Eur Radiol (2008) 18: 1674–1682 DOI 10.1007/s00330-008-0934-9

Sabrina V. Vollmar Willi A. Kalender

Received: 23 July 2007 Revised: 7 January 2008 Accepted: 7 February 2008 Published online: 15 April 2008 # European Society of Radiology 2008

S. V. Vollmar (*) . W. A. Kalender Institute of Medical Physics, Henkestrasse 91, 91052 Erlangen, Germany e-mail: [email protected] Tel.: +49-9131-8522310 Fax: +49-9131-8522824

COMPUTER TOMOG RAPHY

Reduction of dose to the female breast in thoracic CT: a comparison of standard-protocol, bismuth-shielded, partial and tube-current-modulated CT examinations

Abstract We evaluated the potential for reduction of dose to the female breast in computed tomography (CT) of the thorax by using three different techniques: bismuth shielding, partial CT scanning and tube-current modulation (TCM). Measurements and simulations of dose and image quality were performed for a 64-slice CT system using a semi-anthropomorphic thorax phantom with breasts added. Three-dimensional dose distributions were calculated by Monte Carlo (MC) methods. Noise was determined by measurements and simulations. Bismuth shielding resulted in a dose reduction of about 50% for the breast, noise increased up to 40% and image quality was impaired by artifacts. In partial CT scans, not irradiating the breasts directly, dose to the breasts

Introduction The number of diagnosed breast cancer cases has increased steadily for more than two decades and has remained level since 2000 [1]. The augmented awareness for the number one of all cancer diseases in women has led to focused attention on potential predisposing factors such as nutrition, weight, fitness and the radiation due to medical X-ray examinations. When taking a look at the influence of ionizing radiation on breast cancer risk, a number of studies can be found which evaluated the impact on the possibility of malignant tumor development in women exposed to high doses. The incidence of breast cancer in atom bomb survivors [2], in women subjected to multiple fluoroscopic examinations [3] or treated by radiotherapy [4] was found to be increased. Accordingly, a high radiation sensitivity of

was reduced typically by 50%. To sustain a constant noise level, an increase of irradiation in the anteroposterior position resulted in a higher dose to the spine. Reduction of dose to the breasts of about 10% was achieved with TCM; distribution of noise was homogeneous and image quality uniform. Reduction of dose to the female breast was achieved by using all adapted CT methods. Bismuth shielding may compromise image quality, increase noise level and introduce streak artifacts. Partial and TCM examinations reduced dose to the breast without influencing image quality. Keywords Spiral CT . Chest . Breast . Radiation dosimetry

breast tissue was stated. This is also reflected in the new assignments of relative tissue risk factors by the International Commission of Radiation Protection (ICRP): the weight factor for the breast was increased from 0.05 [5] to 0.12 in the newest ICRP report, which is expected to go into effect in 2008 [6]. In diagnostic thoracic computed tomography (CT) the breast is always included in the scan field, but it is rarely or never the organ of interest. Given the radiation sensitivity of the glandular tissue and the fact that the breast is not the object of interest in imaging, any easy to use and effective means for reducing the radiation dose is to be welcomed. In this study we compared three methods of reducing dose to the breast in thoracic CT in regard to their effect on dose, practicability and impact on image quality. One technique with which the radiation dose to the breast can be

1675

reduced is the use of in-plane breast shielding [7, 8]. With in-plane shielding, the X-ray beam is partially blocked to reduce the dose to the underlying tissue while allowing enough X-rays to pass in order to be able to generate a diagnostic CT image. Bismuth garments are commercially available in different sizes for different applications. They are mainly used for breast, thyroid and eye-lens shielding in CT to protect the organs at risk from direct exposure. Another technique is a partial CT scan centered posteriorly. Compared with the other methods this technique is not routinely available on any CT system today, and the idea of this study is to analyze its potential for dose reduction to the female breast and whether it would be useful to have it offered on CT machines for routine clinical practice. In this case the radio-sensitive organ is not exposed directly to the X-ray beam, which is realized by switching off the X-ray tube or at least lowering the tube current to a minimum for a certain range of projections in an anteroposterior direction [9, 10]. A loss of image quality is not to be expected if the scan range per rotation is 180° plus fan angle. In our case we used 180° plus the fan angle of 52°. For the partial CT scan the information for the image reconstruction is complete, due to symmetry, and no artifacts will result. Noise distribution in the z-direction will vary depending on the center of the image. The third technique of reducing the dose not only to the breast but in general is tube-current modulation (TCM). The modulation is chosen dependent on the attenuation for each projection angle, i.e., the tube current is lowered for projections with a shorter path through tissue, e.g., in the anteroposterior direction for the shoulder and thorax, and kept constant for projections with the highest attenuation [11, 12]. Several studies have shown that this method leads to an overall dose reduction without compromising image quality [13–16]. Our goal in this study was to determine the possibility of decreasing radiation exposure of the breast in thoracic CT without decreasing image quality in the organs of interest such as the lung or the heart.

larger patient cross-sections [18]. As different patient sizes had little influence regarding the specific differences between the approaches for breast dose reduction compared here, results presented in this paper are only given for the 25 cm by 35 cm phantom shown in Fig. 1. Inside the phantom, the lungs, the heart and the vertebra are modeled out of tissue-equivalent material. In addition, the phantom provides five bore holes for dose measurements. They are positioned centrally and at the 12, 3, 6, and 9 o’clock positions below the surface in a manner comparable to the phantoms for measuring the CT dose index (CTDI) [12]. For our purpose of evaluating the dose to the female breast, we added material at their positions to the thorax. In the simulations, breasts were modeled by water-equivalent sphere segments and for the measurements we fixed waterfilled bags with a bra to the phantom as shown in Fig. 1. The semi-anthropomorphic physical phantom and its mathematical equivalent were identical in their attenuation properties and shape, only the position and shape of the breasts differed slightly but without a significant influence on the study results. CT measurements Measurements were performed on a standard clinical CT system (Sensation 64, Siemens Medical Solutions, Forchheim, Germany) in spiral mode. Standard protocol The reference CT examination was made at 120 kV with 100 mAs, 200 mA at a rotation time of 0.5 s, and a pitch of 1.2. Slice collimation was 32 × 0.6 mm and the z-flying

Materials and methods Phantoms For this study we used a semi-anthropomorphic thorax phantom (QRM, Moehrendorf, Germany), commonly used in cardiac imaging, in particular for image quality assessments and for standardization efforts [12, 17, 18]. The phantom is 10 cm in the z-direction and has an extension of 20 cm in the anteroposterior and 30 cm in the lateral directions. These dimensions were derived as representative for a collective of healthy volunteers of both sexes and 20- to 80-years old scanned to establish reference values of lung density [19]. Extension rings of 25 cm by 35 cm and 30 cm by 40 cm are available to mimic

Fig. 1 Semi-anthropomorphic thorax phantom with water-filled bags fixed with a bra to simulate breasts and with bore holes for dose measurements (1 vertebra, 2 lung, 3 heart, 4 breasts, 5 bore holes, 5.1 at centre, 5.2 at 12 o’clock, 5.3 at 3 o’clock, 5.4 at 6 o’clock, 5.5 at 9 o’clock)

1676

focus was switched on as in all other scan protocols. The parameters were chosen as standard values of thoracic CT examinations used in our department of radiology.

the tube voltage and current, pitch and slice thickness have to be defined by the user and were modified to match the conditions of the performed CT examinations. We matched the CT protocols for the simulations to the ones presented in the section “CT measurements.”

Bismuth-shielded CT examination For this examination the same protocol as in the reference study was used but with commercially available bismuth garments (AttenuRad Radioprotection, 0.060 mm Pb equivalent, Dyna Medical Corporation, London, Ontario) applied to the breasts to reduce direct irradiation. Partial CT examination The posteriorly centered partial CT scan avoids exposing the breast directly; the X-ray tube is only switched on during 232° out of 360° per rotation. As the X-ray tube cannot currently be switched off on the clinical CT system, we performed this experimental evaluation of image quality by using only 232° of the acquired raw data for the reconstruction. Making this method available on clinical CT systems may be realized by lowering the tube current to a minimum during the posterior part of the rotation. Similar technical solutions are used for ECG pulsing in cardiac CT where the tube current is lowered down to 4%. To achieve the same level of exposure as in the reference study, we increased the intensity of the tube current from 200 mA to 310 mA. This results in the same mAs values [310 mA × 0.5 s × (232°/360°) = 100 mAs] as in the reference study.

Dose simulations Three-dimensional (3D) dose distributions were calculated with a commercially available Monte Carlo tool (ImpactMC, VAMP, Erlangen, Germany), which was validated in a separate study [20]. In the validation of the software tool, a maximum deviation between measurements and calculations was found to be 10%. The calculations were performed on 3D voxel volumes that can either be filled with a measured CT data set, for example a patient CT examination, or with a mathematically described phantom. Additionally, information about CT protocol parameters, such as the tube voltage and current, pitch, slice thickness, and further information about the scanner, such as the CT system’s geometry, the shaped filter and the pre-filtration, have to be provided. According to the given parameters the dose distribution in the 3D voxel volume is calculated applying the Monte Carlo method [21]. Dose was assessed for the breasts as well as for the spine in the region directly exposed and for the CTDI-like positions (Figs. 2a, 3a). For comparison purposes, we determined the mean dose over the whole organ for the breast and the mean dose of the bone marrow for the part of the spine exposed. Mean dose over the whole phantom volume was measured as well.

TCM scan For TCM the CARE Dose 4D option (Siemens Medical Solutions, Erlangen, Germany) on the scanner was used. In a first step, CARE Dose4D automatically adjusts the overall exposure level based on the attenuation obtained by the topogram and therefore depending on patient size. The maximum tube current for each slice is defined. In a second step, the tube current is modulated in real-time depending on the attenuation as a function of the projection angle. The nominal tube current-time product was kept at 100 mAs. CT simulations The simulation study was realized with the software tool ImpactSim (VAMP, Erlangen, Germany). The software calculates attenuation line integrals through the mathematically defined phantoms in an analytical way. Information about the CT system’s geometry, the shaped filter and the pre-filtration has to be available and was entered for the given setup. All variable CT protocol parameters, such as

Dose measurements As a validation of the simulations, we measured dose in the five bore holes of the thorax phantom for the standard protocol with a Unidos Electrometer and a 10-cm CTDI pencil-shaped ionization chamber type 30009 (PTW, Freiburg, Germany) in the center of the phantom and in the four peripheral positions (12, 3, 6, 9 o’clock). The measured dose values were calculated as the weighted CTDI (CTDIw) [12]: CTDIW ¼ 1=3CTDIcenter þ 2=3CTDIperiphery

(1)

Assessment of image quality The quantitative evaluation of image quality was limited to measuring noise levels. Noise is defined as the standard deviation of CT values and was determined in two regions of interest (Figs. 4a, 5a) in the central slice of the scanned

1677

Fig. 2 Dose distributions calculated for simulated scans. a Reference scan with ROIs; b bismuth-shielded CT; c partial CT; d TCM

volume representative of the heart and the lung. For the heart, we chose a central 10-cm diameter region of interest (ROI) and for image quality assessment in the lung region we chose a 12 × 4-cm oval ROI in the right lobe of the lung. Noise results are given as the average noise value over every slice of the scanned volume.

Fig. 3 Dose distributions calculated for measured scans. a Reference scan with ROIs; b bismuth-shielded CT; c partial CT; d TCM

Even though noise cannot be easily measured in clinical cases due to structured tissues and inhomogenities, it is the appropriate parameter to be measured in homogeneous phantoms to evaluate image quality objectively. For a qualitative evaluation of image quality the presence and severity of streak artifacts was recorded.

1678

Fig. 4 Simulated CT images. a Reference scan with ROIs; b bismuth-shielded CT; c partial CT; d TCM

Results Dose Measured and calculated CTDI values for the thorax phantom are shown in Table 1; the deviations of typically Fig. 5 Measured CT images. a Reference scan with ROIs; b bismuth-shielded CT; c partial CT; d TCM

10% or less between measurements and simulations can be considered as good agreement. In consequence, further evaluations of different scan protocols and approaches were restrained to Monte Carlo calculations to limit efforts. All Monte Carlo dose values are given as relative values in mGy per mGy air kerma, which means that they have to be

1679

Table 1 Comparison of measured and calculated CTDI values for the reference CT protocol CT scan

Centre 12 o’clock 3 o’clock 6 o’clock 9 o’clock CTDIw

CTDI Measured (mGy)

Calculated (mGy)

Deviation (%)

4.30 6.04 5.83 6.38 5.62 5.41

4.63 6.64 6.29 6.76 6.19 5.85

7.6 9.9 7.8 6.0 10.1 8.2

Table 2 Comparison of organ dose values for simulated CT examinations CT scan type

Reference Bismuth-shielded Partial TCM

Breast

Spine

Rel. dose (mGy/mGy)

Rel. difference to ref. scan (%)

Rel. dose (mGy/mGy)

Rel. difference to ref. scan (%)

0.661 0.350 0.345 0.598

– –47.0 –47.8 –9.5

0.352 0.331 0.415 0.308

– –6.0 17.9 –12.5

Table 3 Comparison of organ dose values for measured CT CT scan type

Reference Bismuth-shielded Partial TCM

Breast

Spine

Rel. dose (mGy/mGy)

Rel. difference to ref. scan (%)

Rel. dose (mGy/mGy)

Rel. difference to ref. scan (%)

0.626 0.358 0.346 0.567

– –42.8 –44.7 –9.4

0.364 0.353 0.432 0.323

– –3.0 18.7 –11.3

Table 4 Comparison of mean dose for the whole phantom CT scan type

Relative dose values (mGy/mGy)

Absolute dose values (mGy)

Difference to ref. scan (%)

Reference Bismuth-shielded Partial TCM

0.417 0.350 0.388 0.407

5.930 4.977 5.517 5.209

– –16.1 –7.0 –12.2

1680

Table 5 Comparison of the noise level in simulated CT images CT scan type

Reference Bismuth-shielded Partial TCM

Heart region

Lung region

Noise level (HU)

Rel. difference to ref. scan (%)

Noise level (HU)

Rel. difference to ref. scan (%)

40.4 56.1 43.8 43.5

– 28.0 7.8 7.1

31.2 42.9 33.5 32.8

– 27.3 6.9 4.9

multiplied by the air kerma values of the reference scan; in our specific case, this amounted to 17.0 mGy. Dose distributions calculated on the basis of simulated CT images are shown in Fig. 2a–d for the reference scan, the bismuth-shielded scan, the partial CT scan and the TCM scan, respectively. All figures are shown on the same color scale. Relative dose values in the organs of interest, the breast and the exposed part of the bone marrow, are given in Table 2. Dose distributions calculated on the basis of measured CT images are shown in Fig. 3a–d with the thorax phantom and water-filled bags added for the different scan protocols; the relative mean dose values for the breast and the spine are listed in Table 3. The dose values (Tables 2, 3) and the dose distributions (Figs. 2, 3) for simulated and for measured CT examinations generally were in good agreement. Dose reduction to the breast was found to be in the range of -47.0 to -42.8% in bismuth shielding, -47.8 to -44.7% in partial CT scans, and -9.5 to 9.4% in TCM for simulations and measurements, respectively. Dose variations in the spine were found to be in the range of -6.0 to -3.0% in bismuth shielding, +17.9 to +18.7% in partial CT scans, and -12.5 to -11.3% in TCM for simulations and measurements, respectively. Values of the mean dose absorbed in the phantom are listed and compared in Table 4 for the different protocols. Bismuth-shielding, partial and TCM examination methods led to an overall dose reduction of 16.1, 7.0 and 12.2% in that order. With an uncertainty of ±10%, we state that the overall dose reduction is in the same order for all methods.

Image quality Figures 4 and 5 show the simulated and the measured CT images provided by the different CT approaches, respectively. Measured noise levels in the heart and the lung are shown in Table 5 for the simulations and in Table 6 for the measurements. Image noise increases in the region of the heart in the range of 28.0-23.5% using bismuth shielding, 7.8–1.2% performing partial CT scanning and 7.1–3.8% with TCM for simulations and measurements, respectively. In the lung the image noise ranges from 27.3 to 27.8% using bismuth shielding, 6.9 to 4.9% performing partial CT scanning and 4.9 to 2.9% with TCM for simulations and measurements, respectively. Noise value variations in the z-direction were about ±3% for the reference CT, bismuthshielded CT and TCM and slightly more heterogeneous with a variation of ±4% for the partial CT scan. This effect is due to reduced sampling in the z-direction during the otherwise unchanged continuous spiral CT trajectory, but did not show significant impact on image quality. In conclusion, image quality was comparable and homogeneous for the reference, the partial and the TCM methods. For the bismuth-shielded CT examination image noise increased over the whole image in the case of simulations as well as in the case of measurements. Streak artifacts were mainly observed where the shield lay in creases, which was the case for the measurements, and these artifacts were mainly found in the region of the breast where they are not necessarily relevant for diagnosis in

Table 6 Comparison of the noise level in measured CT images CT scan type

Reference Bismuth-shielded Partial TCM

Heart region

Lung region

Noise level (HU)

Rel. difference to ref. scan (%)

Noise level (HU)

Rel. difference to ref. scan (%)

40.7 53.2 41.2 42.3

– 23.5 1.2 3.8

33.2 46.0 34.9 34.2

– 27.8 4.9 2.9

1681

thoracic CT in general. The differences in noise are significant and may influence the diagnostic performance.

Discussion All results shown are given for the example of a semianthropomorphic thorax phantom with 25-cm anteroposterior and 35-cm lateral extension and breasts that are relatively flat, about 2 cm, and have a volume of 300 ml each. This does not mean a limitation of generality. Although absolute dose values and image noise may vary considerably depending on breast size and shape as well as phantom size, this does not affect the analysis of dose reduction efforts and, in particular, the relation of dose values for the different scan protocols. They proved to be valid for a wide range of breast and phantom sizes; detailed results are not included here since this would not lead to new or different conclusions. All dose reduction values for various phantom sizes with smaller, bigger, higher and flatter breasts were in the same order of magnitude. The percentage results are also independent of the scan parameters, such as pitch and mAs, and hold true over a wide range. All methods considered in our study led to a reduction of dose to the breast; these results are in agreement with published data [7, 22]. In the case of bismuth shielding, image quality was impaired. The noise level increased by about 25% in the heart region and about 30% in the lungs for the specific setup. If this noise increase were acceptable, it can be achieved in an easier and better way by reducing the mAs product. This would lead to an overall reduction of dose; for example, an accepted increase of noise by 25% corresponds to a dose or mAs product reduction of over 35% for the whole scanned region and not only for the shielded part. For partial CT scans, dose to the spine and thereby to bone marrow increased by 15–20%. Still, mean dose is lower for the partial CT scan compared with the reference protocol. To apply the idea of partial CT scan in other regions of the human body, the effect of heterogeneous dose distributions has to be analyzed carefully. Dose reduction in one radiosensitive organ can cause an increase of dose in another part of the body, where the radiation sensitivity of the tissue has to be also considered. In general, the partial CT scan approach appears to be of interest for reduction of dose to superficial radiosensitive organs only. Attenuation-dependent TCM is a technical improvement that can considerably reduce radiation dose [13–16]. Tube current is modulated in two respects in today’s CT systems.

One is the z-axis modulation and the other one is the angular modulation. The aim of z-axis modulation is homogeneous image quality; for this the tube current is automatically adapted to changes of the patient crosssection along the longitudinal axis. The angular TCM adjusts the tube current to minimize the X-ray intensity in projections with less attenuation and has little impact on the overall noise level. The modulation is determined in real time by using projection data that lag 180° behind. TCM reduces dose to the breast and to the whole body without a loss of image quality. Both attenuation-dependent tube current modulation and organ of risk-weighted tube current modulation or partial CT scan can help to reduce dose to the breasts and lower the effective dose to the patient without a significant loss of image quality. In summary, there are several viable approaches to reduce the dose to the female breast in thoracic CT. In our assessment, attenuation-dependent TCM appears to be the most practicable of the three investigated approaches. It is available on today’s CT systems and easily feasible in clinical routine. It reduces dose without a loss in image quality and does not require any additional efforts in patient preparation. For the future it would be of interest to have the option for partial CT scanning offered for thorax CT on commercial machines. We do not recommend the use of bismuth shielding because of their impact on image noise. In the case that an increase of image noise is accepted for diagnosis we would rather recommend lowering the tube current, which does not demand additional effort. In addition to the techniques investigated here, spectral optimization efforts and combinations of the approaches are to be considered in further work. And, no doubt, a general reduction of the mAs product whenever higher pixel noise appears acceptable remains the method of choice irrespective of the approaches discussed above.

Acknowledgements This work was financially supported by the EC-EURATOM 6 Framework Program (2002–2006) and forms part of the “Safety and Efficacy of Computed Tomography (CT): A broad perspective” project (contract no. FP/002388). We also would like to thank Tanja Gagliardi, M.D., Inselspital Bern, for fruitful discussions, inspirations and support with this topic. Many thanks to our colleagues Paul Deak, Oliver Langner and Michael Knaup from the Institute of Medical Physics in Erlangen for their support in dose calculations, dose measurements and image reconstruction.

References 1. American Cancer Society (2006) Cancer facts and figures 2006. American Cancer Society, Atlanta

2. Tokunaga M, Land CE, Yamamoto T (1987) Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950–1980. Radiat Res 112:243–272

3. Boice JD, Monson RR (1977) Breast cancer in women after repeated fluoroscopic examinations of the chest. J Natl Cancer Inst 59:823–832

1682

4. Shore RE, Hempelmann LH, Kowaluk E (1977) Breast neoplasms in women treated with x-rays for acute postpartum mastitis. J Natl Cancer Inst 59:813–822 5. International Commission on Radiological Protection (ICRP) (1991) 1990 Recommendations of the International Commission on Radiological Protection. ICRP Publication 60, Ann ICRP 21(1–3). Pergamon Press Oxford 6. International Commission on Radiological Protection (ICRP) (2008) Recommendations of the ICRP. ICRP Publication 103, Ann ICRP 37(2–3). Elsevier, Oxford 7. Geleijns J, Salvado Artells M, Veldkamp WJ, Lopez Tortosa M, Calzado Cantera A (2006) Quantitative assessment of selective in-plane shielding of tissues in computed tomography through evaluation of absorbed dose and image quality. Eur Radiol 16:2334–2340 8. McLaughlin DJ, Mooney RB (2004) Dose reduction to radiosensitive tissues in CT. Do commercially available shields meet the users’ needs? Clin Radiol 59:446–450

9. Schmidt B, Kalender WA (2002) Zusammenhang zwischen Reduktion der Patientendosis und des Röhrenstom-Zeit-Produktes bei unterschiedlichen Arten der Röhrenstrommodulation bei CTUntersuchungen. Fortschr Röntgenstr VO52.3 10. Schmidt B. Dosisberechnungen für die Computertomographie (2001) In: Kalender WA (ed) Berichte aus dem Institut für Medizinische Physik. Bd. 7. Shaker, Aachen 11. Kalender WA, Wolf H, Süß C, Gies M, Greess H, Bautz WA (1999) Dose reduction in CT by on-line tube current control: principles and validation on phantoms and cadavers. Eur Radiol 9:323–328 12. Kalender WA (2005) Computed tomog raphy, 2nd edn. Publicis, Erlangen 13. Greess H, Nömayr A, Wolf H et al (2002) Dose reduction in CT examination of children by an attenuation-based on-line modulation of tube current (CARE dose). Eur Radiol 12:1571– 1576 14. Kalra MK, Maher MM, Toth TL et al (2004) Strategies for CT radiation dose optimization. Radiology 230:619–628 15. Graser A, Wintersperger BJ, Suess C, Reiser MF, Becker CR (2006) Dose reduction and image quality in MDCT colonography using tube current modulation. AJR Am J Roentgenol 187:695–701 16. Mulkens TH, Daineffe S, De Wijngaert R et al (2007) Urinary stone disease: comparison of standard-dose and lowdose with 4D MDCT tube current modulation. AJR Am J Roentgenol 188:553–562

17. Ulzheimer S, Kalender WA (2003) Assessment of calcium scoring performance in cardiac computed tomography. Eur Radiol 13:484–497 18. McCollough CH, Ulzheimer S, Halliburton SS, Shanneik K, White RD, Kalender WA (2007) Coronary artery calcium: a multiinstitutional, multimanufacturer international standard for quantification at cardiac CT. Radiology 243:527–538 19. Kalender WA, Rienmüller R, Behr J, Seißler W, Fichte H, Welke M (1990) Quantitative CT of the lung with spirometrically controlled respiratory status and automated evaluation procedures. In: Fuchs W (ed) Advances in CT. Springer, Heidelberg, pp 85–93 20. Deak P, van Straten M, Shrimpton PC, Zankl M, Kalender WA (2007) Validation of a Monte Carlo tool for patientspecific dose simulations in X-ray computed tomography. Eur Radiol (in press) 21. Schmidt B, Kalender WA (2002) A fast voxel-based Monte Carlo method for scanner- and patient- specific dose calculations in computed tomography. Physica Medica XVIII(2) 22. Mulkens TH, Bellinck P, Baeyaert M et al (2005) Use of an automatic exposure control mechanism for dose optimization in multi-detector row CT examinations: clinical evaluation. Radiology 237:213–223