Eur J Nucl Med Mol Imaging (2010) 37:722–727 DOI 10.1007/s00259-009-1349-9
ORIGINAL ARTICLE
68
Ga-DOTA-NOC PET/CT in comparison with CT for the detection of bone metastasis in patients with neuroendocrine tumours
Valentina Ambrosini & Cristina Nanni & Maurizio Zompatori & Davide Campana & Paola Tomassetti & Paolo Castellucci & Vincenzo Allegri & Domenico Rubello & Giancarlo Montini & Roberto Franchi & Stefano Fanti
Received: 5 August 2009 / Accepted: 1 December 2009 / Published online: 27 January 2010 # Springer-Verlag 2010
Abstract Purpose To retrospectively evaluate the sensitivity, specificity and accuracy of 68Ga-DOTA-NOC PET/CT and CT alone for the evaluation of bone metastasis in patients with neuroendocrine tumour (NET). Methods From among patients with NET who underwent 68 Ga-DOTA-NOC PET/CT between April 2006 and November 2008 in our centre, 223 were included in the study. Criteria for inclusion were pathological confirmation of NET and a follow-up period of at least 10 months. PET and V. Ambrosini : C. Nanni : P. Castellucci : V. Allegri : G. Montini : R. Franchi : S. Fanti Department of Nuclear Medicine, S. Orsola-Malpighi University Hospital, Bologna, Italy M. Zompatori Department of Radiology, S. Orsola-Malpighi University Hospital, Bologna, Italy D. Campana : P. Tomassetti Department of Internal Medicine, S. Orsola-Malpighi University Hospital, Bologna, Italy D. Rubello Department of Nuclear Medicine, S. Maria della Misericordia Hospital, Rovigo, Italy S. Fanti (*) Unità Operativa di Medicina Nucleare, Padiglione 30, Policlinico S.Orsola-Malpighi, Azienda Ospedaliero Universitaria di Bologna, Via Massarenti 9, 40138 Bologna, Italy e-mail:
[email protected]
CT images were retrospectively reviewed by two nuclear medicine specialists and two radiologists, respectively, without knowledge of the patient history or the findings of other imaging modalities. PET data were compared with the CT findings. Interobserver agreement was evaluated in terms of the kappa score. Clinical and imaging follow-up were used as the standard of reference to evaluate the PET findings. Results PET was performed for staging (49/223), unknown primary tumour detection (24/223), restaging (32/223), restaging before radioimmunotherapy (1/223), evaluation during therapy (12/223), equivocal findings on conventional imaging (4/223 at the bone level; 61/223 at sites other than bone), and follow-up (40/223). A very high interobserver agreement was observed. CT detected at least one bone lesion in only 35 of 44 patients with a positive PET scan. In particular, PET showed more lesions in 20/35 patients, a lower number of lesions in 8/35, and the same number in 7/ 35. The characteristics of the lesions (sclerotic, lytic, mixed) on the basis of the CT report did not influence PET reading. PET revealed the presence of at least one bone metastasis in nine patients with a negative CT scan. Considering patients with a negative PET scan (179), CT showed equivocal findings at the bone level in three (single small sclerotic abnormality in two at the spine level, and bilateral small sclerotic abnormalities in the humeri, femurs and scapula). Clinical follow-up confirmed the PET findings in all patients; thus there were no falsepositive or false-negative findings. Considering all patients, PET detected more lesions than CT (246 vs. 194). As compared to CT, on a patient basis PET showed a higher sensitivity (100% vs. 80%), specificity (100% vs. 98%), positive predictive value (100% vs. 92%), and negative predictive value (100% vs. 95%).
Eur J Nucl Med Mol Imaging (2010) 37:722–727
Conclusion In conclusion, 68Ga DOTA-NOC PET was more accurate than CT for the identification of bone lesions and led to a change in clinical management in nine patients with a negative CT scan. Keywords 68Ga-DOTA-NOC . PET . CT . Neuroendocrine tumours . Bone lesions
Introduction PET with 68Ga-DOTA-NOC has recently been reported to be very accurate for the assessment of patients with neuroendocrine tumours (NET). The routine diagnostic work-up of patients with NET includes morphological imaging techniques such as computed tomography (CT), ultrasound (US) and magnetic resonance imaging (MRI) [1] combined with functional imaging, namely whole-body somatostatin receptor scintigraphy (SRS) [2]. However, small NET lesions and variable anatomical location are the major diagnostic challenges [3–5]. Recent papers described the superior value of PET in the diagnosis of NET over both CT and SRS [6, 7]. PET has been increasingly used to study patients with NET following the introduction of new PET tracers (68Ga-DOTA peptides) that specifically bind to somatostatin receptors (SSR) on the surface of NET cells. In previous studies the results using 18F-FDG PET in well-differentiated NET have been unsatisfactory as a consequence of their slow growth rate [8]. On the contrary, differentiated NET overexpress SSR on the cell surface and can therefore be visualized on PET scans using somatostatin analogues labelled with gallium. Among the different DOTA peptides (TOC, TATE, NOC), 68Ga-DOTA-NOC shows the broadest SSR subtype affinity, binding to SSR2, SSR3 and SSR5 [9], and a more favourable dosimetry [10]. The aim of the present study was to retrospectively evaluate the accuracy of 68Ga-DOTA-NOC PET/CT for the detection of bone metastatic lesions and to compare PET data with the corresponding CT findings and clinical follow up.
Materials and methods We retrospectively reviewed all 68Ga-DOTA-NOC PET/CT studies carried out between April 2006 and November 2008. Criteria for inclusion were pathological confirmation of NET (on either the primary or a metastatic site) and a follow-up period of at least 10 months. 68 Ga-DOTA-NOC was synthesized at the Radiopharmacy of the Nuclear Medicine Unit. 68Ga was eluted from a 68 Ge/68Ga generator (produced by Cyclotron, Obninsk, Russia) and DOTA-NOC was labelled with 68Ga following
723
the procedure described by Zhernosekov et al. [11]. PET/ CT with 68Ga-DOTA-NOC (intravenously injected dose 120– 185 MBq, uptake time 60 min) was carried out using a dedicated hybrid PET/CT tomograph (Discovery STE and Discovery LS; GE Medical Systems, Waukesha, WI). PET scan emission images were recorded for 4 min per bed position (Discovery LS) or 3 min per bed position (Discovery STE); for nonuniform attenuation correction, CT images were used (acquisitions parameters: 140 kV, 90 mA, 0.8 s tube rotation, 5 mm thickness). PET images were acquired from the skull base to the middle part of the thigh. For the evaluation of PET studies, any area with an intensity greater than background that could not be identified as physiological activity (pituitary gland, spleen, liver, adrenal glands, head of the pancreas, thyroid, urinary tract) was considered to be indicative of tumour tissue; for the purposes of the study we only considered PET findings at the bone level. In particular, PET findings regarding the site and number of bone lesions were compared with the corresponding CT data. Corresponding PET and CT images were retrospectively reviewed by two experienced nuclear medicine specialists and by two experienced radiologists unaware of the patients’ medical history and of the results of the other imaging modalities. The total number of bone lesions was calculated in each patient for both PET and CT. Multiple and diffuse uncountable lesions were arbitrarily classified as ten lesions. Radiologists were also asked to describe the characteristics of bone lesions (lytic, sclerotic, mixed) on the basis of visual analysis. Each reviewer interpreted the studies independently and the interobserver agreement was evaluated in terms of the kappa score (κ < 0, poor; κ = 0–0.2, slight; κ = 0.2–0.4, fair; κ = 0.4–0.6, moderate; κ = 0.6–0.8, substantial; and K = 0.8–1.0, almost perfect) [12]. Clinical and imaging follow-up data (16 months, range 10–18 months) were used as the reference standard to finally evaluate the PET results as true-positive, true-negative, false-positive and falsenegative. Follow-up included the repetition of the 68GaDOTA-NOC scan, or the confirmation of the same lesions by any imaging procedure such as CT, MR or Octreoscan. The sensitivity, specificity, positive predictive value and negative predictive value of PET and CT scans were calculated using established methods and were on a perperson basis. The McNemar test (SPSS 13.0 version) was used to compare the sensitivity and specificity of PET and CT.
Results 68
Ga-DOTA-NOC PET and CT images from 223 patients with NET (107 men, 116 women; mean age 58 years, range 26–80 years) were retrospectively reviewed to evaluate the accuracy of PET for the detection of NET bone lesions.
724
Eur J Nucl Med Mol Imaging (2010) 37:722–727
PET was performed for staging (49/223), unknown primary tumour detection (24/223), restaging (32/223), restaging before radioimmunotherapy (1/223), evaluation during therapy (12/223), equivocal findings on conventional imaging (4/223 at the bone level; 61/223 at sites other than bone), and follow-up (40/223). The primary tumour sites included the pancreas (64/223), gastroenteric sites other than pancreas (55/223; colon 2, appendix 12, duodenum 5, stomach 11, ileum 25), lung (44/ 223), thyroid (15/223), paraganglioma (5/223), prostate NET (2/223), retroperitoneal carcinoid (1/223), unknown primary tumour (18/223), adrenals (8/223), hypophysis (2/223), ear (1/223), ovary (1/223), kidney (2/223), skin (2/223; 1 Merkel carcinoma, 1 melanoma), testicle (1/223), thymus (1/223), and central nervous system (1/223). The kappa score, applied as a measure of interobserver agreement between the two nuclear medicine specialists and the two radiologists, was equal to 1 in both cases. Considering all cases, PET detected at least one bone lesion in 44 patients and was negative in 179 patients. The sites of secondary bone lesions detected by PET and CT are listed in Table 1. CT detected at least one bone lesion in 35 (80%) of the 44 patients with a positive PET scan. Among the 35 patients with positive PET and CT scans, PET detected a higher number of lesions in 20 (57%; 14 sclerotic, 6 mixed) and a lower number in 8 (23%; 5 sclerotic, 2 mixed, 1 lytic), and PET and CT were concordant in 7 (20%; 6 sclerotic, 1 mixed). PET led to the detection of at least one bone metastasis in nine patients with a negative CT scan (Fig. 1). Sites of bone metastasis identified by PET in patients with a negative CT scan included: temporal bone (three lesions), vertebrae (four), ilium (three), ribs (one) and humerus (one). In two of the three lesions identified by PET at the temporal bone level, PET led to the detection of
viable tumour at the periphery of the surgical area (SUVmax 25 and 29.7, respectively, 3 weeks after surgery of the primary tumour). In this subgroup of nine patients with at least one bone metastasis detected by PET and a negative CT scan, the PET findings led to a change in the clinical management: the detection of NET bone lesions upstaged the disease in four patients and excluded surgery in two patients, and as a result of the demonstration of SSR expression, three patients received radionuclide receptor therapy. Considering
Table 1 Sites of secondary bone lesions detected by PET and CT Site
Ribs/scapula Vertebrae Head bones Pelvis Sternum Long bones Diffuse (uncountable)a Total a
Number of lesions PET
CT
24 52 5 38 2 15 110 (11 patients) 246
17 75 2 45 5 10 40 (4 patients) 194
Patients with more than ten diffuse lesions were assumed to have ten lesions.
Fig. 1 68Ga-DOTA-NOC PET/CT image shows the presence of a very small rib lesion while the CT image is negative
Eur J Nucl Med Mol Imaging (2010) 37:722–727
a
Fig. 2
725
b
c
68
Ga-DOTA-NOC PET/CT images show multiple bone lesions not visible on the CT-alone images (a vertebrae, b ileum, c humerus)
all 223 patients, PET showed more lesions (246 vs. 194) than CT (Fig. 2). Of the 179 patients with a negative PET scan, CT was negative for bone lesions in 176 but showed equivocal findings at the bone level in 3 (single small sclerotic abnormality in two patients at the spine level, and bilateral small sclerotic abnormalities in the humeri, femurs and scapula). Clinical and imaging follow-up confirmed the PET findings in all patients; thus there were no falsepositive nor false-negative PET findings. Considering all patients, PET showed a higher sensitivity than CT (100% vs. 80%; p=0.008), while specificity was comparable (100% vs. 98%; p=0.2). The positive predictive value and negative predictive value of PET were also higher than those of CT (100% vs. 92% and 100% vs. 95%, respectively).
Discussion Both morphological imaging techniques (CT, US, MRI) and functional imaging modalities (SRS) have been routinely employed in the routine diagnostic work-up of patients with NET. The value of conventional imaging procedures in the assessment of NET is mainly limited by small lesion size and variable anatomical location and in the evaluation of residual tumour after therapy. Among nuclear medicine procedures, SRS shows good accuracy for whole-body imaging and is routinely used [2]. However, the low spatial resolution of the gamma camera is the major limitation in the use of SRS, particularly for the assessment of small lesions and for the evaluation of organs with a high physiological uptake (e.g. liver).
726
More recently, PET has been increasingly used for the evaluation of patients with NET. It is well known that 18FFDG has several limitations for the identification of differentiated NET [8]. Therefore alternative tracers (18FDOPA and in particular 68Ga-labelled peptides) have been suggested and successfully employed. Belonging to the APUD cell system, NET cells can be visualized on 18FDOPA PET scans, but the high cost and the more difficult synthesis are the major factors that limit its use in the evaluation of patients with NET. In a series of 23 patients with advanced stage NET, 18F-FDOPA has been found to be able to accurately detect skeletal lesions (sensitivity 100%; specificity 91%), and PET identified bone metastasis in 40% of patients with a negative CT scan [13]. Increasing reports indicate that 68Ga-DOTA-peptides as very promising tracers to visualize NET [6, 14, 15]. The mechanism of uptake is receptor based: 68Ga-DOTApeptides can bind to SSR expressed on NET cells with variable affinity [16, 17]. Therefore PET with 68Ga-DOTApeptides is not only useful from a diagnostic point of view to detect NET lesions but also to provide data on tumour cells receptors expression, an indirect measure of cells differentiation [18]. Several different somatostatin analogues (DOTA-TOC, DOTA-TATE, DOTA-NOC) have been described. Although DOTA-TOC, DOTA-TATE and DOTA-NOC can bind to SSR2 (the predominant receptor type in NET) and to SSR5, only 68Ga-DOTA-NOC also shows good affinity for SSR3 [9]. Another advantage of the use of 68Ga-DOTA-NOC is its dosimetry, since the dose delivered to organs is comparable to, or even lower than, the doses of analogous diagnostic compounds [10]. To our knowledge, this is the first study to compare 68Ga-DOTA-NOC PET and CT for the detection of NET secondary bone lesions. Our data support the superiority of 68Ga-DOTA-NOC PET for the evaluation of NET secondary bone lesions as compared to CT. PET showed bone involvement in almost 20% of our patients, detecting the presence of bone lesions in nine patients with a negative CT scan. Preliminary published data in a limited patient population also support the superiority of 68Ga-DOTA peptides PET over other imaging modalities for the detection of NET bone metastasis. In the largest published series, including 84 patients with advanced NET studied by 68Ga-DOTATOC PET, SRS and CT, PET identified 116 bone lesions as compared to 84 of 116 identified by SRS and 58 of 116 by CT [6]. In a more recent study including 51 patients, Putzer et al. found that 68Ga-DOTA-TOC was more accurate (sensitivity 97%, specificity 92%) than CT and bone scan for the early detection of NET bone lesions [19]. With regard to 68Ga-DOTA-NOC, in a limited population of 13 patients with gastroenteropancreatic tumours and lung NET, PET detected bone lesions in ten patients in contrast to CT
Eur J Nucl Med Mol Imaging (2010) 37:722–727
alone which detected bone lesions in five patients [15]. In another study including 11 patients with lung NET, PET with 68Ga-DOTA-NOC identified more bone lesions in two patients than CT [14]. The accurate evaluation of bone involvement has been reported to be particularly relevant since the presence of extrahepatic secondary lesions is associated with poorer survival and is a contraindication for extended surgical resection [20]. In a population of 156 patients with gastroenteropancreatic NET, the survival rate was lower (50.1%) in those with extrahepatic secondary lesions (half of them at the bone level) than in those with only hepatic (73.3%) or nodal (77%) involvement or those with only local disease (96%). The presence of extrahepatic sites of disease has been suggested to be a marker of a subgroup of patients with a worse prognosis and shorter survival who would benefit more from aggressive therapeutic approaches [21]. Considering the patients positive by both PET and CT, PET detected a higher number of lesions in almost twothirds. Overall, PET showed a higher sensitivity, positive predictive value and negative predictive value than CT while specificity was comparable. CT, bone scans and SRS have long been considered the most important procedures for the assessment of bone lesions, but have several limitations for the assessment of NET bone lesions. The low spatial resolution of the gamma camera is the major limitation of SRS and bone scintigraphy. Moreover, active metastases cannot always be distinguished from a repair process in bone scintigraphic images. With regard to CT, signs of bone involvement may be subtle and difficult to interpret only on an anatomical basis, and the presence of viable residual tumour may be difficult to interpret after treatment. In our study the characteristics of bone lesions as described on the CT report (sclerotic, lytic, mixed) did not influence reporting and did not influence the PET reading results since sclerotic, lytic and mixed lesions were identified in all patient subgroups irrespective of whether PET identified more, fewer or an equal number of lesions than CT. This finding is particularly relevant since small sclerotic lesions are quite frequently encountered and may represent a challenge for CT image interpretation. Our data suggest that these findings are not to be considered as an unequivocal sign of the presence of metastatic disease. On the other hand, 68Ga-DOTA-NOC PET may give falsepositive findings in patients with an inflammatory process (since activated lymphocytes express SSR), while low or absent SSR expression (for example, in those with dedifferentiated metastasis) may lead to false-negative results. In this latter case the use of metabolic tracers such as 18F-FDG and 18F-DOPA may offer advantages over 68 Ga-DOTA peptides.
Eur J Nucl Med Mol Imaging (2010) 37:722–727
Although the accuracy for the detection of welldifferentiated NET bone lesions is higher for PET with 68 Ga-DOTA peptides, the employment of both conventional and functional imaging procedures may prove very helpful and provide complementary information in equivocal cases, in which the possibility of describing different lesion characteristics (anatomical location, receptor expression pattern, metabolic activity) may offer a clearer understanding of the feature of the lesion and of the most appropriate line of intervention.
727
9.
10.
11.
12.
Conclusion In conclusion, our study showed that 68Ga DOTA-NOC PET was more accurate than CT for the identification of bone lesions and led to a change in the clinical management in nine patients with a negative CT scan.
13.
14.
15.
References 1. Sundin A, Garske U, Orlefors H. Nuclear imaging of neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007;21(1):69–85. 2. Ramage JK, Davies AH, Ardill J, Bax N, Caplin M, Grossman A, et al. Guidelines for the management of gastroenteropancreatic neuroendocrine (including carcinoid) tumours. Gut 2005;54(Suppl 4):iv1–6. 3. Kaltsas G, Rockall A, Papadogias D, Reznek R, Grossman AB. Recent advances in radiological and radionuclide imaging and therapy of neuroendocrine tumours. Eur J Endocrinol 2004;151:15–27. 4. Kumbasar B, Kamel IR, Tekes A, Eng J, Fishman EK, Wahl RL. Imaging of neuroendocrine tumors: accuracy of helical CT versus SRS. Abdom Imaging 2004;29:696–702. 5. Montravers F, Grahek D, Kerrou K, Ruszniewski P, de Beco V, Aide N, et al. Can fluorodihydroxyphenylalanine PET replace somatostatin receptor scintigraphy in patients with digestive endocrine tumors? J Nucl Med 2006;47:1455–62. 6. Gabriel M, Decristoforo C, Kendler D, Dobrozemsky G, Heute D, Uprimny C, et al. 68Ga-DOTA-Tyr3-octreotide PET in neuroendocrine tumors: comparison with somatostatin receptor scintigraphy and CT. J Nucl Med 2007;48(4):508–18. 7. Buchmann I, Henze M, Engelbrecht S, Eisenhut M, Runz A, Schafer M, et al. Comparison of (68)Ga-DOTATOC PET and (111)In-DTPAOC (Octreoscan) SPECT in patients with neuroendocrine tumours. Eur J Nucl Med Mol Imaging 2007;34 (10):1617–26. 8. Adams S, Baum R, Rink T, Schumm-Dräger PM, Usadel KH, Hör G. Limited value of fluorine-18 fluorodeoxyglucose positron
16.
17.
18.
19.
20.
21.
emission tomography for the imaging of neuroendocrine tumors. Eur J Nucl Med 1998;25:79–83. Antunes P, Ginj M, Zhang H, Waser B, Baum RP, Reubi JC, et al. Are radiogallium-labelled DOTA-conjugated somatostatin analogues superior to those labelled with other radiometals? Eur J Nucl Med Mol Imaging 2007;34(7):982–93. Pettinato C, Sarnelli A, Di Donna M, Civollani S, Nanni C, Montini G, et al. 68Ga-DOTANOC: biodistribution and dosimetry in patients affected by neuroendocrine tumors. Eur J Nucl Med Mol Imaging 2008;35(1):72–9. Zhernosekov KP, Filosofov DV, Baum RP, Aschoff P, Bihl H, Razbash AA, et al. Processing of generator-produced 68Ga for medical application. J Nucl Med 2007;48(10):1741–8. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74. Becherer A, Szabó M, Karanikas G, Wunderbaldinger P, Angelberger P, Raderer M, et al. Imaging of advanced neuroendocrine tumors with (18)F-FDOPA PET. J Nucl Med 2004;45(7):1161–7. Ambrosini V, Castellucci P, Rubello D, Nanni C, Musto A, Allegri V, et al. 68Ga-DOTA-NOC: a new PET tracer for evaluating patients with bronchial carcinoid. Nucl Med Commun 2009;30 (4):281–6. Ambrosini V, Tomassetti P, Castellucci P, Campana D, Montini G, Rubello D, et al. Comparison between 68Ga-DOTA-NOC and 18F-DOPA PET for the detection of gastro-entero-pancreatic and lung neuro-endocrine tumours. Eur J Nucl Med Mol Imaging 2008;35(8):1431–8. Wild D, Schmitt JS, Ginj M, Mäcke HR, Bernard BF, Krenning E, et al. DOTA-NOC, a high-affinity ligand of somatostatin receptor subtypes 2, 3 and 5 for labelling with various radiometals. Eur J Nucl Med Mol Imaging 2003;30(10):1338–47. Wild D, Macke HR, Waser B, Reubi JC, Ginj M, Rasch H, et al. 68Ga-DOTANOC: a first compound for PET imaging with high affinity for somatostatin receptor subtypes 2 and 5. Eur J Nucl Med Mol Imaging 2005;32(6):724. von Falck C, Boerner AR, Galanski M, Knapp WH. Neuroendocrine tumour of the mediastinum: fusion of 18F-FDG and 68GaDOTATOC PET/CT datasets demonstrates different degrees of differentiation. Eur J Nucl Med Mol Imaging 2007;34(5):812. Putzer D, Gabriel M, Henninger B, Kendler D, Uprimny C, Dobrozemsky G, et al. Bone metastases in patients with neuroendocrine tumor: 68Ga-DOTA-Tyr3-octreotide PET in comparison to CT and bone scintigraphy. J Nucl Med 2009;50 (8):1214–21. Gupta S, Johnson MM, Murthy R, Ahrar K, Wallace MJ, Madoff DC, et al. Hepatic arterial embolization and chemoembolization for the treatment of patients with metastatic neuroendocrine tumors: variables affecting response rates and survival. Cancer 2005;104:1590–602. Panzuto F, Nasoni S, Falconi M, Corleto VD, Capurso G, Cassetta S, et al. Prognostic factors and survival in endocrine tumor patients: comparison between gastrointestinal and pancreatic localization. Endocr Relat Cancer 2005;12:1083–92.