cancer Centre, Nepean Cancer Centre, Blacktown Cancer Centre). ([email protected])

Introduction Over the last 20 years, the practice of radiation oncology has evolved in diverse ways. Radiation oncologists have abdicated the delivery of systemic therapies to medical oncologists and palliative care to physicians, perhaps distancing ourselves from patient interaction. Instead we have focussed on the technical delivery of precision radiation treatments to smaller tumour targets, made possible by increasingly sophisticated information technology systems. Complete trust in medical physicists and radiation therapists to deliver treatment according to expectations has forced the professional groups to work as a triumvirate, not always cohesively. Recent rapid expansion of private radiation oncology practices has created new challenges. In this context, what is the vision for radiation oncology? And is a vision for an Australian radiation oncology department, network or consortium to become a national or international leader a realistic expectation? This presentation reflects on what makes a great radiation oncology department, and the role science and engineering have played in achieving in this. It considers the role of science and engineering in the establishment and operation of a national particle therapy treatment and research centre, and considers the fundamental driver for change – continually improving patient care.

Centre of Medical Radiation Physics, University of Wollongong, Australia. ([email protected]) (Presenting author). 2Radiation Oncology Centre, Australia. (yang.wang@roc-team). 3Centre of Medical Radiation Physics, University of Wollongong, Australia. ([email protected])

1

Introduction EBT3 film is a suitable dosimeter for pre-treatment quality assurance (QA) due to its high spatial resolution, near tissue equivalence and low energy dependence. However, its use has been hindered by lengthy and error-prone calibration and dosimetry procedures. It has been established that dose linearity can be constructed by adjusting scanner settings and that linearity is a more robust condition than calibration curve [1]. Therefore, relative dosimetry can be conducted without establishing a calibration curve beforehand. RODOMS (Radiation Oncology DOsimetry Management System) was developed to utilize dose linearity and enable efficient and accurate QA procedures. Method An IAEA example case for SRS commissioning was selected where the PTV is a sphere of 3 cm diameter. The QA plan was calculated in RayStation treatment planning system and delivered in a cheese phantom using VMAT modality. Dosimetry film and unexposed film scanned at linearity condition as well as DICOM dose file were imported into RODOMS. Background correction was applied to reduce scanning non-uniformity. Intensity-based automatic image registration was applied both translationally and rotationally to align film image to dose map. A regression best-fit scaling factor was automatically generated based on film image and dose map to convert film intensity readings to dose. QA analysis included profile comparison and gamma analysis with a 0.2 mm calculation grid and 10% threshold. Results RODOMS successfully registered the film image with dose map and scaled film intensity readings to dose. Figure 1 displays the result showing excellent profile agreement and more than 95% gamma pass rate using 2 mm/2% criteria.

Medical Physics is a highly interdisciplinary field at the intersection between physics and medicine. Medical Physics innovations have had and continue to have profound impact by developing improved imaging and treatment technologies, and help advancing our understanding of the complexity of the disease. Medical Physics has been highly successful in implementing these innovations in several areas, such as radiation therapy and diagnostic imaging - the foundation of the well-established medical physics profession, which has and will continue to have significant impact on patient care. However, besides radiation therapy and diagnostic imaging, plenty of opportunities lie ahead in translating existing medical physics innovations into patient care. For example, medical physicists should embark on implementing quantitative imaging methods (e.g., scanner harmonization, reduction of imaging uncertainties, novel image analysis techniques, automatic image interpretation), and actively participate in innovative clinical trials that can better elucidate disease progression and response to therapies. Extensive imaging data available from such trials (e.g., radiomics) should be extended beyond imaging, with better integration of genomic and other molecular data

Fig. 1 Example QA result using RODOMS program

IS09 Medical Physics innovations today – Medical Physics jobs in the future Robert Jeraj1 University of Wisconsin, Madison, USA

123

Australas Phys Eng Sci Med Conclusion RODOMS program simplifies film dosimetry with automatic image registration and dose scaling without using a calibration curve. It proved to be an efficient and accurate tool for routine pre-treatment QA. References 1. Cai M, Archibald-Heeren B, Wang Y, Metcalfe P (2016) Linearization of EBT3 film dose response and virtual film dosimetry for SBRT quality assurance. JPCS. Accepted for publishing.

Table 1 Composition of lung-equivalent gels Composition

Gel type 1 (Foam) Gel type 2 (Styrofoam beads)

Gelatin (300 bloom)

12% (w/w)

8% (w/w)

Methacrylic acid (MAC)

6% (w/w)

6% (w/w)

Bis [tetrakis (hydroxymethyl) 10 mM phosponium] sulphate

20 mM

Sodium dodecyl sulphate (SDS)

0.15% (w/w)

–

Water

82% (w/w)

86% (w/w)

O123 Preliminary investigation of 3D lung-equivalent gel dosimeters T. Alharthi1,2,3,4, L. Holloway2,5,6,7,8,9, D. Thwaites1, S. Arumugam2,5, Y. De Deene10 1

Institute of Medical Physics, School of Physics, University of Sydney, Sydney, NSW, Australia. ([email protected]). 2Ingham Institute for Applied Medical Research, Australia. 3Department of Engineering, Macquarie University, Sydney, NSW, Australia. 4School of Medicine, Taif University, Taif, Saudi Arabia. ([email protected]) [Presenting author]. 5Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Australia. ([email protected]). 6CMRP, UOW, Australia. 7 Institute of Medical Physics, University of Sydney, Australia. 8 Sydney Medical School, University of Sydney, Australia. 9South Western Sydney Clinical School (SWSCS), University of New South Wales, Australia. ([email protected]). 10 Department of Engineering, Macquarie University, Sydney, NSW, Australia. ([email protected]) Introduction The accuracy of dose calculations for radiotherapy requires a precise knowledge of the internal geometry and electron density of different tissues. This becomes significant in low-density lung tissues in which radiation interactions occur at different scales compared to other water equivalent tissues. The aim of this study was to investigate the feasibility of two approaches in generating a homogenous low-density polymer gel dosimeter including an assessment of the resulting electron density. Method Building on previous work by De Deene et al [1,2] on lungequivalent gels, two approaches have been considered to reduce the density either by beating the gel solution into foam (Type1) or by embedding Styrofoam beads in the gel solution (Type 2). Two spherical phantoms (each 250 ml) were prepared as listed in (Table 1) and CT scanned (120KV Philips) to obtain electron density from regions of interest (ROIs). Results Both gels showed electron density relative to water ranging from 0.292–0.359 and 0.405–0.452 for type 1 and type 2 gels, respectively. These values are close to what has been reported for lung tissue [3]. Gel (Type 1) demonstrated a homogeneous structure (Fig. 1 (a)) with average electron density relative to water of 0.336 (SD = ±0.02) and intra-slice variations (microstructure related variations in electron density) of 2.3% while gel (Type 2) was more homogenous with an average electron density of 0.425 (SD = ±0.01) and intra-slice variations of 1.2% (Fig. 1 (b)). Conclusion The two lung-equivalent gel types have electron densities in the same range as lung tissue. Because of the broader size distribution of the gas bubbles in type 1 gel as compared to the Styrofoam beads in type 2 gel, the latter was more homogenous. Further

123

Fig. 1 Axial x-ray CT image of gels, (a) Foam, (b) Styrofoam beads

experiments are underway to verify dose reproducibility and accuracy of these lung equivalent 3D dosimeters. References 1. De Deene Y, Vergote K, Claeys C, & De Wagter C. (2006) Three dimensional radiation dosimetry in lung-equivalent regions by use of a radiation sensitive gel foam: Proof of principle. Medical physics, 33(7), 2586–259. 2. De DeeneY, Vandecasteele J, & Vercauteren T. (2013) Lowdensity polymer gel dosimeters for 3D radiation dosimetry in the thoracic region: A preliminary study. In Journal of Physics: Conference Series (Vol. 444, No. 1, p. 012026). IOP Publishing. 3. Aarup LR, Nahum AE, Zacharatou C, Juhler-Nøttrup T, Kno¨o¨s T, Nystro¨m H, & Korreman SS. (2009) The effect of different lung densities on the accuracy of various radiotherapy dose calculation methods: implications for tumour coverage. Radiotherapy and oncology, 91(3), 405–414.

O124 Calibration and energy corrections for LiF based themoluminescence dosimetry (TLD) neutron measurements S. Keehan1,2, P. Lonski3,4, M. L. Taylor4, T. Kron3,5, R. D. Franich5 1

School of Science, RMIT University, Victoria, Australia. 2Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]). 3Physical Sciences, Peter MacCallum Cancer Centre, Victoria, Australia. 4 School of Science, RMIT University, Victoria, Australia.

Australas Phys Eng Sci Med of Technology, Brisbane, Australia. ([email protected]), ([email protected] [Presenting author]), ([email protected]). 3 Radiation Oncology Mater Center, South Brisbane, Australia. ([email protected])

Kerma per incident neutron (MeV/g)

10000 1000 100 10 1 0.1 0.01 0.001 0.0001

0.00001 1E-9

Natural lithium fluoride Lithium-6 fluoride Lithium-7 fluoride Ratio: lithium-6 fluoride/lithium-7 fluoride

1E-7 1E-5 1E-3 1E-1 Incident neutron energy (MeV)

1E+1

Fig. 1 The Kerma per incident neutron as a function of neutron energy. The Kerma in the material is assumed to be directly proportional to its TL response ([email protected]), ([email protected]). 5School of Science, RMIT University, Victoria, Australia. ([email protected]), ([email protected]) Introduction The photonuclear effect causes the production of contaminant neutrons in high energy radiotherapy. In order to assess the risk to patients, accurate neutron detection is required. TLDs enriched in 6Li or 7Li respond differently to neutron exposures. If the response difference can be correlated with the degree of neutron exposure, TLD pairs can be used to quantify neutron exposure even in mixed fields. Method Analytical and MCNP Monte Carlo calculations have been used to determine the Kerma in 6LiF and 7LiF from neutrons over a range of energies. The response of the TLD is assumed to be directly proportional to the Kerma. Energy correction factors for the neutron spectra of several calibration sources have been determined. The energy spectrum of neutrons reaching the patient plane in radiotherapy has also been calculated. Results The Kerma deposited in LiF with different lithium isotope concentrations is dependent on the energy of the incident neutrons. 7 LiF is often assumed to be insensitive to neutron radiation, but Fig. 1 shows that its response to neutrons cannot be ignored above *2 keV. AmBe is a common calibration source, with mean energy of 4.05 MeV. The energy correction for the response difference for AmBe neutrons is 2.3. If this factor is omitted from a calibration factor determined using AmBe, the dose will be over-estimated by 130%. The energy spectrum incident on a patient undergoing high energy radiotherapy has an energy correction of 1.00 ± 0.03. Conclusion 7LiF TLDs are not insensitive to high energy neutron radiation. The energy dependence of the responses of 6LiF and 7LiF TLDs needs to be considered for accurate calibration.

Introduction 3D printing can be used to make phantom materials for use in imaging and dosimetry studies [1–4]. This study seeks to provide a comprehensive evaluation of the radiological properties of 3D printed materials in kVCT and MVCT beams. Models were printed with a variety of materials, including plastic, wood and metal filaments as well as photopolymer resins. Method Cylindrical inserts were printed using ABS, PLA, photoluminescent PLA, woodfill, copperfill and bronzefill filaments as well as standard and flexible photopolymer resins. In order to simulate the attenuation of a range of body tissues, inserts were printed with various infill ranging from 10% to 100%. The inserts were imaged in a Gammex Model 467 Tissue Characterization phantom (Gammex Inc., Middleton, USA) with a Toshiba Aquilion CT scanner (Toshiba Medical Systems Corporation, Otowara, Japan) at 80 KVp and 120 kVp. MVCT imaging was carried out on a Tomotherapy Hi-Art unit (TomoTherapy, Madison, USA). Results The CT number was observed to increase with infill, as shown in Fig. 1 for the plastics. Lung, soft tissue, tumour, bone mineral and cortical bone were all accurately modelled by the printed materials (Figs. 2, 3). In Fig. 2, there are two distinct linear regions in each graph – one up to 0HU and another after 0HU. The metal filaments provided a good surrogate for cortical bone in the MVCT images, but exhibited streaking and cupping artifacts in the kVCT images [5]. Conclusion This study showed that 3D printed materials can model a wide variety of body tissues, including lung, soft tissue, tumour, bone

Fig. 1 CT numbers of the plastics with kVCT and MVCT imaging protocols

O125 Radiological properties of 3D printed materials in kilovoltage and megavoltage photon beams O. L. Dancewicz1,2, S. R. Sylvander1, T. S. Markwell3, S. B. Crowe1,2, J. V. Trapp2 1

Royal Brisbane and Women’s Hospital, Herston, Australia. ([email protected]). 2Queensland University

Fig. 2 CT number - Electron Density (CT-ED) curves for the printed inserts derived from the kVCT scans. The metallic inserts are marked with an arrow

123

Australas Phys Eng Sci Med 1

Shoalhaven Cancer Care Centre, Nowra, NSW. ([email protected]) (Presenting author), ([email protected]), ([email protected]), [email protected]. 2Illawarra Cancer Care Centre, Wollongong, NSW, Australia. 3Current address: Department of Medical Physics, St George Cancer Care Centre, Kogarah, NSW. ([email protected]). 4Current address: Radiation Oncology, Prince of Wales Hospital, Randwick, NSW. ([email protected])

Fig. 3 CT-ED curve for the printed inserts derived from the MVCT scan. The metallic inserts are marked with an arrow

mineral and cortical bone. These tissue types are accurately modelled in kVCT and MVCT imaging. This data will be of use to medical physics researchers and clinicians seeking to manufacture their own phantoms in-house that will accurately model the interaction of body tissue with radiation. Acknowledgements The authors would like to acknowledge Queensland University of Technology, Royal Brisbane and Womens Hospital (RBWH) Department of Nuclear Medicine and Specialised PET Services Queensland and RBWH Cancer Care Services for use the 3D printers. The authors would also like to acknowledge Craig Lancaster from RBWH Cancer Care Services for facilitating use of the Tomotherapy unit.

Introduction Plane-parallel ionisation chambers are regularly used to conduct relative dosimetry measurements for kilovoltage beams [1]. The polarity effect has not been previously quantified for these chambers in kilovoltage applications. Method The polarity effect was quantified for Advanced Markus and Roos ionisation chambers (PTW-Freiburg, Germany) at various depths and field sizes in solid water. Measurements acquired for kilovoltage beams between 100 kVp (half-value layer (HVL) = 2.88 mm Al) and 250 kVp (HVL = 2.12 mm Cu) using field diameters of 3–15 cm for 30 cm focus-source distance (FSD) applicators and 4 9 4 cm2-20 9 20 cm2 for 50 cm FSD. Results The maximum variation in the polarity effect for the Advanced Markus with depth was 0.6%, compared to 0.5% for the Roos. This consistency negates the need for polarity corrections during depth dose measurements. The changes in polarity effect with field size for 30 cm and 50 cm FSD applicators are shown in Figs. 1 and 2 respectively. Polarity effects up to 9.6% were observed for the Advanced Markus chamber, with a maximum 0.5% for the Roos chamber. A large fraction of the dose deposition in kilovoltage beams is due to backscatter, which is field-size dependent [2]. The differences in backscatter between applicators means that polarity effects will not

References 1. Kairn T, Crowe SB, Markwell T (2015) Use of 3D printed materials as tissue-equivalent phantoms in World Congress on Medical Physics and Biomedical Engineering June 7–12, 2015, Toronto, Canada, IFMBE Proceedings 51 2. Ehler ED, Barney BM, Higgins PD, Dusenbery KE (2014) Patient specific 3D printed phantom for IMRT quality assurance, Phys Med Biol 59:5763–5773 3. Kiarashi N, Nolte AC, Sturgeon GM, Segars WP, Ghate SV, Nolte LW, Samei E, Lo JY (2015) Development of realistic physical breast phantoms matched to virtual breast phantoms based on human subject data, Med Phys 42:4116–4126 4. Miller MA, Hutchins GD (2007) Development of anatomically realistic PET and PET/CT phantoms with rapid prototyping technology in 2007 IEEE Nuclear Science Symposium Conference Record Oct 26-Nov 3, 2007, Honolulu, USA, pp 4252–425 5. Barrett JF, Keat N (2004) Artifacts in CT: recognition and avoidance. Radiographics 24(6):1679–1691

O126 Potential discrepancies in relative dose measurements in kilovoltage photon beams: Consequences and implications of polarity effects in plane-parallel ionisation chambers S. Dowdell1, J. McNamara1, M. Tyler1, K. Sloan2, A. Ceylan2,4, A. Rinks1

123

Fig. 1 The polarity effect for (a) Advanced Markus and (b) Roos ionisation chambers measured at 2 cm depth in solid water for different kilovoltage beam energies and circular 30 cm focus-source distance applicators of various diameter

Fig. 2 The polarity effect for (a) Advanced Markus and (b) Roos ionisation chambers measured at 2 cm depth in solid water for different kilovoltage beam energies and square 50 cm focus-source distance applicators of various sizes

Australas Phys Eng Sci Med cancel out when measuring output factors, necessitating clinical values to be corrected. Conclusion Depending on the individual chamber, polarity corrections may be necessary in kilovoltage beams to ensure the correct patient doses.

O128 Development of an eye lens dosimetry solution for use during interventional procedures N. K. Thorpe1, E. H. Da Silva2, F. Vanhavere3, D. L. Cutajar1, M. Pitney4, A. B. Rosenfeld1

References 1

1. Hill R, Healy B, Holloway L, Kuncic Z, Thwaites D, Baldock C (2014) Advanced in kilovoltage x-ray beam dosimetry Phys. Med. Biol. 59 R183-R231 2. Grosswendt B (1990) Dependence of the photon backscatter factor for water on source-to-phantom distance and irradiation field size Phys. Med. Biol. 35 1233–1245

O127 9 years of tissue reaction data – Trends in Interventional Fluoroscopy T. A. Ireland1 1

Biomedical Technology Services, Gold Coast University Hospital. ([email protected])

Introduction Peak skin doses are of high concern in interventional fluoroscopy due to the chance of resultant tissue reactions. Over the past 10 years many developments in structured dose reporting and computational methods have led to improvements in peak skin dose calculation accuracy and recommendations to increase thresholds at which peak skin dose calculations are required. Data collected over the past 9 years has been summarised to review trends in high risk procedures Method Using Radiation Dose Structured Reports and system generated dose reports, accurate estimates of Peak Skin Doses were achieved. Retrospective analysis was conducted to compare: • • • • •

Total number of examinations Percentage of examination (by procedure type) with reported PSD [ 3 Gy Difference between reported PSD compared with calculated PSD Percentage of examinations with calculated PSD [ 3 Gy that resulted in tissue reactions Patient follow up rate

Results Total number of complex interventional fluoroscopic procedures is increasing, however with improving technique and technology, the percentage of exams leading to PSD [3 Gy is decreasing. Percentage of examinations with calculated PSD [ 3 Gy that result in tissue reactions remains consistently at around 2%. The structured patient follow methods employed have yielded inconsistent response rates with GPs. When tissue reactions were identified median reported PSD was 7.4 Gy, and median calculated PSD was 4 Gy. Conclusion Trends identified from this facility support recommendations from the ICRP to increase the threshold for follow from reported PSD of 3 Gy to 5 Gy. This facility has identified cerebral coilings and cerebral embolisations as the most common examinations to deliver PSD in excess of 3 Gy.

Centre for Medical Radiation Physics, University of Wollongong, AUS. ([email protected]) [Presenting author], ([email protected]), ([email protected]). 2SCK-CEN, BEL Department of Radiology, Universitair ziekenhuis Brussel, BE. ([email protected]). 3SCK-CEN, BEL. ([email protected]). 4Southern Heart Clinic, Eastern Heart, AUS. ([email protected])

Introduction Clinicians performing interventional procedures have recently been observed to experience radiation induced conditions. Studies performed across France, Finland and Belgium found high incidences of clinicians experiencing radiation induced cataracts. In light of these findings, ICRP has recommended lower eye lens dose limits and there have been calls for closer monitoring of eye lens doses within the clinical environment [1]. This study identified the needs for a dedicated eye lens dosimetry system, investigating the most effective and accurate location for such a system to be placed upon a clinician. Method Measurements were performed using the MOSkin dosimeter. The MOSkin has been previously used to measure eye lens dose in cerebral angiographic procedures [2]. The MOSkin was characterised to standards set by ISO 4037-3 using the Narrow and RQR Series beam qualities. The detectors were placed upon an Alderson Phantom equipped with radioprotective lead glasses. Dosimeter response at each measurement point was compared to the measured eye lens dose for various irradiation angles (Fig. 1). Results The scaled measurement point doses were found to vary from eye lens dose on average by: 4.7% (Under Glasses), 7.95% (Over Glasses), 4.5% (Forehead) and 0.6% (Side Glasses, limited range). Conclusion The Under Glasses and Forehead measurement points were observed to have the closest relationship with the eye lens dose. The Side Glasses measurement point was also capable of reproducing eye lens dose within a limited range. Despite their suitability, each of these measurement points present unique challenges in implementation. The MOSkin dosimeter was capable of accurate low dose measurements within the RQR Series beam qualities. Future MOSkin designs will feature higher sensitivity and wireless real-time readout capabilities. These designs will be suitable for monitoring clinician doses and could be used to prevent clinician overexposure.

Fig. 1 (a) MOSkin Measurement Points on Alderson Phantom, (b) Range of beam incidence angulation

123

Australas Phys Eng Sci Med References 1. Vanhavere F, Carinou E, Gualdrini G, Clairand I, Merce M, Ginjaume M (2009) The ORAMED Project: Optimisation of Radiation Protection for Medical Staff. IFMBE Proceedings 25/3:470–473. doi: 10.1007/978-3-642-03902-7_133 2. Safari M, Wong J, Kadir K, Thorpe N, Cutajar D, Petasecca M, Lerch M, Rosenfeld A, Ng K (2015) Real-time eye lens dose monitoring during cerebral angiography procedures. Eur Radiol, 6(1):79–86. doi: 10.1007/s00330-015-3818-9

O129 Muscle and bone dose from low kVp paediatric digital radiographs

an approximately 600 speed screen-film system and is in the range of sensitivity for digital detector systems.2 Conclusion Although extremity doses remain low, reducing the kVp in paediatric radiology of the extremities can result in significantly higher radiation doses, particularly for thicker limbs. Applying the ALARA principle, low kVp radiography requires justification for use on the extremities. It should not be used for exposures of any other anatomical region. References 1. Huda W, Glanatsios NA (1998) Radiation dosimetry for extremity radiographs. Health Phys 75(5):492–499 2. Cowen AR, Kengyelics SM, Davies AG (2008) Solid-state, flatpanel, digital radiography detectors and their physical imaging characteristics. Clin Radiol 63(5):487–498

Y. V. Matyagin1, D. W. McRobbie2 1

SA Medical Imaging, Royal Adelaide Hospital, Adelaide, Australia. ([email protected]), 2SA Medical Imaging, Flinders Medical Centre, Adelaide, Australia. ([email protected] [Presenting author])

IS10 Optimisation of image quality and patient dose in radiographs of paediatric extremities using direct digital radiography

Introduction The proliferation of digital radiography has led to a reevaluation of exposure parameters and image quality. Currently there is a move towards reducing peak X-ray tube voltage (kVp) in paediatric exposures in order to achieve better contrast and contrast-tonoise ratio. The aim of this study was to evaluate the effect on patient dose of reducing kVp in paediatric limb digital radiography exposures. Previous work considered higher kVp.1 Method Monte Carlo simulation of radiographic exposures on a paediatric limb phantom was performed in the range of 40 to 70 kVp and the total beam filtration (TF) of 2.5, 3.0 and 3.5 mm of aluminium (Al). The modelled limb phantom included muscle tissue and bone segments of 5 different densities in the range of 1.12 to 1.48 g/cm3. The overall thickness of the phantom varied between 1 and 10 cm. Results X-ray tube current-time products (mAs) required to achieve equal detector dose through muscle versus limb thickness for different kVp and TF were calculated. Average muscle and bone doses are shown against limb thickness, kVp (shown for 3 mm Al TF), and bone density in the Fig. 1 for the detector dose (through muscle) of 1.7 lGy that would correspond to the nominal receptor exposure for

A. Jones Medical Physics Specialist, Westmead Hospital, NSW. ([email protected] [Presenting author]) Previously presented at: UK SRP annual conference 2014 – Winner of UK Young Scientist Award, European Society of Paediatric Radiology 2014, International Radiation Protection Association 2016 – Cape Town – Winner of International Young Scientists and Professionals Award Introduction The purpose of this study was the evaluate effect of beam quality on the image quality (IQ) of ankle radiographs of paediatric patients in the age range of 0–1 years whilst maintaining constant effective dose (ED). Methods Lateral ankle radiographs were taken of an infant foot phantom at a range of tube potentials (40–64.5 kVp) with and without 0.1 mm Copper filtration using a Trixell Pixium 4600 detector. ED to the patient was computed for the default exposure parameters using PCXMC v2.0 and fixed for other beam qualities by modulating the mAs. The contrast-to-noise ratio (CNR) was measured between the tibia and adjacent soft tissue. The IQ of the phantom images was assessed and scored by three radiologists and a reporting radiographer. Four IQ criteria were defined each with a scale of 1–3, giving a maximum score of 12. Finally, a service audit of clinical images at the default and optimum beam qualities was undertaken. Results The measured CNR for the 40 kVp/ no Copper image was 12.0 compared to 7.6 for the default mode (55 kV 0.1 mm Copper). An improvement in the clinical IQ scores was also apparent at this lower beam quality. Conclusion Lowering tube potential and removing filtration improved the clinical IQ of paediatric ankle radiographs in this age range with no additional dose penalty.

IS11 The importance of high quality radiotherapy S. Kry1 1

Fig. 1 Muscle and bone doses for equal detector dose v limb thickness for various kVp and bone densities

123

The University of Texas MD Anderson Cancer Center. ([email protected] [Presenting author])

Australas Phys Eng Sci Med For radiation therapy to be effective at controlling tumors while minimizing acute and long-term side effects, the correct dose must be delivered to the correct location. This has been historically known based on dose response curves, but has been reinforced with more modern studies, particularly in the clinical trial setting. These studies have given rise to the classic recommendation that the radiotherapy dose must be accurate within 5% to achieve the desired effects. Achieving a 5% overall accuracy typically requires a 2% accuracy on the dose calculation. However, there are major obstacles that are impeding our ability to deliver high quality radiotherapy. First, based on IROC Houston onsite dosimetry audits, most treatment planning system beam models provide sub-optimal descriptions of the radiotherapy beam. Problems are particularly rampant for small fields, where approximately 2/3 of institutions show a 3% or larger discrepancy between modeled and measured output. Second, it can be very challenging to change the dose calculation process, or a patient’s treatment. This often stems from limited time for physics review in clinical workflow, and demands that physics be empowered to effect necessary changes. Through empowered physics and a focus on critical dose calculation issues, both of these issues can be controlled, leading to improved outcomes for our patients.

O130 The perirectal space as the principal organ at risk for GI toxicity following pelvic radiotherapy M. A. Ebert1,2, S. Gulliford3, S. Ghose4, A. Kennedy1, J. Mitra5, J. Dowling6, J. W. Denham7 1

Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia. ([email protected]). 2School of Physics, University of Western Australia, Western Australia. ([email protected] [Presenting author]). 3Joint Department of Physics, Institute of Cancer Research and Royal Marsden National Health Service Foundation Trust, Sutton, United Kingdom. ([email protected]). 4Department of Biomedical Engineering, Case Western Reserve University, 10900 Euclid Ave, Cleveland, OH. ([email protected]). 5Department of Biomedical Engineering, Case Western Reserve University, 10900 Euclid Ave, Cleveland, OH. ([email protected]). 6Australian e-Health Research Centre, CSIRO, Brisbane, Queensland. ([email protected]). 7School of Medicine and Public Health, University of Newcastle, Callaghan, NSW. ([email protected]) Introduction Gastrointestinal (GI) toxicity limits the dose and dose distributions that can be applied in pelvic radiotherapy. Whereas many studies have focused on rectal bleeding as a principal endpoint, patients typically prioritise bowel-control like symptoms as having a greater impact on quality of life. Evidence suggests that dose to organs other than the rectum are causally related to control-like symptoms. Method This study specifically examined the perirectal space (PRS), which is the region (of mostly fat) that the rectum can expand into during normal GI activity. A multi-atlas based autosegmentation method, combined with unsupervised clustering of CT information, was developed and validated on a test dataset. The method was then used to generate PRS structures in the CT data of 100 RADAR trial patients, selected due to reported control-like symptoms. The impact of DVH data for the PRS region, as well as the rectum, was assessed for both control and bleeding symptoms. Results The autosegmentation method showed 99% voxel-wise accuracy. 79 of the resulting datasets were considered useable. Atlases of complication incidence (ACI) demonstrated that control-

like symptoms were related to the dose distribution to the PRS which was confirmed with a Wilcoxon signed rank test at multiple dose levels. No relationship was found for incidence of bleeding and ACI for the PRS. Bleeding was, however, as expected on the basis of previous investigations, strongly determined by high dose values to the anal canal. Conclusion This study has demonstrated the contribution to GI symptoms, observed following pelvic radiotherapy, of radiation dose to non-GI anatomy. This result suggests a rethink of the way in which dose distributions are optimised for pelvic radiotherapy in order to minimise treatment-related toxicity. Acknowledgements We gratefully acknowledge the support of the Australian National Health and Medical Research Council (grants 1006447 and 1077788).

O131 Modelling urinary dysfunction following external beam radiotherapy of the prostate based on bladder dose-surface maps: Evidence of spatially-variable response of the bladder surface N. Yahya1, M. A. Ebert2, M. J. House2, A. Kennedy3, J. Matthews4, David J. Joseph3,5, James W. Denham6 1

School of Health Sciences, National University of Malaysia, Kuala Lumpur, Malaysia. ([email protected]). 2School of Physics, University of Western Australia, Western Australia, Australia. 3 Department of Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia, Australia. ([email protected] [Presenting author]). 4Department of Radiation Oncology, Auckland City Hospital, Auckland, New Zealand. 5School of Surgery, University of Western Australia, Western Australia, Australia. 6School of Medicine and Public Health, University of Newcastle, New South Wales, Australia Introduction Normal tissue outcome studies rely on dose-histograms, degenerate to spatial information. In this study, we adapted the methods used to develop dose-surface maps of the rectum through organ unfolding to produce maps visualising the dose received by the bladder surface. From the developed maps, pixel-wise outcome analyses were performed. Method The bladder dose-surface maps of 754 participants from the TROG 03.04-RADAR trial were generated from the volumetric data by virtually cutting the bladder at the sagittal slice intersecting the bladder centre-of-mass through to bladder posterior and projecting the dose information on a two-dimensional plane. Pixel-wise dose comparisons between patients with and without symptoms (dysuria, haematuria, incontinence and increase of C10 International Prostate Symptom Score (DIPSS10)) were performed and results with and without permutation-based multiple-comparison adjustments generated. Results The associations of the spatially-specific dose measures to urinary dysfunction were found to be dependent on specific symptoms. Doses received by the anterior-inferior and, to lesser extent, posterior-superior surface of the bladder were found to have the strongest relationship to the incidence of dysuria, haematuria and DIPSS10, both with and without adjustment for clinical factors (Fig. 1). For doses to the posterior-inferior region corresponding to the area of the trigone, the only symptom showing significance was incontinence. Conclusion Spatially variable response of bladder surface to dose was found for symptoms of urinary dysfunction. Dose surface maps have the potential to provide evidence for spatially-defined dose constraints.

123

Australas Phys Eng Sci Med fitted to the data. Additionally, the EUD was replaced by a LDE weighted dose parameter (EUDlog) and the mean-lung dose (MLD). A random model (p) was used to test for dose-effect relation. The Akaike Information Criteria (AIC) was used for model comparison. Model with the least AIC is considered the best. Results The best-fit parameter values obtained with 95% confidence intervals were 50.1 Gy (32.7–274.7), 0.50 (0.44–0.58) and 26.2 Gy (19.7–57.6), 0.40 (0.32–0.52) for TD50 and ‘m’ respectively for MLD and EUDlog models. The TD50, ‘m’ and ‘n’ parameters in the EUD model were 23.7(7.7–195) Gy, 0.52(0.44–0.60) and 1.9 (0.48–?) respectively. The random model parameter p was 0.05. The AIC values of the MLD, EUD and random model were 2.8, 3.9 and 7.7 higher compared to the EUDlog model. Conclusions EUDlog was found to be the best model to describe the incidence of RP dose effect relation. References

Fig. 1 Visualisation of the results on maps. Dose difference maps (black lines represent area with Wilcoxon signed-rank unadjusted p \ 0.05, pink lines represent area with adjusted p \ 0.05) for dysuria (A), haematuria (B), incontinence (C) and International Prostate Symptom Score (IPSS) increase of C10 (D) Acknowledgements This study was supported by NHMRC grants 1006447 and 1077788.

O132 NTCP modelling of radiation-induced pneumonitis in SBRT based on local dose-effect relationship based on lung perfusion changes derived from SPECT

1. Scheenstra, Alize EH, et al. ‘‘Local dose–effect relations for lung perfusion post stereotactic body radiotherapy.’’ Radiotherapy and Oncology 107.3 (2013): 398–402. Acknowledgements This research was partially supported by Elekta through a research grant with all institutions being members of the Elekta Lung Research Group. This work and these data, however, are the intellectual property of the individual group members and their sponsoring institutions.

O133 Laturality of secondary malignancies in Hodgkin Lymphoma patients E. Kyriakou1, T. Kron1, G. Wheeler2 1

1

3

4,5

J. Selvaraj , Andrew Hope , Matthias Guckenberger , Maria Werner-Wasik6, Jose´ Belderbos1, Inga Grills7, Jan-Jakob Sonke1 1

Department of Radiation Oncology, Netherlands Cancer Institute, Amsterdam, The Netherlands. 2Medical Physics and Radiation Engineering, The Canberra Hospital, ACT, Australia. ([email protected] [Presenting author]). 3Department of Radiation Oncology University of Toronto and Princess Margaret Cancer Center, Toronto, Canada. 4Department of Radiation Oncology, University of Wuerzburg, Wuerzburg, Germany. 5 Department of Radiation Oncology, University Hospital Zurich, Zurich, Switzerland. 6Department of Radiation Oncology Thomas Jefferson University Hospital, Philadelphia, PA, USA. 7Department of Radiation Oncology Beaumont Hospital, Royal Oak, USA Introduction Radiation pneumonitis (RP) is an inflammatory response of the lungs following radiotherapy representing a possibly severe and dose limiting toxicity. Several NTCP models for RP have been published, mostly for conventional treatments. NTCP models for RP following SBRT are less established. Recently, we found a sigmoid local dose effect (LDE) relation for perfusion loss after SBRT [1]. This study aims to investigate if this LDE improves the NTCP model fit of SBRT induced RP. Methods and materials Multi-institutional data of 1015 patients of which 51 have developed grade C2 RP were used in the analysis. Number of fractions and prescription doses in the cohort varied from 3 to 10 and 15 to 60 Gy respectively. Dose distributions were converted to EQD2 with an a/b of 3 Gy. The LKB NTCP model was

123

Department of Physical Science, Peter MacCallum Cancer Centre, Melbourne. ([email protected] [Presenting author]). 2 Department of Radiation Oncology, Peter MacCallum Cancer Centre, Melbourne Introduction Radiotherapy is an effective treatment for Hodgkin Lymphoma (HL). However, given the young age of many HL patients, secondary cancers and other radiation induced late effects are of considerable concern. The objective of this study is to determine if radiation dose can be associated with the location and risk of secondary cancer development. Method The Peter MacCallum Cancer Centre Late effects clinic has a database of long term follow up of HL patients. We reviewed records between 1964 and 2010: 217 patients had been diagnosed with HL. Of the 742 patients treated with radiotherapy 94 were women and 84 men. Subsequently, 14 women developed a secondary breast cancer and an additional 15 patients were also treated for breast cancer as a result of their earlier radiation treatment due to HL, however treated at a hospital other than PeterMac. Results From the cohort of 29 patients treated with radiotherapy for HL who developed a secondary breast cancer, a pattern emerged between their treatment field and location of secondary malignancy. Of patients who received full mantle irradiation developed breast cancer in both breasts (n = 5) and all patients treated with a mini mantle field developed breast cancer in one of the inner quadrants of the breast (n = 19). Patients who were treated with half mantle or involved field where the field of irradiation was only to one side, developed breast cancer on the ipsilateral breast (n = 7). As can be seen in Fig. 1 the incidence of breast cancer in prepuberty female patients was low while the incidence in patients receiving radiotherapy at age 14 and older increased significantly.

Australas Phys Eng Sci Med Method A multidisciplinary approach has been taken to develop strategies, protocols and education to appropriately apply the radiation protection principles and minimise the risk of radiation induced complications. Measures to justify and optimise each case have been adopted including: a tailored patient selection process, simple beam arrangements to minimise scattered radiation, low dose regimes, proximity of radiosensitive organs and evaluation of symptom improvement. Results Several case studies will be presented for patients with conditions such as plantar fasciitis and Dupuytren’s contracture. Conclusion International publications have provided us with an opportunity to learn from and adapt the European recommendations1 to provide treatment for local patients References 1. Taylor R et al ‘A review of the use of radiotherapy in the UK for the treatment of benign clinical conditions and benign tumours.’ London: The Royal College of Radiologists, February 2015. Fig. 1 Age at diagnosis of secondary malignancy as a function of age at treatment for women treated with radiotherapy for Hodgkin Lymphoma

IS12. The 4 ‘R’s: Responsibilities, roles and remaining relevant Sandra Turner1 1

Senior staff specialist in radiation oncology, Crown Princess Mary Cancer Centre, Westmead Hospital, Sydney NSW. ([email protected])

Fig. 2 Local-dose effects for MLD and EUD normalized to EUDlog Conclusion The late effects database is a unique resource. From understanding past practice and impact thereof we hope to be able to improve current RT techniques (Fig. 2).

Radiation therapy is an effective cancer treatment involved with 40% of all cancer cures. It would be effective in cure or palliation of around half of all cancer patients at some point in their treatment pathway yet around 1/3 of patients (at best) actually receive it, even in wealthy countries. One of the major reasons for this is lack of awareness and lack of knowledge about modern radiation therapy, and/or fear and misconceptions of patients, the general community and health professionals, including GPs. It is proposed that all health professionals working in the area of radiation oncology could have a role in ensuring that patients continue to get the optimal benefit of this effective treatment. This may mean taking on unfamiliar roles and responsibilities including patient advocacy, and re-thinking elements of training programs. This session will explore in an interactive and entertaining manner, the patient experience with radiation therapy and the potential for medical physicists to shape the future of radiation oncology outside conventional technical and clinical research endeavours.

O134 Development of Local Practises for Radiation Therapy of Benign Diseases K. M. Harrison1, K. Buman1, J. Martin2 1

Genesis Cancer Care, Newcastle NSW. ([email protected] [Presenting author]), ([email protected]). 2Calvary Mater Newcastle and Genesis Cancer Care, Newcastle NSW. ([email protected])

P02 Dose reconstruction with a 2D transmission silicon diode array ‘‘Magic Plate’’ in 10 MV radiation field Z. A. Alrowaili1,2, M. L. F. Lerch2, M. Carolan2,3, M. Petasecca2, P. Metcalfe2, A. B. Rosenfeld2 1

Introduction Patients with benign disease often have associated pain or functional deficiency which decreases their quality of life. Conventional treatments (such as surgery or steroids) are not always an effective option for some patients or the associated risks may not be acceptable. The use of radiation therapy for treatment of benign conditions is well established practise in many countries, particularly throughout Central and Eastern Europe.

Physics Department, Aljouf University, Aljouf, Saudi Arabia. Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]). 2Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 3Illawarra Cancer Care Centre, Wollongong Hospital, Wollongong, Australia

2

123

Australas Phys Eng Sci Med

P04 Quality management initiative in radiation oncology D. Banjade1, A. Mishra1, R. Hammond1, M. Fuller1 1

Fig. 1 The 40 9 40 cm2beam dose profiles measured by the CC13 IC and reconstructed doses at dmax 21 mm in case of 10 MV. The SwD used was 1 mm

Introduction The photon irradiation response of a 2D solid state transmission detector array Magic Plate (MP) mounted in a linac block tray is used to reconstruct the projected 2D dose map in a homogenous phantom along rays that diverge from the X-ray source and pass through each of the 121 detector elements. The aim of this study is to quantitatively demonstrate reconstruction of the radiation dose from the irradiation response of the MP operated in Transmission Mode (MPTM). Method The divergence of the radiation field with distance from the X-ray source must be taken into account when reconstructing the dose at any given depth in the solid water phantom from the response of the MPTM. The optimum (i.e. independent of field size) response-todose conversion factor for the central diode of MPTM was found at a specific depth in the phantom. This depth is referred to as the ‘‘Sweet Depth’’ (SwD) and the reconstructed dose at any position and depth can then be calculated [1]. Results Figure 1 shows a comparison of the measured (red line) dose profile using an ionization chamber and several reconstructed point doses (blue dots) at dmax from the MPTM data for a 10 MV photon beam. Excellent agreement exists between the two sets of in-field data. The percentage difference between the reconstructed dose and the measured dose was less than 3%. Conclusion We have demonstrated in a 10 MV linac field that based on MPTM response the dose can be reconstructed along rays projected from the photon source through each MPTM detector element and is in a good agreement with dose measured by IC.

Department of radiation Oncology, Central West Cancer Care Centre, Orange Hospital NSW 2800. ([email protected] [Presenting author]) Implementation of a quality management (QM) system in Radiation Oncology (RO) will improve work efficiency in radiotherapy (RT) and prevent the prospective risk associated in the process. Traditionally, the QM in RT has been focused on a device centric approach, however, many of the safety and quality issues in RT have been identifying as human factors content (?). Therefore, assurance of quality and safety of the department is not only the quality assurance (QA) of the equipment and instruments but also requiring implementation of the QA programs covering personnel and procedures as a whole. Founding an interdisciplinary team with full cooperation in the department to focus on mitigating the process-related errors can establish a risk-based QM program. Various activities, procedures and work performance can be strengthened by formulating QM system in RT emphasizing a proactive response to near misses rather than waiting for the events to occur, which is not acceptable in RT. Identifying the likelihood of occurrences, likelihood of failure being undetected and outcome severity of the errors event and preventing them on the paths of a fault tree analysis (FTA) process will prevent any harm to the patient during the RT process. The implementation of total QM approach with procedures, rules and regulations will detect each failure mode on each process and pave the way to deliver safe and quality treatment. This presentation will review the insight of the QM scenario in RT including error propagation and risk assessment. The QM tool of failure mode and effect analysis (FMEA) and FTA will also be discussed with examples. In addition the presentation will explore the initiative and practical approach of QM in Radiation Oncology.

References

References

1. Alrowaili, Z.A., et al., 2D mapping of the MV photon fluence and 3D dose reconstruction in real time for quality assurance during radiotherapy treatment. Journal of Instrumentation, 2015. 10(09): p. P09019.

1. Bruce Thomadsen-editor (2013) Quality and Safety in Radiotherapy: Learning the New Approaches in Task Group 100 and beyond. AAPM, Medical Physics Monograph No. 36

123

Australas Phys Eng Sci Med 2. Noel C. E., Santanam L, Parikh P. J. and Mutic S. (2014) Processbased quality management for clinical implementation of adaptive radiotherapy. Med Phys; 41(8) 3. Laura Atwell-Contributed (2010) Prescribing, Recording, and Reporting Intensity-Modulated Photon-Beam Therapy (IMRT). ICRU Report 83 4. A tripartite (RANZCR, AIR, ACPSEM) initiative (2013) Radiation oncology practice standards and Supplementary Guide

Introduction This work presents a derivation of a base function for GAF Radiochromic film. For GAF Radiochromic film, it has been demonstrated (Bennie and Metcalfe [1]) that the optical response with dose is linear to approximately 100 Gy: Dose ¼ Constant  ððUnexposed=ExposedÞ  1Þ Method The film was exposed to a known dose. Optical response was measured using an Epson V800 scanner. Results

P05 Flattening filter free planning: Is one beam line sufficient? R. P. Short1, J. Begg2 1

Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centres, Sydney, Australia. ([email protected]). 2Liverpool and Macarthur Cancer Therapy Centres, Sydney, NSW, Australia. ([email protected]) [Presenting author] Introduction The promise of Flattening Filter Free (FFF) treatments is for reduced treatment times, particularly for hypo-fractionated treatment regimens. Commonly known as Stereotactic Body Radiation Therapy, SBRT is an expanding treatment technique employed across a range of treatment sites ranging from the brain, prostate, lung, liver, and spine. The linear accelerator vendors offer two FFF beam lines, 6FFF and 10FFF. With the additional work involved in commissioning the accelerator and the treatment planning system, plus the on-going QA maintenance of a photon beam line, it is preferable that only one beam line be clinically available. 10FFF is preferred given its higher dose rate of *2000 MU/min versus 6FFF at *1200 MU/min. Method The treatment sites of brain, lung and prostate were chosen for this initial planning study, with five clinical cases chosen for each site. The clinical case was treated with a 6MV VMAT plan. Additional plans were developed using 6FFF and 10FFF and compared using local dosimetric protocols. All plans were delivered and beamon times recorded. Results All three beam lines generated clinically acceptable plans, with marginal differences between them for both target and OAR coverage. Beam-on time is largely inversely linear with maximum dose rate, so for example as the dose rate doubles, the beam-on time is halved. As such, the 10FFF beam-on time was 30 to 40% of the original 6MV plan. Conclusion Based on these clinical cases, 10FFF produces plans of the same quality as 6MV and 6FFF. In addition, 10FFF reduced the beam-on time by the largest amount. As such, only the 10FFF will be released clinically, unless there is dosimetric evidence to indicate a clinical benefit for 6FFF.

P06 Derivation of linear optical response of GAF radiochromic film N. Bennie1,2, P. Metcalfe1

Discussion Note at exposures greater than 100 Gy the graph tends to a saturation value. It would suggest for the linear portion that the conversion followed a function such as: Optical Mass ¼ lnð1 þ Constant  DoseÞ Substituting this into the equation for exponential optical attenuation will result in the raw optical output decreasing as a simple hyperbolic function as the dose increases. If the value of the transmission for the unexposed film is known, then the hyperbolic function can be manipulated to give the processed value above. This can be explained as follows: For the active medium of GAF film the equivalent of the grains in Silver Halide film are the crystals of Di-acetylene. However these crystals are not of equal size. An exponential type distribution is expected as related to the grain size distribution. The progression of the conversion of the active medium starts predominantly with the larger crystals, which contain a large proportion of the total mass, then progresses with the smaller crystals which contain a smaller proportion of the total mass. The overall result is that the progression is similar to the ln (1 + X) function.

1

Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. 2North Coast Cancer Institute, Lismore, NSW, Australia. ([email protected] [Presenting author])

Conclusion This can be demonstrated by modelling the conversion of an Exponential distribution and comparing with the ln (1 + X) function.

123

Australas Phys Eng Sci Med References 1. Bennie, N., & Metcalfe, P. (2016). Practical IMRT QA dosimetry using Gafchromic film: a quick start guide. Australasian Physical & Engineering Sciences in Medicine, 1–13.

P07 An integrated Monte-Carlo model for heterogeneous glioblastoma multiforme treated with boron neutron capture therapy Leyla Moghaddasi1,2, Eva Bezak2,3,4, Wendy Harriss-Phillips2,5 1

Department of Medical Physics, Adelaide Radiotherapy Centre, Adelaide, Australia. 2School of Physical Sciences, University of Adelaide, Adelaide, Australia. ([email protected]). 3Sansom Institute for Health Research, University of South Australia, Adelaide, Australia. 4 International Centre for Allied Health Evidence, University of South Australia, Adelaide, Australia. ([email protected]) [Presenting author]. 5Department of Medical Physics, Royal Adelaide Hospital, Adelaide, SA, Australia. ([email protected])

P08 The clinical implementation of an electron insert factor spline model using routine measurements S. Biggs1, M. Sobolewski1, J. Kenny2 1

Introduction Glioblastomas (GBM) of the brain, characterized by extensive infiltration into the brain, are highly resistant to radiotherapy. Boron Neutron Capture Therapy (BNCT), a biochemicallytargeted radiotherapy based on thermal neutron capture by 10B atoms, represents an alternative therapy. The objective of the current work is to develop a BNCT GBM treatment modelling framework to investigate the efficacy of BNCT in terms of cell Survival Fraction (SF) following a treatment. Methods The GBM BNCT modelling framework, developed in this work, is a cell-based dosimetry model (i.e. determining the dose deposited in individual GBM/normal cells) using GEANT4.9.6.p02, integrated with in-house developed Microscopic Extension Probability (MEP) and epithermal neutron beam models. The system was defined as a cubic phantom divided into 20 lm-aside voxels and irradiated with an epithermal neutron beam. Typical 10B concentrations in GBM and normal brain cells were obtained from the literature. Each cell was assigned a material composed of a brain (ICRP-based NIST database) material and a 10 B concentration depending on its MEP status. Heterogeneous radiosensitivity was simulated using a range of a/b values (linearquadratic model parameters) associated with different GBM cell lines. Results from the cell-based dosimetry model and the insilico GBM model with MEP distributions were combined to evaluate survival fractions (SF) for typical CTV margins of 2.0 & 2.5 cm. Calculated SFs were compared with those obtained for x-ray radiotherapy (XRT). Results and Conclusion Following BNCT treatment of heterogeneous hypoxic GBM, the calculated SFs within the beam region were smaller by more than two orders of magnitude as compared to XRT. However, compared to XRT, the change in SF as a result of CTV extension showed a reduction of only * 20% using BNCT. While BNCT results in more efficient cell kill, extension of the CTV margin may not increase the treatment outcome significantly.

123

Riverina Cancer Care Centre, New South Wales 2650, Australia. ([email protected]) [Presenting author]. 2Epworth Radiation Oncology, Epworth HealthCare, Victoria 3121, Australia Introduction There are many methods available to predict electron output factors however many centres still measure the factors for each irregular electron field. Previous work presented an electron output factor prediction model that approaches measurement accuracy – but uses already available data and is simple to implement [1]. This work presents the clinical implementation of an empirical spline model for output factor prediction that requires only the measured factors for arbitrary insert shapes.

Fig. 1 An approximation of the dependence of uncertainty on the amount of data collected given zero, one, or two hypothetical outliers with magnitude of +/-2% of measurement. Determined using the extended 12 MeV data set collected in [1]

Australas Phys Eng Sci Med BQ Check vs. Measured Energy Checks 0.742

10 MV Beam Quality

0.741

0.740 BQ Check 10 MV

0.739 Measured Plastic Water TRP20,10

0.738

0.737

0.736 13-Jun

14-Jun

15-Jun

16-Jun

17-Jun

18-Jun

Date

Fig. 1 BQ Check vs. Measured in plastic water

Method A clinical model was created by inputting previously measured electron insert factors and modelling them with respect to their patient shapes. These shapes were extracted from RT-DICOM exports and parameterised into equivalent ellipses taking into account changes in lateral scatter, bremsstrahlung produced in the insert material, and scatter from the edge of the insert. Results The maximum recorded deviation between measurement and prediction over the range of energies and applicators for which there was sufficient data was 1.0%. This is in agreement with the results of the previous work which showed that one may expect an approximate uncertainty of 0.5% (1SD) when as little as eight data points are used (Fig. 1) and recommends an optimum placement of these eight measurements (Fig. 2). Conclusion The level of accuracy combined with the ease with which this model can be generated demonstrates its suitability for clinical use. A full working demonstration of this method is available at http://simonbiggs.net/electrondemo. This demo works best in a modern browser such as Chrome or Firefox.

research1,2 to being commercially available for busy radiation oncology departments. To measure the beam quality of photons, measurements using the PTW Starcheck array are taken under the diagonal wedges of the phantom. A depth dose curve can then be created, from which the TPR20,10 calculation may be made by interpolation. A calibration factor relating the BQ check measurement to actual TPR20,10 is created, and is applied to all BQ Check measurements taken for routine QA. Method As an initial concept study, the beam quality of the 10 MV photon beam for an Elekta Axesse was measured using plastic water (TPR20,10), as well as using the BQ Check phantom. A comparison of the beam quality indices from the two methods was plotted. This study was subsequently extended to other energies and modalities. . Results The BQ check beam quality results were similar to the measured beam quality to within 0.5% over several days (see Fig. 1 for 10 MV results). The results were reproducible, and followed the energy change over time. Conclusion The BQ check phantom has shown strong potential to become the constancy check for photon beams. For electrons, the double Al wedges technique for energy checks is expected to be as sensitive as using a water-based methods over the energy range of 6 to 20 MeV2. (should the superscript 2 be there?)The BQ check would transform QA procedures in the radiation oncology department by shortening QA times, with a fast and simple set-up, especially for electron beam constancy checks.

References

References

1. Biggs, S., Sobolewski, M., Murry, R., & Kenny, J. (2015). Spline modelling electron insert factors using routine measurements. Physica Medica: European Journal of Medical Physics, Volume 32, Issue 1, 255 - 259. doi:10.1016/j.ejmp.2015.11.002.

1. M. K. Islam et al. (1993) A simple method of producing depth ionization data for electron energy constancy check. Med Phys. 20: 187 2. D. M. Wells et al. (2003) Electron energy constancy verification using a double wedged phantom. J Appl Clin Med Phys. 4: 3.

Fig. 2 An example set of the minimum required number of measurements that could be taken during commissioning for a 10 9 10 cm applicator. To produce model redundancy a few clinical shapes should also be measured

P09 Using the PTW BQ Check wedged phantom for energy constancy checks S. Ibrahim1, P. H. Charles1,2 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected]). 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected] [Presenting author]) Introduction This work introduces the PTW BQ Check, which allows the quick measurement of photon and electron beam energies using one simple measurement set-up for each modality. The concept of a double wedge phantom for energy checks has moved from

P10 A detailed analysis of the effects of source occlusion on dose distributions and measurements in small photon fields P. H. Charles1,2, M. A. Barry3 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected] [Presenting author]). 3Radiation Oncology Department, Calvary Mater Newcastle, Australia. ([email protected]) Introduction Small field dose distributions are complicated by lateral electronic disequilibrium (LED) and occlusion of the primary source

123

Australas Phys Eng Sci Med (PSO). Between differing linac designs the effects of LED are consistent (for the same nominal energy)[1]; unlike PSO effects, which are strongly influenced by linac design. The aim of this study is to explicitly quantify the effect of PSO using two common linac designs (Elekta and Varian). Method The EGSnrc Monte Carlo package was used to simulate cross-axis profiles and output factors in water from a Varian 21iX [2] for a number of square field sizes (side length from 5 to 50 mm). Simulations using a pencil beam as the initial electron source size (IESS) were used as a comparative gold standard. The linac simulations were altered so that parameters matched that of an Elekta Agility (E.g. the IESS changed from 1.2 mm [2] to 2.0 mm FWHM [3]). Combinations of source size, collimator position and field size were altered to find where PSO had negligible influence on small field dose distributions. Results For the 5 mm field size (depth = 5 cm, SSD = 95 cm); the distance from central-axis to 80%-dose in the cross-axis profile was decreased significantly by PSO (from 2.2 mm [ideal] to 1.7 mm [Varian jaws]). However, the figure below shows that on a Varian 21iX, PSO was negated for the 5 mm profile by using collimators at least 65 cm from the source. These results varied as function of spot size and field size. 1 Varian 21iX jaws Jaws 55 cm from source

0.8

Relave dose

Jaws 65 cm from source 0.6

Ideal (no PSO)

0.4

0.2

0 -0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

Introduction Radiology images are viewed for many reasons using displays of varying quality. The most important user is perhaps the radiologist and reporting requirements are well described by RANZCR [1]. Clinical review is also important, especially when the diagnostic report is not yet available. ACR-AAPM-SIIM [2] states ‘Images seen by technologists during acquisition, by radiologists during interpretation, and by physicians as a part of patient care should have similar appearance.’ Physicians often use consumer grade displays of unknown quality and calibration status. Clinical review requirements [3] can be difficult to achieve under high ambient lighting conditions. There is a real risk that abnormalities may be missed and patient management compromised. Method It has been demonstrated that DICOM calibration yields a more efficient and improved radiologist performance [4]. It is difficult to measure the performance of physicians viewing displays of variable quality in a range of ambient conditions. This paper compares a typical consumer display to a factory calibrated DICOM display (Philips C240P). Technical performance is analysed, and anecdotal comments from physicians are reported. Results The C240P uses In Plane Switching liquid crystal technology, with a 178 viewing angle, calibrated luminance of 300Cd/m2, and a factory DICOM mode. Typical consumer displays are based on Twisted Nematic technology with poor colour and contrast performance at high viewing angles, maximum luminance is around 230Cd/ m2, and software calibrations cannot be locked. The C240P DICOM mode is within 10% of GSDF, luminance and viewing angle performance are superior to the consumer display. An ambient luminance of 2.2Cd/m2 is assumed, higher than the recommended 1.59 minimum [3] but providing usable contrast from low greyscale values under high ambient lighting. Physicians report improved visualisation of anatomy and abnormalities, and improved angular viewing for group reviews. Conclusion The C240P gives improved image quality for clinical review when compared to a generic consumer monitor.

0.7

Off axis distance (cm)

Conclusion Reducing initial source size, using collimators further from the source and using larger field sizes all reduced the effects of PSO on dose distributions. Importantly, this study quantified exactly how PSO affected small field dose distributions from two common linacs, and provides explicit guidance on how it may be minimised in general linac design. References 1. C. McKerracher and D. I. Thwaites (2008). Phantom scatter factors for small MV photon fields, Radiother. Oncol. 86, 272–275 2. P. H. Charles, S. Crowe, T. Kairn, R. T. Knight, B. Hill, J. Kenny, C. M. Langton, and J. V. Trapp (2013), Monte Carlo-based diode design for correction-less small field dosimetry, Phys. Med. Biol. 58, 4501–4512 3. P. Francescon, S. Cora, and N. Satariano (2011), Calculation of kQclin,Qmsr for several small detectors and for two linear accelerators using Monte Carlo simulations, Med. Phys. 38, 6513–6527

References 1. RANZCR (2014), Standards of practice for diagnostic and interventional radiology version10 http://www.ranzcr.edu.au/ component/docman/?task=doc_download&gid=510 2. ACR-AAPM-SIIM (2014), Technical standard for electronic practice of medical imaging http://www.acr.org/*/media/ AF1480B0F95842E7B163F09F1CE00977.pdf 3. AAPM TG18 (2005), Assessment of display performance for medical imaging systems https://www.aapm.org/pubs/reports/ OR_03.pdf 4. Krupinski EA, Roehrig H, The influence of a perceptually linearised display on observer performance and visual search. Acad. Radiol, 7: 8–13. Doi: 10.1016/S1076-6332(00)80437-7

P12 Gamma KnifeÒ chamber-specific solid-phantomto-water correction factors E. C. Cosgriff1,2, K. Y. T. Biggerstaff3, P. H. Charles1,3

P11 Clinical review displays: What is reasonable image quality and how to achieve it N. J. Cook1 1

Medical Physics and Bioengineering, Christchurch Hospital, NZ. ([email protected] [Presenting author])

123

1

Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. 2Present address: Nepean Cancer Care Centre, Nepean Hospital, Kingswood, Australia. ([email protected]) [Presenting author]. 3Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia ([email protected]), ([email protected])

Australas Phys Eng Sci Med Introduction A proposed dosimetry protocol for small and nonstandard fields applicable to the Leksell Gamma Knife (Elekta Instruments AB, Stockholm, Sweden) requires the establishment of chamber-specific correction factors to account for the unique geometry and a solid phantom. Currently, only Monte-Carlo-calculated correction factors are available for a limited number of chambers and two phantoms. Correction factors for the Gamma Knife Perfexion unit were experimentally determined using five different ionisation chambers and three solid phantom materials. Method Chamber readings were obtained in liquid-water, Solid Water, poly(methyl methacrylate), and RW3 phantoms at a depth of 40 mm. Chambers were irradiated by a subset (a single beam sector) of the 192 60Co beams, which allowed the use of an open liquid-water phantom. A total of nine correction factors were calculated from a ratio of the corrected chamber reading in the liquid-water phantom to that in the solid phantom. Absolute dose measurements using the determined correction factors and four different chambers were carried out following a common protocol. Results Three correction factors obtained with the Solid Water phantom were found to agree the Monte-Carlo-calculated correction factors to within experimental uncertainties. The other six factors have been established for the first time. The average dose rate agreed with that determined using the Monte-Carlo-calculated correction factors, but the variation in the average dose rate was larger. Conclusion Experimental verification of the correction factors needed for absolute dosimetry contribute to the goal of establishing a universal dosimetry protocol for the Gamma Knife.

Fig. 1 Exploded view of assembly for measuring dose in the vicinity of a pacemaker

P13 Characterising and estimating doses near cardiac devices during radiotherapy treatments S. C. Peet1,2, R. Wilks1, T. Kairn2,3, S. B. Crowe1,2 1

Cancer Care Services, Royal Brisbane and Women’s Hospital, Herston, QLD, Australia. ([email protected]). 2 School of Chemistry, Physics, and Mechanical Engineering, Queensland University of Technology, Brisbane, QLD, Australia. ([email protected]), ([email protected] [Presenting author]). 3Genesis Cancer Care Queensland, Auchenflower, QLD, Australia. ([email protected]) Introduction Cancer and heart disease are both associated with the elderly, and so an ageing population will result in more cardiac devices being encountered in the radiotherapy clinic. This study characterised the dose delivered in the vicinity of a real pacemaker under authentic scatter conditions. Comparisons were then drawn with common dose estimate techniques including treatment planning system (TPS) calculations, out-of-field dose reference data, and invivo skin dose measurements. Method An implantable pacemaker was installed onto an anthropomorphic Rando phantom and supported by a 3D printed frame, allowing Gafchromic EBT3 film to be placed at the skin surface, anterior and posterior surfaces of the device, and 5 mm posterior to the device (Fig. 1). Square fields were then delivered at increasing distances, as well as several clinical treatment plans including volumetric modulated arc therapy, helical tomotherapy, and 3D conformal treatments. Results A small systematic difference was observed between the dose surrounding the battery and the circuitry of the device (Fig. 2). Square field deliveries revealed that the dose at the anterior surface of the

Fig. 2 Contrast boosted film scan following delivery of a clinical treatment plan device dropped exponentially with distance from the field. Both Eclipse and TomoTherapy TPS substantially underestimated the dose, whereas Oncentra TPS correctly calculated the dose for the 3D conformal plan. Skin dose measurements were approximately two times greater than the dose at the anterior surface of the pacemaker for all plans. Conclusion Treatment planning systems were found to be mostly unreliable in estimating the dose in the vicinity of a pacemaker. However, there was a predictable relationship found between in-vivo skin dose and pacemaker dose, and so in-vivo measurements are recommended during the beginning fractions of a treatment. References 1. Stovall, M., Blackwell, C. R., Cundiff, J., et al. (1995). Fetal dose from radiotherapy with photon beams: report of AAPM Radiation Therapy Committee Task Group No. 36. Med Phys, 22(1):63–82.

123

Australas Phys Eng Sci Med

P14 Clinical application of data mining: Evaluation and comparison of radiotherapy treatment plans for breast cancer

author]). 2Royal Brisbane & Women’s Hospital, Brisbane, Australia. ([email protected]), ([email protected]). 3Queen Alexandra Hospital, Portsmouth, United Kingdom. ([email protected])

T. Kairn1,2, S. B. Crowe2,3

Introduction Commercially available wearable glass beads have physical characteristics such as small physical size, inert nature and low cost which, combined with their thermoluminescent properties, make them suitable for in vivo dosimetry. The purpose of this study was to investigate the dosimetric characteristics of glass beads for electron beams and test their application as radiation detectors in invivo dosimetry for patients undergoing total skin electron irradiation treatment (TSEI). Method Glass beads manufactured in Japan (Mill Hill) with diameters of 3 mm were cleaned in an acetone solution in an ultra-sonic bath. Linearity, reproducibility, fading, and dose rate dependence were characterised using an electron beam produced on a Varian 21iX linear accelerator. Dose measurements were performed for a TSEI treatment using a Stanford six field technique, with OSLD measurements used for verification. A Harshaw 3500 TLD reader was employed for readouts. Results Radiation response differed from bead to bead, with a standard deviation in mean signal of 8%. Dose response was linear to approximately 6 Gy. The rate of fading was found to be 12% within 24 h after irradiations, whereas the signal decay observed when the readouts are done within the first 40 mins after irradiation was found to be within 0.9%. No dose rate dependence was observed, including at the high dose rate mode used for TSEI (approximately 2000 MU/ min). Large variations were observed between OSLD and glass bead measurements for the TSEI treatments, exceeding the calculated uncertainty of 7.7%. Conclusion Great care needs to be taken in the preparation and use of glass beads for dosimetry. Glass beads may be immediately suitable for the detection of gross errors in TSEI treatments, however, further work is required before these beads can be used for precision dosimetry.

1

Genesis Cancer Care Queensland, Brisbane, Australia. 2Queensland University of Technology, Brisbane, Australia. ([email protected]). 3Royal Brisbane & Women’s Hospital, Brisbane, Australia. ([email protected] [Presenting author]) Introduction This study describes a bulk, retrospective analysis of 1,137 breast and chest wall radiotherapy treatment plans as a demonstration of the potential breadth and value of the information that may be obtained from clinical data mining. Method The treatments evaluated in this study were planned at five radiotherapy centres belonging to one organisation, in 2013–2015. All DICOM plan, structure and dose files were simultaneously imported into Crowe et al’s treatment and dose assessor code (TADA) [1] and automatically evaluated, in a process which took nearly 60 hours, using a standard desktop PC. The production of clinically useful results required several modifications to be made to the TADA code, to redefine some structures (eg. separating left and right lungs, excluding markers from dosimetric analysis, etc.) and thereby clean up the ‘‘dirty data’’. Results The treatment plans exhibited relatively good agreement between treatment planning teams at different geographical locations, with all centres enthusiastically adopting a forward-planned IMRT (field-in-field) breast treatment planning technique (\1% of plans did not use this method). Plans generally met organ-at-risk limits well enough that tighter planning tolerances could be recommended in future. Heart doses calculated in left breast and chest wall treatments were significantly higher (p \ 0.001) than heart doses calculated in right breast and chest wall treatments, while the small differences in lung doses between the left and right breast plans were not so statistically significant, confirming the potential value (and lack of serious detriment) of using breath-hold techniques to increase the volume of lung separating the heart from the chest wall in left breast treatments. Conclusion The results of this study exemplify the use of bulk treatment planning data analyses to evaluate plan quality, inform ongoing treatment planning practises and provide general advice about the effects of anatomical asymmetry on achievable organ-atrisk sparing. References 1. Crowe, S. B. et al. (2013). Retrospective evaluation of dosimetric quality for prostate carcinomas treated with 3D conformal, intensity-modulated and volumetric-modulated arc radiotherapy. J. Med. Radiat. Sci. 60(4): 131–138.

P15 In-vivo dosimetry for total skin electron treatments using wearable glass beads S. K. Nabankema1, D. Binny2, S. M. Jafari3, S. R. Sylvander2, S. B. Crowe1,2 1

Queensland University of Technology, Brisbane, Australia. ([email protected]), ([email protected] [Presenting

123

P16 Comparative study of total skin electron therapy (TSET) treatments of Mycosis Fungoides A. G. Orth1, S. R. Sylvander2, S. B. Crowe1,2 1

Queensland University of Technology, Brisbane, Australia. ([email protected]). 2Royal Brisbane & Women’s Hospital, Brisbane, Australia. ([email protected]), ([email protected] [Presenting author]) Introduction Mycosis Fungoides (MF) is a form of T-cell Lymphoma and its most common type. MF cells can cover the entire skin surface up to a depth of a couple of centimeters, making it suitable for treatment with low-energy electron beams. The aim of this work is to investigate the effectiveness of Stanford 6-field, the 8-Field, the rotational and an additional method (a 12-Field method) in order to provide clinical recommendations for optimal patient setup geometries. Method OSLD and EBT3 film measurements were made on the surface and at depth, respectively, on a humanoid phantom positioned at approximately 400 cm from the radiation source, a Varian iX linac producing a 6 MeV electron beam at a high dose rate (2000 MU/min). The phantom was rotated manually to the 6-, 8- and 12-field treatment positions, and placed on a continuously rotating platform for the delivery of the rotating treatment.

Australas Phys Eng Sci Med Results There were no significant differences between the superficial doses and dose falloffs with depth, between the four treatment setups. The only observed difference between the treatment setups was an increase in measured bremsstrahlung photon dose at depth from the rotating treatment. Conclusion While the rotating phantom treatment was easier to setup and reproducibly deliver, it offered no obvious dosimetric advantage compared to the more-common 6- and 8- field treatments.

P17 Quantification of the Sensitivity Variation of PTW Octavius 1000 SRS for several photon beams R. de Chavez1, P. H. Charles1,2, P. O’Connor1 1

Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]). 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected])

Introduction The sensitivity variation of PTW Octavius 1000 SRS 2D array was investigated for flattening-filter (FF) and flatteningfilter-free (FFF) static beams at reference conditions. Moreover, the difference in the sensitivity of the central detector compared to the outer-most edge detector was also evaluated. Method The 1000 SRS array is a liquid-filled ionization chamber array with 977 detectors in an 11 9 11 cm2 measurement area. Each detector size is 2.3 9 2.3 9 0.5 mm (volume of 0.003 cm3) and their spacing in the inner area of 5.5 9 5.5 cm is 2.5 mm center-to-center, whereas in the outer area it is 5 mm center-to-center. Measurements were performed using an Elekta Agility linear accelerator in reference conditions (10 9 10 cm2 field size, SSD 90 and 10 cm effective depth, using plastic water build-up) with the central detector at the centre of the beam. The relative response was assessed for 6 and 10 MV FF beams and also 6 and 10 MV FFF beams. All measurements were repeated with the outer-most edge detector at beam central axis. The sensitivity correction factor was derived by getting the ratio of 1000 SRS with that of absolute dose measurements with a reference Farmer ion chamber (set up separately, under the same conditions). Results The values of the sensitivity correction factors (relative to Co-60 – provided by the manufacturer) were 0.96 for 6X FF, 0.93 for 10X FF, 0.98 for 6X FFF, and 0.97 for 10X FFF showing variation with beam energy. The relative response of the central and edge detectors were found to be identical. Conclusion The 1000 SRS was found to have sensitivity variation with the beam energies investigated. The derived sensitivity correction factors must be used to correct for beam quality. The relative detector response did not vary across the measurement area.

Introduction Microbeam Radiation Therapy (MRT) is a preclinical radiotherapy modality using micron-sized arrays of high brilliance, low divergence, for the treatment of tumours [1]. Typical MRT configurations consist of multiple high dose rate (*20 kGy/s) ‘peaks’ (25–50 lm FWHM) separated by lower dose rate ‘valleys’ (100–400 lm pitch). These conditions necessitate a dosimetry system with high spatial resolution to resolve peaks, and large dynamic range to monitor the large dose rate change between the peaks and valleys. Method The Centre for Medical Radiation Physics (CMRP) has produced a novel n-type silicon single strip detector – the 3D MESA SSD (Fig. 1) – that was designed to meet the requirements of MRT dosimetry. The device topology was examined at the Australian Institute of Innovative Materials (AIIM) using an electron microscope. Characterisation and dosimetry of an MRT field (50 lm FWHM, 400 lm pitch – denoted ‘50–400’) was performed at Australian Synchrotron Imaging and Medical Beamline (AS IMBL) hutch TM 2B after calibration using a Pinpoint ionisation chamber. Measurements were taken in an RMI-457 solid water phantom at 24 mm depth for a spectrum produced by a 1.4 T wiggler. Results Topological examination showed that devices matched the design and no noticeable defects from fabrication process were observed. A sample MRT profile obtained is shown in Fig. 2. Mean peak-to-valley dose ratio (PVDR) was calculated as (41 ± 5), and mean FWHM as (56 ± 2 lm). The FWHM is expected from convolution of the 50 lm Multi-Slit Collimator (MSC) slit width, beam divergence and SSD dimensions. Conclusion Novel n-type silicon detector, the 3D MESA SSD, was used to acquire MRT dose profiles at the AS IMBL hutch 2B. Excellent spatial resolution suggests these detectors are suitable for MRT dosimetry.

Fig. 1 (a) Schematic of 3D MESA SSD. (b) Topology of 3D MESA SSD

P18 Characterisation of a novel silicon strip detector for quality assurance of microbeam radiation therapy M. Cameron1, M. Petasecca1, A. Dipuglia1, J. A. Davis1, V. Perevertaylo2, A. B. Rosenfeld1, M. L. F. Lerch1 1

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected]), ([email protected]), ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]). 2SPA-BIT, Kiev, Ukraine. ([email protected])

Fig. 2 Left Measured 50–400 MRT dose profile. Top right Comparison of valley to baseline counts. Bottom right Snapshot of three individual peaks

123

Australas Phys Eng Sci Med 1

References 1. Lerch, M. Petasecca, M. Cullen, A. et al. (2011) Dosimetry of intensive synchrotron microbeams. Radiat Meas 46.12: 1560–1565. Acknowledgements CMRP thanks AIIM, ANSTO, and the ASIMBL team (A Stevenson, J Livingston, C Hall, and D Hausermann). CMRP authors acknowledge the support of the NH&MRC (APP1093256) and the Australian Synchrotron (AS153/IM/10045).

P19 Dose verification for liver target volumes using respiratory management 1,2

2

1

E. Dyce , D. Cutajar , S. Downes 1

Department of Radiation Oncology, Prince of Wales Hospital, Australia. ([email protected]). 2Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]) Introduction Respiratory motion during radiotherapy has a significant impact on the dose delivered to the target and surrounding tissue. Under motion, 4D imaging can aid localisation of the treatment volumes to be targeted. The accurate treatment of liver tumours near the diaphragm can be difficult due to the large respiratory movement generally found in this region, as well as the differing tissue densities at the liver/lung interface. The aim of this study was to establish the accuracy of dose delivered to liver tumours using an in-house respiratory phantom designed to emulate the lung/liver interface. Method The phantom consisted of adjacent slabs of water (liver substitute and lung equivalent materials) and a cam drive system to mimic respiratory motion of amplitude 2.6 cm. Cavities for an IBA Scanditronix cc04 ionization chamber and Gafchromic EBT3 film were used to perform dosimetric measurements. Three metal spheres, 5 mm in diameter were placed in the surrounding ‘liver’ material to represent surgical clips to assist with localisation during treatment alignment. An average sized PTV was defined on the 4D CT data set acquired with a Toshiba Aquilion LB scanner and a Varian RPM 4D tracking system with breathing rates of 10, 14 and 23 breathes per minute. Plans were calculated using an Elekta Monaco TPS on the phase average and exhale phase study sets for both conformal and VMAT techniques. Two generic organs at risk were created adjacent to the PTV to force modulation in the VMAT plan. Results/Conclusions The shape of the liver CTV was spherical and 4.5 cm in diameter. As a result of linear respiratory motion of 2.6 cm, the total PTV volume was defined to be 94.6 cm3 which is the approximate average PTV size of the patient cohort treated by the department. Full results of planned and measured 2D dose analysis will be presented.

P20 Data accumulation, processing and assessment for a clinical trial to establish models of kidney toxicity following radiotherapy J. Lopez Gaitan1, M. A. Ebert1,2, M. House1, P. Robins3, J. Boucek3, P. Podias2, K. Ciantar1, G. Waters2, T. Leong4, N. A. Spry2,5

123

School of Physics, University of Western Australia, Western Australia. ([email protected]), ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2 Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia. ([email protected]), ([email protected]). 3Nuclear Medicine, Sir Charles Gairdner Hospital, Western Australia. ([email protected]), ([email protected]). 4 Radiation Oncology, Peter MacCallum Cancer Centre, Melbourne. ([email protected]). 5School of Medicine and Pharmacology, University of Western Australia, Western Australia. ([email protected]) Introduction RAPRASI is an observational trial being conducted at Sir Charles Gairdner Hospital and the Peter MacCallum Cancer Centre [1]. Its aim is to utilise anatomical/functional imaging data to allow direct correlation of radiotherapy dose with spatio-temporal changes in kidney function. We report here on patient accrual, data accumulation, and developed approaches to image processing and data analysis. Method Participants with upper gastrointestinal cancer, undergoing conformal radiotherapy that will involve incidental radiation to one or both kidneys, are being recruited. Kidney function is being assessed at baseline and at time-points up to 78 weeks after the first radiotherapy fraction, using two measures: Glomerular Filtration Rate measured using 51Cr-EDTA clearance; and regional kidney perfusion via SPECT/CT imaged uptake of 99mTc- DMSA. Velocity AI and Matlab are being used to co-register subsequent SPECT images to the planned 3D dose distribution. SPECT signal intensity is being used as a measure of local renal function. Results 40 participants have been recruited, collectively having had 107 SPECT/CT follow-up scans. Due to participant mortality and withdrawals, only 6 participants have reached the planned final follow-up. Rigid registration has proven to be more reliable than nonrigid approaches. Follow-up images are being resampled into the planning CT coordinate space. All registered imaging data and planned dose distributions are being exported via DICOM to Matlab. Developed analysis tools include generation of histograms correlating dose, time-point, volume, SPECT signal and signal-change, as well as spatial representation of dose-signal correlation. Conclusion Methods for processing and analysing sequential functional imaging data have been developed and shall be used to generate kidney toxicity response models once follow-up has completed. The large number of follow-up datasets obtained over shorter follow-up periods will provide adequate study power for acute kidney response; late response will need to be inferred from other published studies on toxicity progression. Acknowledgements To Brenton O’Mara (Nuclear Medicine, Sir Charles Gairdner Hospital, Western Australia), Sean Bydder (Radiation Oncology, Sir Charles Gairdner Hospital, Western Australia), Julie Chu (Radiation Oncology, Peter MacCallum Cancer Centre, Melbourne) and Michael Hofman (Centre for Cancer Imaging, Peter MacCallum Cancer Centre, Melbourne). RAPRASI is funded by Priority-driven Collaborative Cancer Research Scheme (PdCCRs) grant 10027005 from Cancer Australia and the Australian Commonwealth Department of Health and Ageing, Radiation Oncology Section Reference 1. Lopez-Gaitan et al. (2013) Radiotherapy of abdomen with precise renal assessment with SPECT/CT imaging (RAPRASI): design and methodology of a prospective trial to improve the understanding of kidney radiation dose response. BMC Cancer, 13:381

Australas Phys Eng Sci Med

P21 Comparison of SH EPR-Alanine pellet dosimeters and TLD100H for pre-treatment verification in stereotactic ablative radiotherapy N. Esen1,2, R. Prabhakar1,2, C. Smith1, J. Hagekyriakou2, M. Geso1 1

RMIT University, Melbourne, Australia. 2Peter Mccullum Cancer Centre, Victoria, Australia. ([email protected] [Presenting author]) Introduction The aim of this work is to compare the characteristics of thermoluminescent dosimeters (TLD100H) and synergy health (SH) EPR alanine pellet dosimeters for small field dosimetry and explore the feasibility of using these dosimeters for pre-treatment verification of the dose delivered to patients undergoing stereotactic ablative radiotherapy (SABR). Methods The dose linearity, angular dependence, energy dependence and dose rate dependence of SH Alanine-EPR pellets dosimeters and TLD100H prior to application for SABR patient were studied for 6 MV X-rays. The Bruker EleXsys E500 EPR spectroscopy of 9.5 MHz was used to read the Alanine pellet dosimeters signals and the HarshawQS 5500 automatic TLD reader was used for reading the TLDs. The alanine pellet and TLD dosimeters were placed inside a phantom made of Perspex and was used to perform pre-treatment patient specific quality assurance. Five different SABR cases including Lung, Liver, Scapula, Sternum and Spine were included in this study. The measured dose from SH EPR alanine and TLD dosimeters were then compared to the dose calculated from the eclipse treatment planning system. Results The relationship between dose measured using Alanine-EPR spectroscopy and TLD followed a linear curve (R2 = 0.998) and no signification difference with dose rate and energy was observed for alanine. The differences between the measured and the TPS computed dose were less than 2 and 3% with Alanine-EPR and TLD respectively. Conclusions Our study shows that the SH EPR-Alanine dosimeter and TLD are consistent and agree well between the measured and the calculated dose. It also confirmed that both alanine and TLD are valuable dosimeters for hypo-fractionated radiotherapy pre-treatment quality assurance.

compared the occupational radiation doses to all monitored individuals and professional groups in these departments. Method A total of 6067 dose records of personal torso TLD and InstaDose USB-type dosimeters worn by the monitored workers were registered in the period from 2011 to 2015. The dose records of shortterm (\10 months) staff were excluded from the study. Annual effective whole-body doses were calculated in individuals and compared among 9 professional groups. Extremity monitoring was also provided for the workers involved with high radiation activity procedures. All equivalent dose records of finger TLDs were included for analysis. Results There were no cases of effective whole-body doses exceeding the annual limit of 20 mSv or equivalent doses to fingers exceeding the annual limit of 500 mSv during the 5-year period. The average annual dose of all measurable records was 2.41 mSv in MIaT and 0.40 mSV in Radiology. The radiopharmacy group received the highest exposure with the annual whole-body dose of 4.29 ± 1.52 mSv and finger dose of 193.11 ± 111.97 mSv. Nurse and medical imaging technologists (MIT) were the largest exposed groups with significant individual differences in radiation exposure. MITs in PET received the highest annual doses of 2.86 ± 1.00 mSv, compared with 1.91 ± 0.65 mSv in NM, 0.35 ± 0.23 mSv in Bone Mineral Density Unit and 0.10 ± 0.20 mSv in Radiology. Nurses working in MIaT received annual dose of 1.11 ± 0.65 mSv, significantly higher than those of Radiology nurses (0.13 ± 0.29 mSv). Conclusion Radiation exposure control and continuous occupational dose reduction have been achieved in the investigated departments. The waiting time in identifying the exceptional high effective doses was minimised by employing a new monitoring technique, and therefore correction of the related work practice could be implemented in a more timely fashion.

P23 A convenient verification method of the entrance photo-neutron dose for an 18MV medical LINAC using silicon p-i-n diodes V. Gracanin1, L. Dermikol1, S. Guatelli1, L. Tran1, M. Lerch1, D. Cutajar1,2, R. Gupta2, J. Yuen2, I. Cornelius3, R. Preston4, V. Prevertaylo5, A. B. Rosenfeld1 1

P22 Assessment and comparison of occupational radiation exposure in radiology, molecular imaging and therapy: A 5-year review Sylvia J. Gong1, Lisa Mong2, Danielle Thom3, Paul U2, Graeme J. O’keefe1 1

Department of Molecular Imaging and Therapy, Austin Health, Melbourne, Australia. ([email protected] [Presenting author]). 2Department of Medical Physics, Austin Health, Melbourne, Australia. 3Department of Radiology, Austin Health, Melbourne, Australia. Introduction In compliance with the key aspects of ARPANSA regulations in protecting workers against the hazards of ionising radiation, employees exposed to radiation from diagnostic imaging and therapeutic procedures in the Department of Molecular Imaging and Therapy (MIaT), Department of Radiology, and Surgery and Endoscopy Centre were regularly monitored. This study assessed and

Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected] [Presenting author]), ([email protected]). 2St George Cancer Care Centre, St George Hospital, Kogarah, Australia. ([email protected]), ([email protected]). 32MRD, 196 Faraday St, Carlton, Victoria, Australia. ([email protected]). 4Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia. ([email protected]). 5SPA BIT, Kiev, Ukraine. Introduction Electron Linear Accelerators (LINAC’s) used in radiotherapy treatments produce undesired photo-neutrons when operated at energies [8 MeV. [1,2], which contaminate the therapeutic beam. The biological effects produced by photo-neutrons are difficult to estimate in a mixed pulsed photon-neutron radiation field, passive detectors have been recommended to avoid instrumental problems present with active devices [3]. Method Passive silicon p-i-n diodes, for real time neutron dosimetry based on Si-displacement KERMA estimation in mixed photon– neutron field of an 18 MeV LINAC were investigated. Bulk p-i-n diodes with a base length of 1.2 mm and planar p-i-n diodes with different geometries of the base, N+ & P+ electrodes were investigated.

123

Australas Phys Eng Sci Med

Fig. 1 (a) Schematic of the experimental and simulation setup. (b) Position of the silicon p-i-n diodes on the surface of the solid water phantom Each detector was calibrated in separate: neutron, photon and electron fields to determine their sensitivity factors. A Geant4 simulation was developed to model an 18 MeV linac and study the response of the p-i-n diode to photons, neutrons, electrons and calculate the neuron Displacement KERMA and neutron dose equivalent at different depths in a solid water phantom was measured (Fig. 1a,b). Results The detectors showed excellent discrimination between neutrons and photons, with sensitivity to fast neutrons being *103 times higher than photons per 1 Gy(TE). Planar p-i-n diodes have demonstrated the possibility of obtaining a wide range of sensitivities in a neutron field, but were limited by their high temperature sensitivity. Bulk p-i-n diode response to neutrons, was found to be most sensitive at the surface of the phantom. Good agreement between simulation and experimental results was observed. Conclusion The p-i-n diode neutron sensors were found to have a high neutron to gamma discrimination utilizing a simple readout method which is performed directly after irradiation. The bulk p-i-n diodes with size 1 mm3 were found to be convenient for in field fast estimation of the neutron dose equivalent. References 1. Esposito A, et.al (2008) Determination of neutron spectra around an 18MV medical LINAC with a passive Bonner sphere spectrometer based on gold foils & TLD pairs. Radiation Measurements 43: 1038–1043 2. D’Errico F, et.al (1997) Advances in superheated drop bubble detector techniques. Radiation Protection Dosimetry. 70: 103–108 3. AAPM (1986) Neutron measurements around high energy x-ray machines, Report 19

P24 Estimation of the influence of radical effect in the proton beams K. Haneda Department of Clinical Radiology, Hiroshima International University, Japan. ([email protected] [Presenting author]) Introduction The purpose of this study was to estimate an impact on radical effect in the proton beams using a combined approach with physical data and gel data. The study used two dosimeters: ionization chambers and polymer gel dosimeters. Polymer gel dosimeters have specific advantages when compared to other dosimeters. They can

123

measure chemical reaction and they are at the same time a phantom that can map in three dimensions continuously and easily. Method The projection particles were protons with a kinetic energy of 210 MeV. An SOBP of length 6 cm was formed with a bar ridge filter made of aluminium. First, a depth-dose curve for a 210 MeV proton beam was measured using an ionization chamber and a gel dosimeter. Second, the spatial distribution of the physical dose was calculated by the Monte Carlo code system PHITS: To verify the accuracy of Monte Carlo calculation, the calculation results were compared with experimental data measured using an ionization chamber. Last, to evaluate of the rate of the radical effect against the physical dose. Results The simulation results were compared with the measured depth-dose distribution and showed good agreement. The spatial distribution of a gel dose with threshold LET value of proton beam was calculated by the same simulation code. Then, the relative distribution of the radical effect was calculated from the physical dose and gel dose. The relative distribution of the radical effect was calculated at each depth as the quotient of relative dose obtained using physical and gel dose. Conclusion This approach showed good agreement with the general RBE in SOBP proton beam. The approach may predict proton RBE.

P25 ACPSEM focus newsletter: Old-fashioned initiatives to improve communication K. M. Harrison Genesis Cancer Care, Newcastle NSW. ([email protected] [Presenting author]) Introduction In 2014, the first edition of a combined-branch newsletter was compiled and distributed. The purpose of this was to improve dissemination of information regarding ACPSEM activities to members in quarterly editions. This poster aims to provide an outline of how the newsletter is coordinated. Method The editorial responsibility rotates around the branches to distribute the workload and provide an opportunity for regional characterisation. Results To date, 9 issues of the newsletter have been produced which has provided a forum to inform and celebrate many ACPSEM activities, particularly in the area of fundraising and philanthropy. Conclusion The ACPSEM Focus newsletter is a successful venture and there are many opportunities for members to be involved in the future direction Acknowledgements Nick Hardcastle and Duncan Butler

P26 Use of an effective multidisciplinary approach in implementing the Brainlab Elements software from a management perspective R. Hill1,2, J. Morales1,3, J. Poder1,4, M. Butson1, P. Estoesta1, L. Delaney1, P. Aston1, L. Rezaei1, N. Patanjali1, J. Martyn1, B. Jonker5 1

Department of Radiation Oncology, Chris O’Brien Lifehouse, Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]),

Australas Phys Eng Sci Med ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Institute of Medical Physics, School of Physics, University of Sydney, Australia. 3School of Chemistry, Physics and Mechanical Engineering, QUT, QLD, Australia. 4Centre of Medical Radiation Physics, University of Wollongong, Australia. 5 Central Neurosurgery, Camperdown, NSW, Australia. ([email protected]) Introduction The implementation of new technology into a radiation oncology department is a complex process. There are many factors that require consideration including appropriate commissioning, safe implementation, development of protocols, training and the involvement of all the disciplines. This work describes the process for the recent successful implementation of the Brainlab Elements module for multiple brain metastases by a multidisciplinary team in the Department of Radiation Oncology, Chris O’Brien Lifehouse. Method The project plan and documentation was developed based on standards required to conform to Novalis Circle certification and the Tripartite Radiation Oncology Practice standards. A multidisciplinary team (MDT) was established consisting of medical physicists, radiation therapists (RTs), radiation oncologists and a consultant neurosurgeon and given responsibility for implementing the Elements software. The commissioning plan for Elements was developed by the medical physicists based on current SRS commissioning recommendations as no specific literature is available on the performance or characteristics of this new treatment technique. The RT trainer was included in the commissioning team to ensure ongoing clinical support and training for the RTs. A weekly MDT was scheduled to ensure timely implementation and to ensure all professions provide input into the development of the protocols and reviewing progress. Results By involving multiple members of each of the professions, rapid progress was able to be made through the required steps. The commissioning tests, protocols and training were all completed in time for the first patient. An important factor in the success of the project was setting meeting times to accommodate the staff specialists to ensure their input was readily available. The final result was the successful treatment of the first patient who had four brain metastases. Conclusion A multidisciplinary approach to setting up new technology has been shown to be successful in achieving treatment outcomes and conformity to various practice standards.

Fig. 1 Left – Experimental setup. Right – Geometry-based target identification (blue) from 20 candidates (red) Method Our previous 3D approaches were based on single target tracking with optionally state-dependent patterns. This was enhanced by separating anatomical structures into a network of nC3 smaller target volumes. The resulting target network allows for 6dof localization and offers the possibility for deformation assessment. We identify a set of k candidates per target inside the network which are then validated against a model of the original target shape including an uncertainty threshold for target deformation. The 6dof position of the matching networks is calculated using [4]. In a final step, the networks are either ranked by their reprojection errors (6dof) or their single target matching qualities (deformation) to determine the best match. An industrial robot moving an ultrasound phantom with 4 targets (Fig. 1) to 1000 positions (±10 degrees, ±20 mm) was used for 6dof accuracy assessment. Results Qualitative analysis confirmed the feasibility of the proposed method. For latencies below 5 ms we used small templates and k = 5 candidates. This fast, the geometry validation was successful in 94% of the cases with 0% false positives. The mean matching error was 0.76 mm (with 0.22 mm precision). Conclusion We have developed a low-latency, highly accurate method for 6dof ultrasound tracking. In a next step the tracking information will be used for landmark-based deformation of the target volume, projection into beam’s eye view and compensation with the MLC. References

P28 Towards 6dof tracking of deformable objects for 4D ultrasound-guided radiation therapy S. Ipsen, R. Bruder, A. Schweikard Institute for Robotics and Cognitive Systems, University of Luebeck, Germany. ([email protected] [Presenting author]), ([email protected]), ([email protected]) Introduction Motion compensation is an essential topic in modern image-guided radiotherapy. We have modified a 4D ultrasound station for real-time tracking of anatomical structures [1]. This system was interfaced with a CyberKnife (Accuray Inc.) [2] and an MLC tracking system (Varian Medical) [3] for 3D motion compensation. However, both systems hold the potential for six-dimensional (6dof) motion compensation, MLCs might even handle deformation.

1. Ralf Bruder, Floris Ernst, Alexander Schlaefer and Achim Schweikard, A Framework for Real-Time Target Tracking in Radiosurgery Using Three-dimensional Ultrasound, in: Proceedings of the 25th International Congress and Exhibition on Computer Assisted Radiology and Surgery (CARS’11), Berlin, Germany, pages S306–7, 2011 2. Oliver Blanck, Philipp Jauer, Floris Ernst, Ralf Bruder and Achim Schweikard, Pilot-Phantomtest zur ultraschall-gefu¨hrten robotergestu¨tzten Radiochirurgie, in: 44. Jahrestagung der DGMP, DGMP, Cologne, Germany, pages 122–3, 2013 3. Svenja Ipsen, Ralf Bruder, Rick O’Brien, Paul Keall, Achim Schweikard and Per Poulsen, First integration of online 4D ultrasound guidance with MLC tracking for real-time motion compensation in radiation therapy, abstract submitted to EPSM 2016, Sydney, Australia. 4. Berthold K. P. Horn, Closed-form solution of absolute orientation using unit quaternions, in Journal of the Optical Society of America. Vol. 4, pages 629–42

123

Australas Phys Eng Sci Med

P29 Comparison of volumes calculated by the Leksell Gamma Plan (LGP), Eclipse, Monaco, Velocity and Pinnacle treatment planning systems (TPS) C. E. Jones, L. K. Webb Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]) Introduction When evaluating radiation therapy plans, the oncologist relies on structure volumes and dose volume histograms (DVH) calculated by the TPS. In this study, we compare structure volumes calculated by five commercial TPS’s to determine whether inter-TPS variations may affect apparent structure size and DVH’s. Method The AAPM TG244 head & neck (Case 5) IMRT/VMAT validation dataset was downloaded from AAPM webpage1. This consisted of a CT and structure dataset in DICOM format. The CT and structures were imported into the LGP, Eclipse, Monaco, Velocity and Pinnacle TPS’s, where the volumes of the structures were computed. The volumes of these 33 structures range from the order of 0.1 (LT LENS) to 10000 (BODY) cm3. In addition, an inhouse calculation of structure volume was performed where the volume included the entire volume of voxels on the surface of the contour (the simplest volume calculation). Results LGP usually provided the lowest calculated volume, followed by Eclipse (although in 6 of the 33 cases Eclipse was lower). Monaco and Velocity provided similar volumes. Pinnacle returned the highest volume calculation in all cases. Comparison of the Pinnacle and inhouse calculated volumes indicated that Pinnacle ‘‘rounds-up’’ the volume by including the full size of surface voxels (i.e. voxels intersected by the structure’s contour), while the other four TPS’s compute a partial volume for the voxels intersected by the structure’s contour. Small structures were affected to a greater extent. For example comparison of the smallest (LT LENS) and largest (BODY) between the two extreme TPS’s (LGP and Pinnacle) were 52.6 and 0.9% respectively. Conclusion The discrepancy between volume calculated by different TPS’s is due to the treatment of the surface voxels. Therefore small structures (with a larger surface to volume ratio) are affected to a greater extent. This should be considered when evaluating treatment plans. Reference

Method Increasingly computationally-expensive simulations of a photon beam through a rectilinear Perspex phantom were performed, and the simulation times recorded. The photon source and irradiated volume remained identical. These simulations are as follows. • • •

•

•

Single Perspex detector with 1 cm margin of photon cross section enhancement (XCSE). Single detector with margin of XCSE that would enclose a 2D array of the 1681 of the above detectors (volume = 121 cm3). As above with square matrices of up to 1681 detectors,but scoring only performed in the central detector. (I.e. the simulation volume being divided into increasing number of regions). As above with scoring performed in all 1681 detectors, but with the Russian Roulette (RR) cavity only enclosing the central detector. As above with a union of RR cavities, so that all the detectors are enclosed by an RR cavity.

Results The simulation time of the 1681 detector array was over 4,000 times greater than for a single detector. In descending order, the contributions to the increased simulation time were: • • • •

Large number of regions (85). Large XCSE volume (10). Scoring of multiple detectors (2.5). Large multiple-region RR cavity (2).

Conclusion Simulation times for realistic 2D array is much larger than those corresponding to a single identical detector. The dominant contributor is the large number of regions in the simulation volume. Reference 1. Wulff J, Zink K, and Kawrakow I. Efficiency improvements for ion chamber calculations in high energy photon beams. Medical Physics, 35(4), pp 1328–1336 (2008)

P31 Impact of tissue heterogeneity on dose calculation in skin brachytherapy. An initial investigation of the new TG-186 dose calculation algorithm R. Jones, P. Simpson

1. www.aapm.org/pubs/MPPG/TPS/ Calvary Mater Hospital, Newcastle, Australia. ([email protected] [Presenting author]), ([email protected])

P30 Evaluation of usage of the EGSnrc, egs_chamber user code for simulations through 2D arrays C. E. Jones Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]) Introduction In the event of discrepancy between TPS-calculated and measured dose distributions, Monte Carlo simulations of beams through arrays may enable separation of TPS dose calculation errors and machine delivery errors. The EGSnrc, egs_chamber1 user code may enable efficient calculation through arrays. In this study, we compare simulation times of a realistically-sized, multi-detector array to the corresponding time of a single identical detector in the same simulation volume.

123

Introduction Skin Brachytherapy is a routine method of treating skin lesions. Conventionally, treatments have been performed using a HDR 192Ir source, with dwell times calculated using the TG-43 treatment planning algorithm. This algorithm assumes total scatter conditions; in practice this does not reflect the clinical environment. The introduction of the collapsed cone algorithm TG-186, seeks to improve the accuracy of brachytherapy dose calculations by incorporating tissue heterogeneity information. This work compares the two treatment planning algorithms for two simple cases and for a clinical skin brachytherapy treatment plan. Method A virtual water phantom (40 9 40 9 40 cm3) surrounded by an air volume (60 9 60 9 60 cm3) was created within the Elekta Oncentra Brachytherapy treatment planning system. Using a Flexitron 192 Ir source model, a single dwell was positioned centrally on the phantom surface. Dose was calculated for this scenario and then repeated for the addition of 2, 3.5, 5 and 10 cm backscatter material. A second simple case was planned with full scatter conditions and

Australas Phys Eng Sci Med

P31A Novel moving MR and LINAC compatible platform for moving target treatment verification and simulation A. George1, Paul J. Keall2, Ricky T. O’Brien2, Gary Liney3,4,5,6, Shrikant Deshpande1, Lois Holloway5,6,7,8 1

Cancer Therapy Department, Liverpool Hospital, Sydney Australia. ([email protected]), ([email protected]). 2Sydney Medical School, University of Sydney, Camperdown, New South Wales, Australia. ([email protected] [Presenting author]), ([email protected]). 3Liverpool and Macarthur Cancer Therapy Centres, Australia. 4Ingham Institute of Applied Medical Research, Australia. 5South Western Sydney Clinical School, University of New South Wales, Australia. 6Centre for Medical Radiation Physics, University of Wollongong, Australia. ([email protected]). 7Ingham Institute and Liverpool and Macarthur Cancer Therapy Centres, Liverpool, Australia. 8 Institute of Medical Physics, School of Physics, University of Sydney, Camperdown, Australia. ([email protected]) Introduction Tumour motion during hypo-fractionated radiotherapy is a major challenge for accurate dose delivery and treatment verification is performed using 4D commercial phantom (add reference). These phantoms are designed for conventional linac based treatment and not compatible in magnetic environment. This work reports development and characterisation of prototype moving platform design and software interface which is treatment modality neural. This prototype can be used in conjunction with any commercial dosimeter and phantom. This will give us the ability to choose; • • • • •

Range of movement the motion to be any function set ultimately to program any patient breathing pattern any phantom to be moved any insert of a phantom

Method The first version was made using the driving mechanism of a wedge of an Elekta linac. It was successful and used in a 3T MRI. The second version, a Polulu stepper motor (USYD) was used in conjunction with a 2D stage (Siemens). In testing, it is capable of moving the ArcCheck with ease. Yet to be tested on LINAC and MRI. Future of the device to move in 3D ie x,y as well as rotation. Results The phantom was found to be functional in the presence of an MR field with a phantom of weight up to 30 kg. The accuracy of the

phantom was found to be within 1 mm in the x direction and the reproducibility to be within 1 mm. Conclusion A prototype MRI compatible moving platform was created which can be used with a variety of phantoms to simulate breathing during patient specific QA. Currently the movement is 1D, only a 10 cm and restricted to limited phantoms. Further developments would include expanding the movement to 2D/3D, with different programmable amplitudes and frequency of movements.

P32 ACDS level Ib audits: FFF, directly measured kQ S. Keehan, J. Lye, A. Alves, M. Shaw, F. Gibbons, I. Williams Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction The ACDS offers on-site linac output checks for new linacs. The TRS-398 protocol is used to determine dose to water in the facility water tank using ACDS equipment. The level Ib audit has been updated to include FFF dosimetry. ACDS has also moved to using directly measured chamber specific kQ values, as recommended by TRS-398. Method The suitability of dosimetry protocols for FFF beams has been investigated and the ACDS has implemented a non-uniformity correction, kn, to be used with the TRS-398 protocol [1]. Results On-site measurements of 23 6FFF and 13 10FFF beams have been performed by the ACDS. The average dose variations shown in Fig. 1 for 6FFF and 10FFF beams were -0.7 and -1.0%. A survey of facilities found that everyone was using the TRS-398 protocol without any additional corrections for FFF beams. The bias in the results can be explained by the ACDS use of a beam profile non-uniformity correction (6FFF: 0.3% and 10FFF: 0.45%) and a random small bias in the dataset. The random bias can be corrected using the offset in the corresponding beam with the flattening filter (see Fig. 2). The average difference between chamber specific kQ and TRS-398 tabulated kQ for ACDS’s 11 farmer chambers is up to 1.2% for a TPR20,10 of 0.758. Conclusion The ACDS now offers LIb audits of FFF beams. The ACDS calculates dose to water using chamber specific kQ measured during calibration by ARPANSA. 3.0

6FFF 10FFF

Pass (Action Level) Out-of-tolerance

2.0 Dose variaition (%)

dose was calculated using the two algorithms and compared. Finally, a clinical shoulder skin brachytherapy case was calculated with each algorithm and compared. Results A notable difference in dose calculated between TG-43 and TG-186 was found for the first simple case. With air abutting the source, TG-186 calculated dose 6% lower than TG-43 for a point located 6 cm beneath the water surface. The magnitude of this difference steadily decreased as backscatter material was added. In the second case, the dose calculated by the two algorithms agree at the water surface but some calculation difference was apparent at depth. For the complex clinical skin brachytherapy case, a dose calculation difference of [3% between the algorithms was found. Conclusion The new TG-186 algorithm shows a marked decrease in calculated dose under incomplete scatter conditions compared with TG-43. This is an important consideration in the clinical environment, particularly in skin brachytherapy.

1.0 0.0 -1.0 -2.0 -3.0

Random

Fig. 1 The dose variation ((Facility stated – ACDS measured)/ACDS measured) for on-site measurements of FFF beams

123

Australas Phys Eng Sci Med 3.0

Pass (Action level)

6FFF/kn - 6FF 10FFF/kn - 10FF

Pass (Action Level) Out-of-tolerance

4.0

1.0

Dose variation (%)

Dose variaition (%)

2.0

0.0 -1.0

0.0 -2.0

-6.0 6

Random

Fig. 2 The dose variation was recalculated without the non-uniformity correction applied to the ACDS measured dose and adjusted for any offset in the corresponding beam with the flattening filter Reference 1. Lye JE, Butler DJ, Oliver CP, Alves A, Lehmann J, Gibbons FP, Williams IM (2016) Comparison between the TRS-398 code of practice and the TG-51 dosimetry protocol for flattening filter free beams. Phys. Med. Biol. (accepted)

P33 An empirical correction factor for using OSLDs in a perspex phantom for absolute electron beam dosimetry S. Keehan, A. Alves, J. Lye, M. Shaw, F. Gibbons, I. Williams Australian Clinical Dosimetry Service (ACDS), ARPANSA. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected]) Introduction The ACDS conducts both on-site (level IB) and mailout (level I) audits for electron beams. The on-site audit uses a Roos chamber to measure electron dose and the mail-out audit uses OSLDs in perspex phantoms. Method A number of correction factors must be applied to the OSLD signal in order to obtain the dose to water under full scatter conditions. Each of these correction factors has an associated uncertainty which contributes to the overall uncertainty in an OSLD measurement. A review of 687 electron beam audits has been used to assess the accuracy of the ACDS OSLD audit. (does this mean 687 linacs?) Results The average of the dose variation of 285 electron beams measured with Roos chambers by the ACDS in level IB audits is 0.3% with a standard deviation of 1.0%. This indicates that the offset from zero is not statistically significant and implies there is no nationwide systematic dosimetry error or bias affecting electron beams. The dose variation offsets observed in the level I OSLD audit results shown in Fig. 1 are therefore an artefact of the ACDS calculation of dose to water from the OSLD measurement of dose in a perspex phantom. Conclusion A correction factor determined from the average offset for each electron beam energy will be applied to the dose calculations to improve the accuracy of the ACDS mail-out audit.

123

2.0

-4.0

-2.0 -3.0

Out-of-tolerance

6.0

8

9 10 12 15 16 Nominal electron beam energy (MeV)

18

20

Fig. 1 The average dose variation for electron beam audits. A total of 687 electron beam OSLD audit results are included. The error bars are the standard uncertainties. The largest uncertainties are associated with the 8 and 10 MeV beams, which are not commonly found in Australia. Those points are the average of only ten beams each

P33A CT image simulation of XCAT phantoms for quantification of deformable image registration errors J. R. Supple1, S. Siva2, M. L. Taylor1, T. Kron1,3, R. D. Franich1 1

School of Applied Sciences and Health Innovations Research Institute, RMIT University, Melbourne, VIC, Australia. ([email protected]), ([email protected]), ([email protected] [Presenting author]), ([email protected]). 2Department of Radiation Oncology, Peter MacCallum Cancer Centre, East Melbourne, VIC, Australia. ([email protected]). 3Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne, VIC, Australia. Introduction Quantification of deformable image registration (DIR) spatial errors is a difficult task, primarily due to the lack of a known ground truth deformation between patient images. Mathematically defined phantoms incorporating deformation models, for example XCAT (Segars, 2010), may be used to overcome this limitation with the drawback of the phantoms being defined by precise surfaces and omitting features such as tissue inhomogeneities and noise inherent to the imaging system. The exclusion of these features affects the quality of the DIR result and so conclusions drawn about the accuracy of a DIR algorithm are not transferrable to patient data. Here we present an image processing method for incorporating realistic image features with XCAT phantoms allowing meaningful, high resolution quantification of DIR spatial errors. Method Respiratory correlated 4D planning CTs of the torso were simulated. An average frame spanning 0.2 s was output from XCAT to incorporate breathing and cardiac motion over the typical acquisition time for a single phase. Tissue inhomogeneities were then introduced by applying a multiplication filter made from a smoothed, dilated salt and pepper noise. Finally, quantum noise is simulated by adding Gaussian noise such that the standard deviation of voxel values (incorporating variations from the salt and pepper noise) for a given tissue match that typically observed in patient images. Results Figure 1 shows three orthogonal slices through a simulated 4DCT image phase at end-exhalation.

Australas Phys Eng Sci Med

Fig. 1 Three orthogonal slices through a simulated 4DCT image phase at end-exhalation Conclusion Image processing of XCAT attenuation phantoms was used to simulate 4DCT image phases. The ground truth deformation maps output by the phantom software can then be used to quantify the DIR spatial error arising from performing registration of the simulated images. Reference 1. Segars W, Mahesh M, Beck T, Frey E, Tsui B (2008) Realistic CT simulation using the 4D XCAT phantom. Med Phys 35:3800–3808

P34 Four-dimensional deforming dosimetry with DEFGEL phantoms J. R. Supple1, P. Lonski2, M. L. Taylor1, T. Kron1,2, R. D. Franich1 1

School of Applied Sciences and Health Innovations Research Institute, RMIT University, Melbourne, VIC, Australia. ([email protected]), ([email protected]), ([email protected]). 2Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne, VIC, Australia. ([email protected]), ([email protected] [Presenting author]) Introduction Quantifying the accuracy of deformable image registration (DIR) based dose-warping and deformable dose accumulation is an inherently difficult task as there is generally no way of knowing the true underlying deformation. As the accuracy of these calculations is largely dependent on the accuracy of the DIR used, much work has been done in assessing the spatial accuracy of DIR algorithms (Latifi, 2013; Yeo, 2013). A novel approach to the problem is to directly compare the results of dose-warping calculations to measurements taken with a deformable gel dosimeter (DEFGEL). Measuring dose to a dynamically deforming integrated phantom/dosimeter presents several challenges. It is critical to have reproducible geometry in regards to the dosimeter, patient/phantom, and source. We have previously shown that DEFGEL is able to be reproducibly position within the phantom and that the deformation applied is reproducible over many cycles (Franich, 2014). Method A Quasar respiratory motion phantom was used to dynamically deform a DEFGEL dosimeter by driving a piston applying force to the dosimeter in the superior-inferior direction. The phantom was CT scanned in 3D for the control case (no deformation) and respiratory correlated 4D for dynamic cases. A simple cross-shaped field applied laterally (perpendicular to the direction of the applied

Fig. 1 Control dose distribution measured in the absence of deformation (a), and the dose distribution measured when the same fields were delivered to a dynamically deforming DEFGEL (b) deformation force) was delivered to each case. Doses were calculated using the anisotropic analytical algorithm in Eclipse. Dose-warping was performed using the open source DIRART software package. Results Figure 1a shows three orthogonal planes through the control dose distribution where no deformation was applied. Figure 1b shows the dose distribution for the dynamically deforming DEFGEL. The coronal and sagittal planes are through the centre of the field and the transverse planes at Dmax. Conclusion Using DEFGEL dosimeters in conjunction with a modified Quasar respiratory motion phantom the, first high resolution 4D measurements of dose were taken. References 1. Franich RD, Supple JR, Lindsay B, Yeo UJ, Lonski P, Smith RL, Taylor ML, Dunn L, Kron T (2014) Reproducibility assessment of dynamically deforming DEFGEL in a respiratory motion phantom. Journal of Physics: Conference Series 573: 012024 2. Latifi K, Zhang G, Stawicki M, van Elmpt W, Dekker A, Forster K (2013) Validation of three deformable image registration algorithms for the thorax. J App Clin Med Phys 14: 19–30 3. Yeo UJ, Supple JR, Taylor ML, Smith RL, Kron T, Franich RD (2013) Performance of 12 DIR algorithms in low-contrast regions for mass and density conserving deformation. Med Phys 40: 101701

P34A A study of dose calculation algorithms using an IPSM phantom with different density materials for in-field and out-of-field conditions Prabhakar Ramachandran1,2, Abdulrahman Tajaldeen1,2, Karl Roozen1, Derrick Wanigaratne1, David Taylor1, Tomas Kron1 1

Physical Sciences, Peter MacCallum Cancer Centre, Vic, AU. ([email protected]), ([email protected]), ([email protected]), ([email protected]), ([email protected] [Presenting author]). 2RMIT University, Vic, AU. Introduction Prior to clinical use, the accuracy of dose calculation algorithm needs to be validated against measurement including for different inhomogeneous conditions. In this study, we have compared four different dose calculation algorithms: Pencil Beam Convolution

123

Australas Phys Eng Sci Med

P35 Is Individual patient QA of 4DCTs used for SABR planning warranted? Rachitha Antony, Elena Ungureanu, Peta Lonski, Tomas Kron Peter MacCallum Cancer Centre Melbourne Victoria. ([email protected]), ([email protected]), ([email protected]), ([email protected] [Presenting author]) Introduction 4DCT is an important tool in radiotherapy planning of moving structures to inform margins for the target volumes that vary depending on the patient’s breathing pattern. In our institution 4DCT is acquired for all stereotactic ablative body radiotherapy (SABR) patients having lung, kidney or liver lesions. We report our experience of the impact of medical physics 4DCT reviews prior to planning based on data from 51 consecutive patients. Method Patients were immobilized using the Elekta Body fix system and 4DCT images acquired with 140 kVp in helical mode on a Phillips Brilliance wide bore CT scanner. Breathing tracks were recorded using the Philips bellows system. Depending on the breathing frequency, the pitch factor is adjusted by the operator for each patient. The images were reconstructed for 10 equidistant breathing phases and a full review of the scan was done by a radiation oncology medical physicist. Results 4DCT scans of 51 patients with different treatment sites and breathing patterns were reviewed. Their impact on treatment volumes with different breathing patterns was assessed and advice on margin and contouring provided where deemed appropriate. In several cases a rescan was recommended by the physicist due to excessive artefacts or lack of recorded tumour motion. The results of the reviews are shown in the figure. It can be seen that irregular breathing predicts for more interventions.

123

Breathing pattern VS impact 30 25 NO.OF CASES

(PBC), AcurosXB, Anisotropic Analytical Algorithm (AAA), and Collapsed Cone Convolution Algorithm (CCC), then validated the results against measurement using an IPSM phantom. Methods The study was conducted on a Varian 21iX linear accelerator for both 6 and 18 MV X-rays. The measurement was conducted using a 0.6 cc ionization Farmer type chamber in a geometric solid water phantom with inhomogeneity inserts (IPSM phantom). The chamber was placed at six different positions and three different field sizes 3 9 5, 5 9 5 and 10 9 10 were defined for both in-field and out-of-field measurements. The measurements were performed for open, physical wedge and enhanced dynamic wedge fields at 90 and 270 degree gantry angles. Water equivalent, lung, ribs and dense bone inserts were used to study the accuracy of dose calculation algorithms for the above geometries. Dose calculation was performed using PBC, AcurosXB, AAA (all Varian Eclipse) and CCC (Mobius) compared against the measured dose. Results The mean difference between measurement and the dose calculations algorithms with water equivalent insert for PBC, AcurosXB, AAA and CCC were -1.57 ± 0.81, -0.96 ± 0.65, -1.50 ± 0.61 and -0.77 ± 0.69 respectively. Similarly, the mean difference beyond lung insert for PBC, AcurosXB, AAA and CCC were 0.68 ± 1.21, -0.40 ± 1.20, -1.8 ± 0.82 and -0.1 ± 1.03 respectively. Conclusion Our results show that CCC and AcurosXB algorithm are closer to the measurements compared to AAA and PBC for majority of the field conditions for water equivalent, lung and ribs inserts. CCC resulted in better agreement with measurement for out-of-field points as compared to all other algorithms

20 15 10 5 0

regular 24

adequate 7

irregular 2

advice on margin

1

5

6

rescan

2

0

4

no issue

Graph showing impact of different breathing patterns on margin

Conclusion This study confirms that physics 4DCT reviews have an impact on SABR cases where dose is delivered with tight margins. It improves the accuracy of target delineation and reduces the uncertainties in treatment delivery.

P36 Investigation of interlocks over three years of service for 15 Varian 21type linear accelerators Li Zhu, Gerrard Janson, Brad Shilling, Tomas Kron Peter MacCallum Cancer Centre, Melbourne, Australia. ([email protected] [Presenting author]) Background and Purpose Radiation oncology relies on linear accelerators (linacs) for the majority of all cancer treatment. Due to the complexity of modern linacs and their constant use things can go wrong and a comprehensive set of interlocks prevents use of machines if safe patient treatment cannot be assured. It was the aim of the present work to review the frequency and type of interlocks observed on 15 linacs in a single institution. Methods All interlocks that require engineering or medical physics intervention are recorded at Peter MacCallum Cancer Centre using an in-house developed software tool, Track to Treat (T2T). We studied the frequency of interlocks occurring over 3 years on 15 Varian Clinac 21 type machines (two 21ex, eleven 21iX, 2100CD, Trilogy) over 5 campuses. All linacs are serviced in house. Interlock occurrence was linked to machine age and use as identified by the presence of an on-board image (OBI). Results From Nov 2012 until Nov 2015 more than 7000 interlocks were recorded with 85% of them specified in the database. The most common interlock was HWFA (Hardware, n = 1152 or 16%) followed by dose interlocks XDRS and DS12 (11 and 10%, respectively). There was a significant difference of recorded interlocks between machines ranging from 70 to nearly 300 per year with a tendency for interlock frequency to increase with age of the machine as can be seen in Fig. 1. Nine machines have on board imaging and cone beam CT. These linacs showed typically more interlocks, be it for the added complexity or the larger number of IMRT/IGRT patients treated (173 vs 139 per machine per year, respectively). Conclusion Interlocks on linacs report important functionality, safety and reliability issues. Analysing their frequency and type can assist with linac maintenance and quality assurance.

Australas Phys Eng Sci Med

P38 Employing statistical process control charts as evidence for improving linac isocentre quality control and frequency optimisation

1000 T2T all

900 Number of interlocks in 3 years

T2T OBI

800

Linear (T2T OBI)

700

Tom Kupfer, Benjamin Harris, Kym Rykers

600 500

Radiation Oncology Centre, Olivia Newton-John Cancer Wellness & Research Centre, Austin Health, Heidelberg, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected])

400 300 200 100 0 0.0

2.0

4.0

6.0 8.0 10.0 Age of machine (years)

12.0

14.0

16.0

Fig. 1 Interlocks recorded over three years for 15 linear accelerators using in-house developed software ‘Track to Treat’ (T2T) as a function of age at the end of the study. A linear fit for the data for the 9 linacs with OBI is shown (R2 = 0.35)

P37 Moving a large and complex radiotherapy department: The role of an independent plan check E. Ungureanu, A. Mullen, E. Kyriakou, F. Gibbons, C. Fox, R. Prabhakar, T. Kron, A. Haworth Department of Physical Sciences, Peter MacCallum Cancer Centre, East Melbourne. ([email protected]), ([email protected] [Presenting author]) Introduction On June 23, 2016, Peter MacCallum Cancer Centre in East Melbourne will turn off lights, close doors and move to a new site in Melbourne. The new centre opens on 27 June with four new (Varian Truebeam), two relocated linear accelerators (21iX, Truebeam STX), and a full suite of brachytherapy and special services. Seventy patients are part way through treatment at transition. We describe the QA measures put in place to ensure the safe transition and will discuss the success/failure of the processes put in place. Method All machines and planning were commissioned following national and international guidelines and independently checked by the Australian Clinical Dosimetry Service. Of the transfer patients, sixty had treatment plans created using Varian Eclipse (mostly IMRT) and ten ELEKTA Xio. About the same number of new patients is scheduled to start treatment in the first two weeks of operation. All plans are checked using Mobius M3D plan check software and verified by analysing dyna log files of deliveries using Mobius FX. Additional patient specific QA measurements were carried out for all transferred IMRT patients using physical measurements (ArcCheck diode array). Results To date, full or partial patient specific QA results are available for 31 of the IMRT patients. They met acceptance criteria of 93% gamma pass rate on Mobius 3D with a 3%/3 mm tolerance and 10% threshold for plan review. Log file analysis and physical measurements are ongoing. The gamma criterion was based on analysis of patient treatments in our centre. Conclusion Mobius provides an efficient process for plan review (M3D) and verification of treatment delivery (MFX) that is based on independent beam data and incorporates patient specific heterogeneities. The ArcCheck physical measurements provide an additional layer of confidence in the actual deliverability.

Introduction Process control charts are quality assurance (QA) tools that provide succinct graphical representations of process stability (1–4). One such chart was implemented for QA of an Elekta Infinity linac, to characterise isocentric stability and inform optimal QA frequency. Method An image-based, semi-automated and robust Winston-Lutztype QA process was established that measures to sub-millimetre accuracy (a) the positioning accuracy of the Elekta Precise Table, (b) the coincidence of radiation isocentre and imaging isocentre, and (c) the size of the radiation isocentre with gantry rotation, defined as the greatest distance between individual radiation field centres along the major X-, Y- and Z-axes (5). A macro-enabled in-house MSExcel spreadsheet facilitated graphical presentation and statistical analysis of the QA results. Action thresholds were based on published recommendations (6, 7). Results From data gathered over 12 months, the table positioning accuracy along X- and Y-axis was 0.0 ± 0.5 mm (mean ± 2SD) and -0.2 ± 1.0 mm in Z-axis. For the latter, the data was not normally distributed. The coincidence of imaging and radiation isocentre along any axis was stable at \|0.2| ± 0.2 mm. The size of the radiation isocentre in the Y-axis was stable at 0.86 ± 0.11. However, on two occasions after radiation beam steering, the mean size changed by more than 0.35 mm ([3 SD) in the X-axis and Z-axis. This magnitude was obvious on the charts, in one instance prompting corrective action. Conclusion Control charts revealed how beam tuning affects the size of the radiation isocenter and elegantly captured the linac’s isocentric stability. These results provided justification for reducing the frequency of comprehensive linac isocentre checks, as well as evidence for when corrective action was required. Similar benefits are anticipated as control charts are applied to other linac QA processes. References 1. Benneyan JC, Lloyd RC, Plsek PE. Statistical process control as a tool for research and healthcare improvement. Qual Saf Health Care. 2003;12(6):458–64. 2. Sanghangthum T, Suriyapee S, Srisatit S, Pawlicki T. Retrospective analysis of linear accelerator output constancy checks using process control techniques. J Appl Clin Med Phys. 2013;14(1):147–60. 3. Rah JE, Shin D, Oh DH, Kim TH, Kim GY. Feasibility study of using statistical process control to customized quality assurance in proton therapy. Med Phys. 2014;41(9). 4. Larcos G, Collins LT, Georgiou A, Westbrook JI. Nuclear medicine incident reporting in Australia: Control charts and notification rates inform quality improvement. Intern Med J. 2015;45(6):609–17. 5. International Electrotechnical Commission. Medical electrical equipment Part 2.1: Particular requirements for safety – Electron accelerators in the range 1 MeV to 50 MeV. IEC 60601-2-11998.

123

Australas Phys Eng Sci Med 6. Klein EE, Hanley J, Bayouth J, Yin FF, Simon W, Dresser S, et al. Task group 142 report: Quality assurance of medical accelerators. Med Phys. 2009;36(9):4197–212. 7. Solberg TD, Balter JM, Benedict SH, Fraass BA, Kavanagh B, Miyamoto C, et al. Quality and safety considerations in stereotactic radiosurgery and stereotactic body radiation therapy: Executive summary. Pract Radiat Oncol. 2012;2(1):2–9.

P39 Automated treatment planning system quality assurance J. J. Lee, B. Cooper Medical Physics and Radiation Engineering, Canberra Hospital. ([email protected] [Presenting author]), ([email protected]) Introduction Ongoing verification that the treatment planning system (TPS) is operating reliably is an important part of overall treatment planning quality assurance (QA). However, there must be a balance between overly rigorous testing, which can lead to an impractical drain on resources, and being too capricious, which could result in missing potential problems. We present here an efficient tool which can assist in routinely validating functionality and consistency of the TPS over time.

Method The main components of the automated TPS QA system are shown below (Fig. 1). The script automatically generates reference trials in Pinnacle (Philips Healthcare, The Netherlands) based on three modalities: a hinged wedge pair, a 5-field IMRT and a single electron beam. The script also creates a backup of the machine model database and calculates the checksum to verify that the files have not been changed. A separate Python script is used to process all output files and compare these to a designated reference QA session. It then populates a time-stamped HTML page and calculates deviations for any numerical data. Results An initial test has shown good short term agreement for the script applied to two scans of the same phantom. These scans were several years apart. However, in order to test the system holistically, this QA programme can only be effectively tested for the same TPS over a significant time period. Conclusion An automated TPS QA system has been developed which provides a convenient way to establish that the TPS is functional and behaving consistently. This should not be the sole component in the TPS QA programme but should be complemented by other tests mentioned above as well as manufacturer release notes and professional judgement.

P40 Impact of the new eye dose limit on the personal shielding (lead glasses) requirements M. McManus, A. Harvey, B. Khoo Medical Technology and Physics, Sir Charles Gairdner Hospital, WA. ([email protected] [Presenting author]), ([email protected]), ([email protected])

CT Scan Machine DB Files

Script Wedge Plan

IMRT Plan

Electron Plan

Vol Report

Vol Report

Vol Report

RT Plan

RT Plan

RT Plan

Screenshot

Screenshot

Screenshot

Checksum

Checksum Report

QA Folder

Transfer to Physics Drive

Transfer to Physics Drive

Ref Folder

QA Folder

Python Script

Introduction In 2011 the International Commission on Radiological Protection (ICRP) recommended that the dose limit for the eye be reduced from 150 to 20 mSv per year (averaged over defined periods of five years, with no annual dose in a single year exceeding 50 mSv). While this dose limit has not yet been incorporated into the WA legislation, Sir Charles Gairdner Hospital (SCGH) recently decided to set an internal eye limit of 20 mSv per year in the spirit of ALARA. Previously no workers were thought to be at risk of exceeding the eye dose limit. However, because the dose limit reduction is so significant previous, expectations had to be reviewed. The eye doses in a range of departments were investigated and shielding requirements were made based on exposure. Method A number of at-risk departments were identified at SCGH. To determine whether the eye doses of staff were close to 20 mSv limit the standard work practices were investigated and collar TLDs were assessed. Trigger level collar doses were established and shielding requirements made for these. Based on the measured doses, recommendations and requirements have been put in place regarding the use of lead glasses. (state these requirements in results) Results There was a range of measured eye doses that could be categorised into worker type and department. The range of doses are likely to be the result of a varied number of procedures, procedure lengths, different factors used and the spread of workload in different areas. Only a small group of staff are at risk of reaching the much reduced eye dose limit, so in most cases no changes were required. Conclusion There has been a surprisingly small need to increase the shielding to the eyes of SCGH staff who work in fluoroscopic areas. Some specific groups have been identified as requiring lead glasses.

TPS QA HTML Report

Reference Fig. 1 Flowchart diagram of automated TPS QA

123

1. ICRP (April 2011). ICRP statement on tissue reactions.

Australas Phys Eng Sci Med

P41 Investigation of systematic discrepancy in TBI planning vs. measurement A. E. Moggre´, A. T. Cousins Medical Physics and Bioengineering Department, Canterbury District Health Board, NZ. ([email protected] [Presenting author]), ([email protected]) Introduction Investigation of a systematic discrepancy between the total body irradiation (TBI) planning process and measurements is presented. At Christchurch Hospital these treatments are manually planned using patient separation measurements, and off-axis and depth-dose tables. The number of monitor units (MU) required to deliver the prescribed dose to midline is validated by in-vivo measurement of dose for each fraction. Pairs of surface diodes record entrance and exit doses, whilst an ionisation chamber (IC) located close to the midline and between the thighs is used to measure the total dose for fractions 1–3 and hence inform any required adjustments of the MU for fractions 4–6 to achieve the prescribed dose. Method and Results Comparison of plans and dosimetric results from the 14 patients treated since 2012 showed that the original MU calculated for treatment was consistently underestimated by an average of 2.8% (2.2% standard deviation, 6.9% maximum deviation). The chamber position compared to the calculation point resulted in slight variations, but the larger impact was found to be due to the central position of the chamber resulting in ‘‘shadowing’’ by the femur, which causes additional dose attenuation that is not accounted for in the homogeneous water equivalent model used for planning. A patient was modelled in the XiO treatment planning system, which showed an average decrease in dose at the chamber location of 2.3%. Subsequent increase of the MU for the second half of the treatment will therefore consistently result in higher doses than calculated throughout the rest of the body. Conclusion There are currently no clinical indications for a variation in our delivered TBI doses, therefore changes to the TBI process are treated cautiously. Several strategies under consideration to account for the underestimated planning doses will be presented, alongside results from an ongoing study to correlate the IC doses with the corresponding diode entrance/exit doses which may eliminate the need for the IC entirely.

P42 Will the real isocentre please stand up: Comparing isocentre locations from different collimating apertures for stereotactic radiation therapy Z. Moutrie1, T. Chen2, N. Hu2 1

Genesis Cancer Care, The Mater Crows Nest, Crows Nest, NSW. ([email protected] [Presenting author]). 2The Mater Crows Nest, Crows Nest, NSW. Introduction The treatment of small volumes of tissue has often been associated with an escalation in the delivered dose and referred to as stereotactic radiation therapy. The use of medical accelerators to provide these treatments have been well documented[1–4], where a variety of collimation systems are available to provide a highly collimated beam. The coincidence of the image-guidance radiation source and the treatment source must be well understood [5] and any deviation between the two systems minimised. Linear accelerator manufacturers provide tools to quantify these deviations.

Method EPID images produced using Cone and MLC collimated fields were used to determine each MV isocentre using commercial software. Non-coplanar images were acquired to and vendor software used to determine the kV-imaging isocentre location, the deviation from the non-coplanar kV imaging system and MV treatment system was compared. EPID images from MLC and jaw collimated fields were used to determine the location of the MV isocentre using both vendor and commercial software. The deviation of this location from the kV Cone Beam CT imaging system was compared. (reword this sentence or sentences to make sense). Results Small variations in the isocentre location of the collimating apertures observed may be attributed to the field edge search algorithm expecting a well-defined sharp edge. The conical collimation and jaw systems meet only one of those search criteria. No observed variation was outside the measurement uncertainty. Conclusion For well calibrated linear accelerator collimation systems no variation is to be expected between conical collimation, MLC or jaw collimated MV treatment isocentres and their respective kV imaging isocentre. A consistent result is expected in comparing commercial or vendor supplied kV-MV isocentre location verification. References 1. Suh JH, Barnett GH, Sohn JW, Kupelian PA, Cohen BH (2000) Results of linear accelerator-based stereotactic radiosurgery for recurrent and newly diagnosed acoustic neuromas. Int J Cancer 90 (3):145–151 2. Verhaegen F, Das IJ, Palmans H (1998) Monte Carlo dosimetry study of a 6 MV stereotactic radiosurgery unit. Phys Med Biol 43 (10):2755 3. Leavitt DD, Gibbs FA, Heilbrun MP, Moeller JH, Takach GA (1991) Dynamic field shaping to optimize stereotactic radiosurgery. International Journal of Radiation Oncology* Biology* Physics 21 (5):1247–1255 4. Winston KR, Lutz W (1988) Linear accelerator as a neurosurgical tool for stereotactic radiosurgery. Neurosurgery 22 (3):454–464 5. Lutz W, Winston KR, Maleki N (1988) A system for stereotactic radiosurgery with a linear accelerator. International Journal of Radiation Oncology* Biology* Physics 14 (2):373–381

P43 Reduction of planning margins for lung stereotactic body radiation therapy (SBRT) R. Newstead1, R. Short2, J. Dowling3, M. Boxer4, M. Yap4, K. Neville4, J. Veneran4, L. Holloway5 1

University of Sydney, Australia. ([email protected] [Presenting author]). 2Macarthur Cancer Therapy Centre, Campbelltown Hospital. ([email protected]). 3 CSIRO. ([email protected]). 4Liverpool and Macarthur Cancer Therapy Centres. ([email protected]), ([email protected]), ([email protected]), ([email protected]). 5Ingham Institute of Applied Medical Research, Liverpool Hospital. ([email protected]) Introduction SBRT is now a common treatment for early stage NSCLC however toxicities have led to the avoidance of treatment within 2 cm of the proximal bronchial tree [1]. This distance can be impacted by inter-observer contouring variation [2]. With the

123

Australas Phys Eng Sci Med Australian MRI-linac program comes real-time imaging and the possibility to reduce planning margins around tumours and increase eligibility for SBRT. This project aimed to demonstrate that marginless treatment plans can meet standard SBRT protocol requirements [3] and assess the accuracy of auto-contouring for treatment planning purposes. Method Two plans were created for five T1/T2, N0 NSCLC patients with GTVexhale \ 5 cm that were previously ineligible for SBRT due to proximity to various OARs. Plan 1 treated the ITV plus a 5 mm PTV margin and plan 2 treated the GTVexhale with no planning margin. Target doses and normal tissue tolerances for the two plans were analysed. Auto-segmentation of the bronchus was performed and compared to existing manual contours. The distance between the ITV and the proximal bronchial tree was measured for both contours. Results All plans met the requirements in the lung SBRT protocol. The no-margin plans were able to meet the target dose for the GTVexhale whilst resulting in lower doses to the OARs than the PTVmargin plans. The distance between the ITV and the bronchus when using the manual contours ranged from 18.9–67.8 mm compared to a range of 17.9–66.9 mm using the auto-contours. Conclusion The no-margin plans may be considered superior as removing margins means a smaller volume is treated. Similar plans may be suitable for future use with the MR-linac and increase eligibility for SBRT. Clinically a CTV margin may be required but the results for the no-margin situation should translate. Although bronchus auto-segmentation resulted in different contours, there could still be value in auto-contouring for removing inter-observer variation.

results, we developed a 3D small animal phantom to be used to evaluate micro-CT dose. Method We evaluated the photon attenuation of materials used in 3D printing by irradiating them in a Micro-CT . Micro-CT system have an X-ray tube that was 70 lm focal spot, the number of detector pixel was 2240 9 2344, and a pixel pitch was 50 lm. We irradiated bolus and 3D printer materials used to make body of 3D phantom for the range of X-ray energies 30–130 kVp and certified photon attenuation by energy level. And commercial numerical mouse phantom was used to develop 3D small animal phantom. This numerical phantom was converted to DICOM image and the whole body was segmented into bone, tissue, and lung using Mimics software. Each segmented image was reconstructed into a 3D object. After evaluating 3D printer materials and modelling 3D animal phantom, we printed small animal phantom. Results We acquired each of the 30 X-ray images of super-flex bolus and rubber-like. These images were averaged to reduce noise, and the averaged intensities at the central area of 200 pixels 9 200 pixels were extracted to avoid caused scattering at the boundary. With this method, three kinds of intensities, super-flex bolus, rubber-like, and air which is not attenuated were acquired for the range of X-ray energy up to 130 kVp. Conclusion We certified the possibility of using 3D printer materials as 3D phantom materials and developed 3D small animal phantom. Using this phantom, we will evaluate absorbed dose in micro CT Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (No. 2014R1A1A2057004)

References 1. Timmerman, R., McGarry, R., Yiannoutsos, C., Papiez, L., Tudor, K., DeLuca, J., Ewing, M., Abdulrahman, R., DesRosiers, C., Williams, M. and Fletcher, J. (2006). Excessive Toxicity When Treating Central Tumors in a Phase II Study of Stereotactic Body Radiation Therapy for Medically Inoperable Early-Stage Lung Cancer. Journal of Clinical Oncology, 24(30), pp. 4833–4839. 2. Jameson, M., Holloway, L., Vial, P., Vinod, S. and Metcalfe, P. (2010). A review of methods of analysis in contouring studies for radiation oncology. Journal of Medical Imaging and Radiation Oncology, 54(5), pp. 401–410. 3. RT 4.188.1 Lung SABR VMAT Planning Protocol. Liverpool & Macarthur Cancer Therapy Centres, SWSLHD Cancer Services. Issue Date: 27/5/15 using Zhang et al. (2011) Radiation Oncology, 6:152 http://www.ro-journal.com/content/6/1/152

P44 Preliminary study for small animal dosimetry in Micro-CT imaging using 3D small animal phantom Su Chul Han, Se-Ho Lee, Seungwoo Park Division of Medical Radiation Equipment, Korea Institute of Radiological and Medical Sciences, Seoul, Korea. ([email protected] [Presenting author]) Introduction Micro–CT is a powerful tool used to study various models of disease on anesthetized animals . In procedures, we irradiated using Micro-CT to the small animal. The high level doses involved in micro-CT have an effect on results of radiobiological studies. In this study, we evaluated dosimetric characteristics of materials of 3D printer as preliminary study and based on these

123

P45 Bioluminescent evaluation of breast cancer cells in vitro for optimal plan parameters in radiation treatment H. Park1, J. Park2, J. Lee3, D. Lee4, K. Choi5, Y. Oh1, M. Chun1, O. Noh1, O. Cho1, D. Woo6 1

Department of Radiatioin Oncology, Ajou University School of Medicine, South Korea. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Department of Radiation Oncology, University of Florida, USA. ([email protected]). 3Department of Radiation Oncology, Konkuk University Medical Center, South Korea. ([email protected]). 4Department of Radiology, Johns Hopkins University School of Medicine, USA. ([email protected]). 5 Department Radiation Oncology, Anyang SAM Hospital, South Korea. ([email protected]). 6MR Core Lab, ASAN Medical Center, South Korea. ([email protected]) Introduction To evaluate and to find optimal plan parameters of field margins and dose rates on metastatic breast cancer cells, bioluminescent-image (BLI) based cell responses were monitored in radiation treatment set-up. Method The luciferase-positive 4T1 cells were cultured on dishes of 3.5 cm and 6 cm diameter. They were inserted into a dedicated acrylic phantom for accurate and reliable dose delivery of 6 Gy in 3 fractions. When the dish is regarded as a gross tumor, cells were irradiated with fields conforming the dish plus positive and negative margins (0.5 cm and 1 cm) with an assumption that target can be missed in contouring on the images. Because the dose rate effect can be different from the cell’s point of view even under the same

Australas Phys Eng Sci Med

Fig. 1 Bioluminescent responses of 4T1 metastatic breast cancer cells on a dish of 6 cm diameter, when the field with a reduced margin of 1 cm was irradiated to delivery 6 Gy in 3 fractions. On the (a) 7th (b) 27th day from initial monitoring irradiation, bioluminescent responses were evaluated with variable dose rates during one-month monitoring. Dose rates were changed from the maximum (600 MU/min at the source-to-surface distance of 64 cm) to the minimum (100 MU/min at 156 cm) with reducing the ratio of 0.1. Results Cells at the field edge formed distinguishable colonies earlier than unirradiated living cells out of field. The cells on the larger dish, regarded as a relatively large sized tumor, showed earlier survival of residual cells as well as irregular and uneven cell death. The number of colonies and bioluminescent radiance in BLIs were quantified in accordance with relative effects of dose rates (Fig. 1). Conclusion The effect of variable field margins was more critical to cells cultivated on the large dish. They showed earlier responses associated with colonies formed at the field boundaries and irregular cell death. Cell responses were correlated with beam parameters with spatial information by matching the dose distributions with BLIs. References 1. Butterworth KT, McGarry CK, Trainor C et al (2012) Dose, doserate and field size effects on cell survival following exposure to non-uniform radiation fields. Phys Med Biol 57(10):3197–3206. doi:10.1088/0031-9155/57/10/3197. 2. Murphy MJ, Lin PS, Ozhasoglu C (2007) Intra-fraction dose delivery timing during stereotactic radiotherapy can influence the radiobiological effect. Med Phys 34(2):481–484. doi: 10.1118/1.2409750.

P46 Material identification using CT imaging A. Perdomo1, Z. Brady1, J. Crosbie2, R. Franich2, J. Du Plessis2 1

Radiology Department, Alfred Health, Melbourne, Victoria. ([email protected] [Presenting author]), ([email protected]), 2School of Science, RMIT University, Melbourne, Victoria. ([email protected]), ([email protected]), ([email protected]) Introduction The ability to identify a foreign material reproducibly with the use of CT may assist the forensic process. Our objective was to assess a range of elemental rods to investigate the optimal way to analyse each sample. Method The rods were mounted in a Perspex cylinder, and individually scanned in a CT number phantom at 80, 100, 120 and 140 kVp,

as well as using a dual energy (DE) mode. Images were reconstructed using an extended CT number scale (maximum value = 30,710 HU). The minimum/maximum/mean/standard deviation CT numbers were obtained using ImageJ software via three distinct regions of interest (ROI) methods. Results The maximum of the three distinct ROI methods were in agreement but varied from the theoretical CT numbers calculated with the CT number formula. As a single example, the difference between the measured and theoretical value for an 8 mm diameter Titanium rod at 80 kVp is 14,400 HU or 78%. But the three ROI measurements produced results that were within 240 HU or 1.5% of each other. The variation between the measured and theoretical values becomes more marked as the atomic number increases and saturation of the measured values occurs at Z = 30 due to the maximum CT number of the extended CT number scale being reached. Conclusion Although the CT numbers measured in this study did not agree with the theoretical CT numbers, the three distinct ROI methods are in good agreement. Therefore, it is possible to differentiate materials using a number of analytical methods. Our approach could have applications in the forensic and materials identification environment.

P48 An inter-observer delineation comparison of Glandular Breast Tissue visible on imaging (CT and MRI) alone E. M. Pogson1,2,3,4, G. Delaney3,5, V. Ahern6, M. Boxer3, C. Chan7, S. David8, M. Dimigen7, J. A. Harvey9, E.-S. Koh3,4,10,11, K. Lim3, G. Papadatos3, M. L. Yap3,10,11,12,13, V. Batumalai3,4,10, G. Liney3,12, C. Moran14, P. Metcalfe2,15, L. Holloway2,3,5 1

Institute of Medical Physics, School of Physics, The University of Sydney, Australia. 2Centre for Medical Radiation Physics, University of Wollongong, NSW, Australia. 3Liverpool and Macarthur Cancer Therapy Centres, NSW, Australia. ([email protected]), ([email protected]), ([email protected]). 4Ingham Institute for Applied Medical Research, Sydney, NSW, Australia. ([email protected] [Presenting author]). 5SWSCS, University of New South Wales, Australia. ([email protected]). 6Crown Princess Mary Cancer Care Centre, Westmead Hospital, NSW, Australia. ([email protected]). 7Department of Radiology, Liverpool Hospital, NSW, Australia. ([email protected]), ([email protected]). 8 Peter MacCallum Cancer Institute, Melbourne, VIC, Australia. ([email protected]). 9Princess Alexandra Hospital, QLD, Australia. ([email protected]). 10University of New South Wales, Sydney, NSW, Australia. ([email protected]). 11Collaboration for Cancer Outcomes Research and Evaluation (CCORE), Liverpool Hospital, Liverpool, NSW, Australia. ([email protected]). 12Ingham Institute for Applied Medical Research, Liverpool Hospital, NSW, Australia. ([email protected]). 13University of Western Sydney, Sydney, NSW, Australia. ([email protected]). 14 Department of Radiology, Stanford University, CA, America. ([email protected]). 15Liverpool & Macarthur Cancer Therapy Centres & Ingham Institute, Liverpool, Australia. ([email protected])

123

Australas Phys Eng Sci Med Table 1 The GBT volume, overlap statistics and distance measure compared pairwise. The uncertainties given are the 95% confidence limits CT Supine Volume (cm3) DSCpairwise HDpairwise (cm)

MRI Supine

CT Prone

MRI Prone

244 ± 18

261 ± 21

273 ± 20

298 ± 24

0.715 ± 0.011

0.682 ± 0.011

0.717 ± 0.010

0.688 ± 0.012

2.76 ± 0.10

3.24 ± 0.13

2.90 ± 0.11

3.20 ± 0.12

P49 Assessment of the joint AAPM/ESTRO/ABG Working Group’s user guide for commissioning of the OncentraÒ ACE model based dose calculation algorithm M. A. Powers1,2, M. G. Roche1, L. Hamlett1, L. Fourie1, R. D. White2 1

Townsville Cancer Centre, Townsville, Australia. ([email protected]), ([email protected]), ([email protected] [Presenting author]), ([email protected]). 2Physical Sciences, College of Science and Engineering, James Cook University, Townsville, Australia. ([email protected])

Fig. 1 Average DSCPAIRWISE compared to the average fused DSC of all 10 observers. Conformity between modalities (CT and MRI) after registration was higher 0.75 than the inter-observer variation 0.72 for CT and 0.69 for MRI. The MRI GBT volumes were larger than the CT GBT volumes (p \ 0.001). Despite this indication that MRI may indicate more GBT than CT, the inter-observer conformity was higher on CT (DSCPAIRWISE = 0.72) than MRI (DSCPAIRWISE = 0.69). Prone GBT volumes were larger than supine GBT volumes

Introduction Visualisation of the Glandular Breast Tissue (GBT) is qualitatively superior on Magnetic Resonance Image (MRI) compared with Computed Tomography (CT), however this has not been assessed quantitatively for radiotherapy purposes. Method The GBT of thirty three patients were delineated on MRI and CT, in both prone and supine positions. This was not a conventional whole breast radiotherapy planning volume as typical information such as palpation and surface markers were deliberately not utilised. This ensured any differences in GBT volumes arose from the differing modality (CT or MRI) alone. Delineation performed by ten experts (9 breast radiation oncologists and 1 radiologist) was assessed utilising the average volume, HDPAIRWISE, and DSCPAIRWISE (VOI). Results The inter-observer variation was similar for the CT supine, CT prone, MRI supine and MRI prone datasets, DSCPAIRWISE *0.7 as shown in Table 1. The average results for all 33 patients is shown in Fig. 1 and compared to the average DSC between registered CT to MRI datasets. Conclusion Consistency of contouring for GBT is between that for whole breast and boost/partial breast target volumes. There are differences in volumes and consistency between CT and MRI. This volume requires further investigation. Acknowledgements This work is supported by a grant from Cancer Australia and the National Breast Cancer Foundation

123

Introduction The AAPM Task Group 186 [1] provides general guidance for early adopters of model based dose calculation algorithms (MBDCAs). To facilitate uniformity of clinical practice a joint Working Group (WG) between the AAPM, ESTRO and the ABG was commenced. One of the activities of the WG is to develop a small number of virtual phantoms and corresponding reference dose distributions for use in level 1 and 2 commissioning of high dose rate (HDR) Ir-192 sources. The WG has also published a user guide to aid with the commissioning process. This poster will summarise the authors’ experience with this user guide and the MBDCA commissioning process. Method The MBDCA commissioning process involves downloading test cases and associated reference dose distributions from the IROC Houston webpage and importing them into the local Oncentra Brachytherapy (OCB) treatment planning system. Dose distributions are then calculated locally with the ACE MBDCA and compared with the reference dose distributions. The procedure in the user guide was followed and results of the dose comparisons were appropriately reported. Results The process outlined in the user guide was intuitive and easily applied, not only for an experienced user of the OCB treatment planning system, but also for a beginner. The instructions and data were concise and allowed for informative comparisons with the reference data in the local OCB system. Navigation of the IROC Houston website, along with downloading and importation of the test case data, was straight forward and easily achieved. Conclusions The user guide provided by the AAPM, ESTRO and ABG joint WG is extremely useful for level 1 and 2 commissioning of the OCB ACE model based dose calculations algorithm. It will undoubtedly be an invaluable tool for providing a standardised level 1 and 2 commissioning procedure for MBDCAs worldwide. Acknowledgements We wish to extend our gratitude to the collaborators of the joint AAPM/ESTRO/ABG Working Group for their work in making this resource available. We would also like to extend our thanks to A/Prof Annette Haworth of the Peter MacCallum Cancer Centre for her efforts in assisting with this project. Reference 1. Beaulieu L, Carlsson Tedgren A, Carrier JF, Davis SD, Mourtada F, Rivard MJ et al. (2012) Report of the Task Group 186 on model-based dose calculation methods in brachytherapy beyond the TG-43 formalism: Current status and recommendations for clinical implementation. Med. Phys. 39(10):6208–6236.

Australas Phys Eng Sci Med

P50 Markerless tumour tracking during lung cancer radiotherapy using on-board kV imaging Chun-Chien Shieh1, Vincent Caillet2, Paul Keall1, Michelle Dunbar1, Chen-Yu Huang1, Jeremy Booth2, Ilana Feain1 1

Radiation Physics Laboratory, Sydney Medical School, University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]), ([email protected]), ([email protected]). 2Department of Medical Physics, Royal North Shore Hospital, Australia. [email protected], ([email protected])

Introduction Lung tumours exhibit large and unpredictable motions, which if unaccounted for can severely compromise radiotherapy outcomes. Markerless tumour tracking can enable wide access to motion-adaptive radiotherapy as it negates the risk of marker-induced toxicity. A major barrier is the inferior tumour visibility on x-ray images due to overlapping anatomies along the imaging beam. The aim of this study is to develop a markerless tumour tracking method using kV imaging. Method Our markerless tumour tracking method consists of two steps. Firstly, an anatomic model of the patient is built from the pretreatment cone-beam CT (CBCT). This model is used to remove the contribution of overlapping anatomies on kV images. Secondly, tumour 3D position is estimated by a Kalman filter, which accounts for modelling uncertainties and imaging noise. The proposed method was retrospectively validated on (i) 11 sets of CBCT projections from four patients with central tumours,[1] and (ii) a kV fluoroscopic scan during a stereotactic ablative radiotherapy (SABR) treatment.[2] Tracking errors were estimated using the motions of markers or beacons implanted near the tumours. To avoid bias, markers and beacons were removed from the x-ray images used in markerless tracking. Results The mean 3D tracking errors ranged from 1.8 to 4.1 mm. Compared with the standard of care, i.e. a single tumour position estimated from the pre-treatment CBCT, markerless tumour tracking reduced tumour localization error by 0.9–7.9 mm. For the 11 CBCT cases, tracking was successful at all imaging angles. For the SABR case, tracking was not possible near the LR view due to poor image quality caused by MV scatter and large radiological depth (Fig. 1). Conclusion The feasibility and benefits of markerless tumour tracking over the standard of care was demonstrated in 12 lung cancer cases. The clinical implementation of this method can potentially enable wide access to motion-adaptive lung radiotherapy (Fig. 2).

Fig. 2 A comparison of mean tumour localisation errors in LR, SI, and AP between markerless tumour tracking and the standard of care, i.e. a single tumour position estimation based on the pre-treatment 3D CBCT. The standard deviations of the localisation errors were plotted as error bars Acknowledgements (if space is an issue this can be omitted) We thank Assistant Professor G Hugo from the Virginia Commonwealth University for providing the CBCT datasets. We thank Nick Hardcastle and the Royal North Shore Hospital for providing and assisting with the SABR dataset. The software development for this project was built based on the Reconstruction Toolkit developed by Dr Simon Rit. This project is supported by an NHMRC Australia Fellowship, NHMRC project grant 1034060, and US NCI P01CA116602. References 1. Roman NO, Shepherd W, Mukhopadhyay N et al (2012) Interfractional positional variability of fiducial markers and primary tumours in locally advanced non-small-cell lung cancer during audiovisual biofeedback radiotherapy. Int. J. Radiat. Oncol. 83:1566–72. 2. ClinicalTrials.gov [Internet]. Royal North Shore Hospital (Australia). 2015 September – Identifier NCT02514512, Lung Cancer Radiotherapy Using Realtime Dynamic Multileaf Collimator (MLC) Adaptation And Radiofrequency Tracking (LIGHTSABR).

P51 Validation of 3D printed I-125 ROPES eye plaque model using GafChromicÒ EBT3 films L. Sim Radiation Oncology Mater Centre, Princess Alexandra Hospital, South Brisbane QLD 4101, Australia. ([email protected] [Presenting author])

Fig. 1 The workflow of the proposed markerless tumour tracking method

The purpose of this study was to evaluate if MED610 3D printed material can be used as a surrogate for acrylic in the manufacturing of a replacement insert used in an eye plaque brachytherapy applicator.

123

Australas Phys Eng Sci Med Measurement of the dose distributions from a standard acrylic insert were compared with dose obtained from MED610 3D printed replica using GafChromic EBT3 films. The study used a 15 mm Radiation Oncology Physics and Engineering Services, Australia (ROPES) type eye plaque applicator loaded with I-125 (model 6711) seeds. GafChromic() EBT3 films were placed in a solid water phantom and dose distributions were measured three-dimensionally both along and perpendicular to a loaded ROPES eye plaque’s central axis (CAX). Each measurement was performed with the stainless steel plaque backing attached to the eye plaque, to assess the variability of the dose distributions between the acrylic and MED 610 insert. Absolute point dose using calibrated film and relative depth dose and profiles were also compared to AAPM TG 43 homogeneous calculations for validation against our clinical treatment planning practice. Results of dose along the central axis were compared between acrylic and MED610 insert and the results found agreement within 1.5%. Off-axis profiles were also compared between the acrylic insert and MED610 and were found to agree to within 7%. The measured central axis dose were also validated against the planning dose calculation using EBT3 film measurements along and perpendicular to the ROPES eye plaque’s central axis. Agreement between experimental measurements and dose calculations were observed given the film uncertainty of 7%.

P52 The economics of radiation risk reduction in medical imaging I. R. Smith1, G. P. Barbaro2, J. T. Rivers3 1

Physical Sciences, St Andrew’s War Memorial Hospital, Brisbane, AUS. ([email protected] [Presenting author]). 2Australasian College of Physical Scientists & Engineers in Medicine, Sydney, AUS. ([email protected]). 3Queensland Cardiovascular Group, Brisbane, AUS. ([email protected]) Introduction Radiation risk reduction in medical imaging is driven by the need to optimize the clinical benefits to those undergoing diagnosis or treatment while minimizing the risks to those incidentally exposed. In this study we evaluate the economic impact these efforts might deliver. Method Using procedural radiation data from the 2010 Dose Datamed 2 (DDM2) project (Study on European Population Doses from Medical Exposure) a model was developed to simulate the detrimental impact of medical imaging practice in a theoretical Australian environment. The model accounts for the distribution of patient characteristics to adjust for radiographic factors across a range of imaging procedures and considers life and quality of life lost through induction of fatal and non fatal cancers. Cost-utility analysis, which measures the effectiveness of health care projects in terms of the cost per quality adjusted life-year (QALY), is used to evaluate the impact of reductions in radiation risk in diagnostic imaging. A value of $66,000/QALY has been assumed. Results Assuming dose reduction can be achieved without impacting clinical outcome, the QALY gain in using lower procedural doses does not have to be offset with losses through diminished clinical utility. Estimating Australian procedural doses to be similar to the average DDM2 doses, a 10% reduction in radiation delivered would save *$15 M/year to the economy. A shift to the first quartile dose levels (*50% reduction) would save *$50 M/year. For cardiologists and nurses involved in cardiac procedures the savings are less

123

significant with QALY gains of 0.006 and 0.004 respectively for individuals performing an average annual caseload across a working life. Conclusion This analysis suggests that significant financial gain to the community can be delivered through efforts to reduce the radiation risks associated with medical imaging procedures. As medical physicists and biomedical engineers we can contribute through leadership in these efforts.

P53 Assessment of lung tumour motion comparing 4DCT, 4DCBCT reconstructed based on internal or external surrogates, and motion of implanted beacons during imaging and irradiation E. Steiner1, C. C. Shieh1, V. Caillet2, N. Hardcastle2, P. Keall1, J. Booth2 1

Radiation Physics Laboratory, Sydney Medical School – Central, University of Sydney, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Northern Sydney Cancer Centre, Royal North Shore Hospital, Sydney, Australia. ([email protected]), ([email protected]), ( [email protected]) Introduction Moving lung tumours exceeding the observed motion from planning 4D computed tomography (4DCT) can result in a lack of dose coverage in stereotactic ablative body radiation therapy (SABR). 4D cone-beam CT (4DCBCT) provides the option to verify tumour trajectories before each treatment fraction. Using implanted Calypso beacons in the lung, this work aims to assess whether 4DCBCT gives a better estimate for the motion during a treatment fraction than 4DCT. Method 4DCBCT was reconstructed for 1–2 fractions of 5 patients (three implanted Calypso beacons) receiving lung SABR from the projections acquired for treatment setup CBCT with the clinical standard protocol. Two reconstructions per dataset based on the Calypso motion trajectories or an external Bellows belt surrogate signal were created using the prior image constrained compressed sensing (PICCS) method. Calypso beacons were segmented for all 10 bins of the 4DCT and 4DCBCT sets and the centroid position calculated. Beacon centroid motion with respect to reference phase (endexhale) was extracted and compared with the closest surrogate for the actual tumour motion (Calypso centroid signal) during CBCT acquisition and during irradiation. Results Both methods for 4DCBCT reconstruction failed to capture sudden motion peaks during scanning (see Fig. 1), but their performance was similar to the 4DCT. In general, 4DCT and 4DCBCT underestimated the actual beacon centroid motion (see Table 1) and performed best in providing a valid motion estimate for the SI direction (16–28% of the actual motion exceeded range), while performing rather poorly for the AP and LR direction (up to 61% of the motion exceeded range). Conclusion Both 4DCT and 4DCBCT failed to represent the full tumour motion range. For a safe treatment delivery this needs to be accounted for either by sufficient margins or more preferably realtime treatment adaptation using MLC tracking directly tackling motion peaks and unpredictable motion.

Australas Phys Eng Sci Med

Fig. 1 Motion range of and exemplary patient treatment fraction in (a) AP, (b) LR and (c) SI direction. 4DCT at simulation, 4DCBCT reconstructed based on the Calypso centroid signal, 4DCBCT reconstructed based on the Bellows belt signal, Calypso centroid motion during CBCT acquisition and Calypso centroid motion during irradiation

MRI only treatment planning would remove current registration errors, approximately 2 mm [2, 3], reduce patient discomfort and lower the workload and financial cost of additional scans. MRI however lacks electron density information required for dose calculations. This may be overcome by applying manually segmented bulk density corrections for tissue, bone, lung and air, with results reported to be within approximately 2% compared to the same plan applied on a CT dataset [2, 4]. Further work includes an atlas-based electron density mapping method to automatically segment and apply the appropriate bulk density correction, with some success reported for prostate cases [5] Method 10 retrospective H&N patients had clinically acceptable VMAT plans created. These plans were recalculated on the data sets with no density correction (water equivalent) and with a bulk density correction for bone/air/tissue applied. Plans were also reoptimised on these two data sets, and recalculated. These plans were compared for dose calculation and optimization error through point dose comparison, DVH analysis and gamma analysis of dose. Results Bulk density corrections for bone and air provide dose calculations generally within expected uncertainties (2%) of the original treatment plans with appropriate corrections. Conclusion The bulk density approach to electron density correction achieves uncertainties within expected dose uncertainties, however more advanced approaches such as a direct conversion approach may improve accuracy, particularly at interface regions, and should be considered. References

Table 1 Mean percentage of tumour (Calypso centroid) motion during CBCT imaging and during irradiation outside of the motion range of 4DCT and 4DCBCT Calypso CBCT LR

SI

Calypso irradiation AP

LR

SI

AP

4DCT

62.8 ± 18.8 27.5 ± 17.8 36.9 ± 5.0

40.2 ± 32.8 15.6 ± 5.8 46.0 ± 6.6

4DCBCT Calypso

61.1 ± 25.5 27.3 ± 18.9 38.0 ± 9.4

39.7 ± 15.7 21.7 ± 3.8 44.8 ± 6.0

4DCBCT Belt

47.7 ± 13.4 23.9 ± 15.6 39.1 ± 12.4 35.9 ± 14.3 16.2 ± 4.8 51.2 ± 10.1

P54 Assessment of electron density effects for MRI only treatment planning Tony Young1,2, David Thwaites2, Lois Holloway1,2,3,4 1

Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute, Sydney, Australia. 2Institute of Medical Physics, School of Physics, University of Sydney, Sydney, Australia ([email protected]), ([email protected] [Presenting author]). 3University of New South Wales, Sydney, New South Wales, Australia. 4Centre for Medical Radiation Physics, University of Wollongong, Wollongong, New South Wales, Australia. ([email protected]) Introduction The gold standard for radiotherapy simulation and treatment planning is Computed Tomography (CT) due to spatial accuracy, bony anatomy definition and electron density information for dose calculations. Initial use of MRI in radiotherapy was to improve visualisation of anatomy for accurate target definition and contouring, via fusion of the MRI scan with the CT scan. This introduces registration error and potential errors from variations in patient setup and internal organ motion [1]

1. Nyholm T, Jonsson J. Counterpoint: Opportunities and Challenges of a Magnetic Resonance Imaging–Only Radiotherapy Work Flow. Seminars in Radiation Oncology. 2014;24(3): 175–80. 2. Karlsson M, Karlsson MG, Nyholm T, Amies C, Zackrisson B. Dedicated Magnetic Resonance Imaging in the Radiotherapy Clinic. International Journal of Radiation Oncology Biology Physics. 2009;74(2):644–51. 3. Roberson PL, McLaughlin PW, Narayana V, Troyer S, Hixson GV, Kessler ML. Use and uncertainties of mutual information for computed tomography/magnetic resonance (CT/MR) registration post permanent implant of the prostate. Medical Physics. 2005;32(2):473–82. 4. Lambert J, Greer PB, Menk F, Patterson J, Parker J, Dahl K, et al. MRI-guided prostate radiation therapy planning: Investigation of dosimetric accuracy of MRI-based dose planning. Radiotherapy and Oncology. 2011;98(3):330–4. 5. Dowling JA, Lambert J, Parker J, Salvado O, Fripp J, Capp A, et al. An Atlas-Based Electron Density Mapping Method for Magnetic Resonance Imaging (MRI)-Alone Treatment Planning and Adaptive MRI-Based Prostate Radiation Therapy. International Journal of Radiation Oncology Biology Physics. 2012;83(1):e5–11.

P55 Treatment planning systems cannot be trusted to calculate MU for electron radiotherapy: Fact or Myth? B. Beeksma1,2, P. Vial1,2 1

Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centre, Australia. 2Ingham Institute for Applied Medical Research, Australia. ([email protected]), ([email protected] [Presenting author])

123

Australas Phys Eng Sci Med Table 1 Number of computed output factors showing \3% difference from measured values in flat phantom geometry Beam energy

Cutout shape Square

Rectangular

Irregular

6

20/20

14/15

1/2

8

20/20

15/15

2/2

10 12

20/20 20/20

15/15 15/15

2/2 2/2

15

20/20

15/15

2/2

Total

100/100

74/75

9/10

(100%)

(99%)

(90%)

Introduction There is a widespread view that for electron radiotherapy Treatment Planning System (TPS) calculated monitor units (MU) are less reliable than manually calculated MU based on measured cut-out factors. There is a lack of evidence supporting this view. In this work we attempt to separate myth from fact by an experimental investigation. Method This study is based on the Pinnacle3 (Version 9.8, Philips Healthcare, USA) TPS which uses a modified version of the Hogstrom electron pencil-beam algorithm. A 100SSD electron beam model was completed for all available electron energies on an Elekta Synergy linear accelerator. TPS calculated MU on both a flat phantom and patient CT were compared to measurement on a flat phantom for a large range of standard and patient electron cut-outs. Results The agreement between TPS and measured cut-out factors on a flat phantom is summarized in Table 1. From the 33 clinical cases, 32 (97%) had \3% difference between TPS and measured cut-out factor in the flat phantom geometry. When computed in patient geometry, differences in MU were \3% in 25/33 (76%) of cases and \5% in 30/33 (91%) cases. Conclusion No significant difference was observed between TPS and manually calculated MUs on a flat phantom. Preliminary evidence supports the case that TPS calculated MUs are more reliable than manual calculation when patient anatomy is included.

P56 Commissioning and clinical implementation of EBT3 films for small field in vivo dosimetry of superficial treatments

Fig. 1 Dose response and calibration set up (left). Placement of the film on the plaster mould (right). Field is collimated by applicator cone and lead cut-out dosimeters are not practical in this scenario where profiles and dose maps are preferable over point dose measurement. In this study, we discuss the commissioning and clinical implementation aspects of EBT3 films to assess the dose delivered to patients planned using small kV fields. Method The energy dependence and dose response of EBT3 films were evaluated for kilovoltage X-ray beams. A dose calibration curve was established using the function OD ¼

aD bD

ð1Þ

where D is the absorbed dose in Gy and a and b are parameters. Initial evaluation was carried out by placing the film on a plaster mould (Fig. 1). The dose response curve was established by carrying measurements on a flat surface with full backscatter. Results The difference between the expected and measured dose was within ±5% inside the entire treatment field. Our results also showed that film could be a valuable tool for verifying the dose fall off outside the field which is highly desirable for small fields. Conclusion The dose response curve can be well defined for kV . Considering the difficulty beams using rational function OD ¼ baD D of the applicator alignment on irregular anatomy and size of the field, the dose given can be evaluated within ±5.0% using EBT3 films. Reference 1. D Wanigaratne, T. Kron, S. Herath, S. Atkins and J. Cramb, ‘‘A robust curve fitting for dose calibration in EBT2 Gafchromic film dosimetry’’ (Australian Phys Eng Sci Med (2011) 34:627.

D. Wanigaratne1, R. Prabhakar1, D. Taylor1, K. Roozen1, T. Kron2 1

Peter MacCallum Cancer Centre, Moorabbin Campus Centre Road East Bentleigh Vic. ([email protected] [Presenting author]). 2Peter MacCallum Cancer Centre, VCCC 305 Grattan Street Melbourne VIC 3000 Introduction The Peter MacCallum Cancer Centre in Moorabbin uses kilovoltage photon beams (tube potential 40–225 kV) to treat superficial tumours. In some instances, the field size is relatively small with a diameter less than 10 mm. The beam is generally collimated by cones and lead cut-outs to set the treatment field on patient’s skin. It is important that the prescribed dose is delivered to the tumour, and dose decreases rapidly outside the field to minimize the radiation damage to healthy tissues. Thermoluminescent

123

P57 Calibration free output factors on the Leksell Gamma-Knife with the Exradin W1 PSD L. K. Webb1, E. K. Inness1, C. E. Jones1, P. H. Charles1,2 Radiation Oncology, Princess Alexandra Hospital, Brisbane, Australia. ([email protected] [Presenting author]), ([email protected]), ([email protected]). 2Science & Engineering Faculty, Queensland University of Technology, Brisbane, Australia. ([email protected])

Australas Phys Eng Sci Med Introduction The small field sizes available (4, 8 and 16 mm) for treatment on the Leksell Gamma Knife (LGK) requires the use of specialised detectors for the accurate measurement of output factors (OF). The Exradin W1 PSD fulfils these requirements, having high spatial resolution, angular independence and water equivalence. Traditionally, in order to utilise the PSD for relative dosimetry a Cerenkov Light Ratio (CLR) calibration should be performed on a standard linac to remove the influence of Cerenkov radiation. This work determined if the PSD could be used for relative dosimetry on the LGK without the need for calibration on a linac. Method Profiles in the x, y and z directions for the 4 mm field size were measured to ensure that the effective point of measurement of the PSD was located at the LGK’s radiation iso-centre. Nine irradiations, each 60 seconds in duration, were performed for each of the 3 LGK field sizes. The PSD was set to the raw (no corrections) 2 channel mode on the associated Supermax electrometer, and the charge collected in both channels was recorded for each irradiation. The OF (16 mm reference) for each field size was then calculated with and without the CLR calibration factor. Results Without CLR calibration the OF’s were found to differ from the nominal values supplied by Elekta by 1.3 and 0.5% for the 8 and 4 mm shot sizes, respectively, and with CLR calibration they were found to differ by 0.3 and -1.4% for the 8 and 4 mm shot sizes, respectively. Conclusion Due to the small field sizes, and hence the proportion of optical fibre irradiated, the impact of Cerenkov radiation on the PSD when used on the LGK was minimal. Therefore it can be used for relative dosimetry checks without first having to undertake CLR calibration on a linac.

P58 Developing a dynamic anthropomorphic phantom for SBRT and DIBH QA Chuan-Dong Wen INWENTECH, Melbourne, VIC 3150 Australia. ([email protected] [Presenting author]) Introduction Aiming to assist the increasing clinical implementations of SBRT and DIBH techniques for lung and breast cancers, a deformable dynamic anthropomorphic phantom has been designed for treatment equipment commissioning and patient-specific QA purposes. This paper describes the process of concept development of an inventive (?innovative)medical device. Method A thorough literature review was conducted over both scientific publications and engineering designs for purpose-built radiotherapy phantoms as well as ant available commercial products. Particular focus of this survey was on finding the suitable phantoms that would satisfy the clinical requirements by SBRT and DIBH commissioning and patient-specific QA. A comparison of both intended design features and technically achievable functions of available devices were conducted against criteria that are clinically significant. By natural progression, a novel design concept was developed in the process. Of importance, a working model based on that concept was constructed and tested clinically. Results It is confirmed that though a few products are available, no single device will satisfy all major requirements of SBRT and DIBH, when considering aspects of tumour motion, organ deformity and true representation of clinical complexity. Therefore, an organ deformable dynamic anthropomorphic thorax phantom is needed and has been designed from the clinician’s perspectives. This device can be used as the end-to-end test phantom during equipment commissioning or as a clinical pre-treatment QA device by the planners and therapists.

Conclusion A comprehensive and purposely-built dynamic phantom is required for the complex, ever increasing requirements of motion management in the clinical treatments delivered by SBRT and DIBH techniques. This author has designed a novel device. References 1. Keall et al. (2006) The management of respiratory motion in radiation oncology report of AAPM Task Group 76 Med. Phys. 33 (10), 3874–900 2. Willoughby et al. (2012) Quality assurance for nonradiographic radiotherapy localization and positioning systems: Report of Task Group 147 Med. Phys. 39 (4) 3. P Steidl et al. (2012) A breathing thorax phantom with independently programmable 6D tumour motion for dosimetric measurements in radiation therapy Phys. Med. Biol. 57, 2235–2250 4. Stine S Korreman (2012) Motion in radiotherapy: photon therapy Phys. Med. Biol. 57, R161–R191

P59 HDR afterloader quality assurance using helical diode array R. Wilks, S. Nilsson Royal Brisbane and Women’s Hospital, QLD. ([email protected] [Presenting author]), ([email protected]) Introduction Brachytherapy treatments involve the delivery of high doses with steep dose gradients to the target volume. It is therefore essential that the afterloaded radioactive source be placed accurately and precisely in the predetermined position for the correct length of time. Regular quality assurance of the afterloading unit and its source is consequently required to verify the performance of the delivery system. The accuracy of the source position and step size, and the dwell time accuracy and linearity, are commonly measured by methods susceptible to human error, such as the use of a stopwatch or tracings and measurements on radiographic film. It is therefore valuable to produce a quality assurance system with minimal setup and measurement uncertainties. Method The Sun Nuclear ArcCHECK was employed to collect data from a Nucletron MicroSelectron HDR afterloader system. The Iridium-192 source was inserted into the centre of the bore where relative measurements were made using the SNC Patient software (V6.2). The movie files were then analysed using in-house Python code to realise the dwell time accuracy, reproducibility and linearity, and the source position reproducibility of the afterloading system. Results The dwell time was found to be reproducible to within 3 data points (i.e. 0.15 seconds), but was consistently overestimated by 0.4 seconds. The dwell time was also measured to be linear, within measurement uncertainty. The ArcCHECK measured that the source position, relative to the central detector, was reproducible with standard deviation of 1.2 mm. Reproducibility of the system could be increased using noise reduction techniques on the measured signal. Conclusions The ArcCHECK helical diode array was successfully used to measure the reproducibility and linearity of the afterloading system. Further data processing, such as noise reduction, is required to more precisely and accurately measure the dwell time, and potentially the step size, of the system.

123

Australas Phys Eng Sci Med

P60 Improvement to the junction dose for the craniospinal irradiation technique at AHRO Reza Alinaghizadeh, Christopher Wong, Fiona Mu, Stephanie Miller Alfred Health Radiation Oncology. ([email protected] [Presenting author]) Introduction Cranio Spinal Irradiation (CSI) treatment technique, is a technically challenging treatment because of the large size (of what?), tissue sensitivity and the required use of multiple junctions. This research proposed a method to improve the junction dose for CSI treatment. Method CSI requires irradiation of the brain and spinal cord, which requires several junctioned fields to totally encompass (encompass what?). At Alfred Health Radiation Oncology (AHRO), junctioned IMRT fields are used for CSI treatment. This study examines the improvement of dose uniformity within the junction region by using overlapping IMRT fields instead of moving junction (feathered) technique used currently. Using overlapping fields means the treatment plan is less sensitive to the set-up position uncertainty and results in a more homogeneous dose distribution. Results Based on TPS modelling, the proposed IMRT with junction overlap technique improves the dose uniformity in the target volume

123

and also reduces the sensitivity of the junction dose to misplacement. Utilising overlapping fields in the junction region demonstrates a reduction in the size of inhomogeneities compared to the moving junction method. A ±3 mm and ±1 mm relative misplacement of the junctioned fields results in a change in the delivered dose distribution of up to ±16 and ±5% respectively. By comparison, the current feathered IMRT technique in the same situation would lead to a dose difference of 36 and 16% respectively. Conclusion The improvements to dose homogeneity and the reduced magnitude of potential errors due to set-up uncertainty is a strong incentive to further investigate and hopefully introduce the proposed overlapping junction IMRT technique for CSI treatment. References 1. Seppa¨la¨, J., Kulmala, J., Lindholm, P., and Minn, H., (2010) ‘‘A Method to Improve Target Dose Homogeneity of Craniospinal Irradiation Using Dynamic Split Field IMRT’’, Radiotherapy and Oncology 96: 209–215 2. Wang, Z., Jiang, W., Feng, Y., Guo, Y., Cong, Z., Song, B., and Guo, Y., (2013), ‘‘A simple approach of three-isocenter IMRT planning for craniospinal irradiation’’, Radiation Oncology 8: 217 3. Hadley, A., and Ding, G.X., (2014) ‘‘A single-gradient junction technique to replace multiple-junction shifts for craniospinal irradiation treatment’’, Medical Dosimetry 39: 314–319