| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Nuclear Medicine and Endocrine Oncology (A.L.G., S.L., J.L., E.B., M.S.), Radiology (D.V., C.D., L.C., N.L.), Biostatistics and Epidemiology (A.A., N.N.T.), Medical Physics Unit (G.B.), and Surgery (D.H., J.-P.T.), Institut Gustave Roussy and Faculté de Médecine Paris-Sud, 94805 Villejuif Cédex, France
Address all correspondence and requests for reprints to: Martin Schlumberger, Nuclear Medicine and Endocrine Oncology, Institut Gustave Roussy, 94805 Villejuif Cédex, France. E-mail: schlumbg{at}igr.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
Patients and Methods: Fifty-five consecutive elevated calcitonin level MTC patients were enrolled to undergo neck and abdomen ultrasonography (US); neck, chest, and abdomen spiral computed tomography (CT); liver and whole-body magnetic resonance imaging (MRI); bone scintigraphy; and 2-[fluorine-18]fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET)/CT scan (PET).
Results: Fifty patients underwent neck US, CT, and PET, and neck recurrence was demonstrated in 56, 42, and 32%, respectively. Lung and mediastinum lymph node metastases in the 55 patients were demonstrated in 35 and 31% by CT and in 15 and 20% by PET. Liver imaging with MRI, CT, US, and PET in 41 patients showed liver in 49, 44, 41, and 27% patients, respectively. Bone metastases in 55 patients were demonstrated in 35% by PET, 40% by bone scintigraphy, and 40% by MRI; bone scintigraphy was complementary with MRI for axial lesions but superior for the detection of peripheral lesions. Ten patients had no imaged tumor site despite elevated calcitonin level (median 196 pg/ml; range 39–816). FDG uptake in neoplastic foci was higher in progressive patients but with a considerable overlap with stable ones.
Conclusion: The most efficient imaging work-up for depicting MTC tumor sites would consist of a neck US, chest CT, liver MRI, bone scintigraphy, and axial skeleton MRI. FDG PET scan appeared to be less sensitive and of low prognostic value.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
Many radiopharmaceuticals have been used but did not prove to be more sensitive than computed tomography (CT) or magnetic resonance imaging (MRI) (8, 9, 10, 11). 2-[Fluorine-18]fluoro-2-deoxy-D-glucose (FDG)-positron emission tomography (PET) is an imaging standard for many cancer types but produced controversial results in MTC patients (12, 13, 14, 15, 16, 17). 18F-Dihydroxyphenylalanine (18, 19, 20), immunoscintigraphy with labeled monoclonal antibodies directed against carcinoembryonic antigen (CEA) (21), and gastrin receptor scintigraphy seem to be promising (22) but are not widely available. Venous sampling catheterization with calcitonin determinations may localize small neoplastic foci (4, 23), and angiography or laparoscopy may demonstrate liver metastases, but these techniques are invasive (24, 25).
Many techniques may thus be used for imaging MTC patients, but no consensus has yet been reported on their selective use. Also, their performances depend on the protocol used and on the standard to which they are compared. Indeed, we recently demonstrated how standardized protocols may improve the diagnostic performances of liver CT scan and MRI in patients with other endocrine tumors (26).
Finally, survival in many other primary tumors is related to the level of FDG uptake (27, 28, 29). However, FDG uptake in MTC patients is usually low, and the usefulness of FDG-PET in such patients has been questioned (16).
To assess the interest of various imaging procedures, we prospectively studied 55 consecutive MTC patients with persistent elevated calcitonin levels and performed a standardized imaging staging with optimized protocols including neck and abdomen ultrasonography (US); neck, chest, and abdomen spiral CT scan (CT); liver MRI and whole-body (WB) MRI; bone scintigraphy (BS); and FDG PET/CT scan (PET). Then we compared results of these imaging procedures and suggested an imaging strategy for these patients. Finally, we assessed the prognostic significance of FDG uptake on PET.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
All histologically proven MTC patients with elevated basal calcitonin levels followed up at the Institut Gustave Roussy were prospectively enrolled in the study (30). Patients were at least 3 months beyond any major surgery, external radiation therapy, or any chemotherapy. Before study, patients were either free of imaged disease (n = 18) or known to have lesions (n = 37).
The study evaluated disease extent in the neck by US, CT, and PET; in the lungs and mediastinum by CT and PET; in the liver by US, CT, MRI, and PET; and in bones by WB-MRI, BS, and PET. All imaging procedures were completed within 6 wk. The protocol was approved by the Institut Gustave Roussy Scientific Committee.
Imaging modalities
US examination was performed with an Aplio machine (Toshiba, Japan). A high-frequency, wide-band probe (8–14 MHz) was used for neck examination. Power Doppler ultrasonography assessed node vascularity. Malignancy criteria were any hypoechoic nodule or mass with peripheral hypervascularization and presence of microcalcifications (31). Persistence of hilum in lymph node was in favor of its benignity. Liver US was performed in B-mode imaging using curvilinear transducers with a center frequency of 2.5–4 MHz. Hypo- or hyperechogenic lesions (hemangioma-like lesions), yet unknown or increasing in size, and calcified lesions were considered to be liver metastases (32).
BS was performed using a double-headed 
-camera equipped with low-energy, high-resolution collimators (Ecam; Siemens, Erlangen, Germany). It consisted of a 15 cm/min speed whole-body bone scan, and in spot lateral views of the ribs and skull and on any equivocal uptake, done 3–4 h after the injection of 740 MBq 99m technetium methylene diphosphonate (Osteocis; CisBio International, Gif-Sur-Yvette, France). Any extraarticular unexplained high bone uptake was considered to be a bone metastasis after exclusion of recent surgery or trauma.
Neck, chest, and abdomen CT was performed with a HiSpeed spiral scanner (GE Medical Systems, Milwaukee, WI). Liver CT images were obtained before and after a monophasic injection of 100 ml of nonionic contrast medium (Xenetix 300, Guerbet, France). Three spiral CT acquisitions were obtained during the hepatic arterial phase, the portal venous phase, and the equilibrium phase at 30, 70, and 300 sec after the initiation of the injection, respectively. Any liver lesion fulfilling the previously described malignancy criteria was considered to be a metastasis (26).
MRI was performed with a 1.5-T whole-body imager (Signa LX; General Electric Medical Systems). Liver MRI was acquired in the axial plane with a phased-array body multicoil. Slice thickness was 7 mm, with a 2-mm intersection gap for all pulse sequences. Fat-suppressed T2-weighted imaging, breath-hold single-shot fast spin-echo, dynamic contrast-enhanced MRI, T1-weighted sequences with fast multiplanar spoiled gradient-recalled echo imaging was acquired. Any liver lesion fulfilling the previously described malignancy criteria was considered to be a metastasis (33). WB-MRI was performed in the supine position with a body coil. To cover the axial and appendicular skeleton, four contiguous coronal acquisitions with two different magnetic resonance sequences were used: short tau inversion recovery and spin echo T1. No respiratory gating was performed during the acquisition. No iv contrast agent was administered. Any bone lesion with a low signal on T1 and a high signal on short time inversion recovery images was considered to be a metastasis (34, 35).
PET was performed on an integrated PET-CT Biograph LSO system (Siemens Medical Solutions, Erlangen, Germany) after the iv injection of 5 MBq/kg of 18-FDG, followed by a 60-min uptake phase. Reconstruction data were acquired with a single slice spiral CT (Somatom Emotion, Siemens Medical Solutions) without iodine contrast injection. The acquisition included the whole body from head to toes. Any unexplained FDG uptake superior to background, corresponding or not to a visible lesion on CT, was considered to be a metastasis. The maximum standardized uptake value (SUV) was determined in metastatic lesions.
|
Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
MRI and CT were reviewed by one experienced radiologist, PET and BS by two experienced nuclear medicine specialists, and US by two experienced echographists, all unaware of imaging findings concerning the patients. Each reader scored the number of metastases depicted at each examination. Only lesions fulfilling malignancy criteria defined above were considered metastases. Imaging discrepancies were resolved by consensus reading, patient follow-up, and for bone lesions by a lesion focused MRI. Neck lesions could be histologically proven in the nine patients who were subsequently submitted to surgery.
A per-patient and per-lesion analysis was done in each area defined as: 1) neck including thyroid bed and cervical lymph nodes, 2) mediastinum, 3) lungs, 4) liver, and 5) bones. Bone lesions were classified as low mechanical risk (ribs, sternum, clavicles, and skull) or high mechanical risk lesions (spine, pelvic bones, long bones).
To assess the procedure sensitivity in each area, we arbitrarily set the reference number to be the sum of the metastases depicted for each patient by all combined modalities. Then we calculated the ratio between the total number of metastases determined by each procedure to the reference number. When more than 10 lesions were depicted in any localization, or in ribs, long or pelvic bones, spine, and skull, 11 lesions were considered for the statistical analysis.
In the 45 patients with demonstrated lesions, the comparison of imaging modalities performed at an interval of 1 yr permitted to classify disease as either progressive in case of appearance of new lesions or an increase by 20% or more of the sum of the longest diameters of the target lesions or as stable disease, according to response evaluation criteria in solid tumors (36). No treatment was given during this study period.
Statistical method
The detection of lesions in each organ by each imaging modality was compared using the McNemar test for matched proportions.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
Fifty-five consecutive MTC patients were enrolled in the study from January 2005 until August 2006. There were 21 women and 34 men, with a median age of 56 yr (range 14–83 yr). Five patients had a multiple endocrine neoplasia 2a syndrome, three had a multiple endocrine neoplasia 2b syndrome, four had a familial MTC, and finally 43 had a sporadic MTC. According to the 2002 pathological tumor node metastasis (pTNM) classification (37), the thyroid tumor was classified as pT1 in five patients, pT2 in 14, pT3 in 11, pT4a in 12, pT4b in nine, and pTx in four patients; pN0 in four patients and pN1 in 51; M1 in 14 patients (in bones in seven patients, lungs in five, and the liver in eight). Median follow-up after initial surgery was 79 months (range 3–309 months). At the time of the study, plasma calcitonin level ranged from 21 to 247,000 pg/ml (median 1250), and CEA level ranged from 1 to 5845 ng/ml (median 37).
Imaging results
The standardized imaging work-up led to the diagnosis of local recurrence or metastases in 45 of 55 (82%) patients with a median calcitonin level at 1534 pg/ml (range 21–247,000 pg/ml) but found no lesion in the remaining 10 (18%) patients with a median calcitonin level at 196 pg/ml (range 39–816). Unknown lesions were discovered in eight patients using all these procedures.
FDG-PET alone evidenced metastases in only 32 of 55 (58%) patients with a median calcitonin level at 2311 pg/ml (range 51–247,000) but demonstrated no lesion in the remaining 23 patients with a median calcitonin at 654 pg/ml (range 21–24,600). In no patient was the FDG-PET the only abnormal tool.
Disease extent is summarized in Fig. 1
and results per site with each procedure in Table 1. Forty-five patients had known tumor foci, in the neck in 32 patients, the mediastinum in 17, the lungs in 19, the liver in 25, bones in 25, and at other sites in four. Lesions were located only in the neck or mediastinum in 10 patients (in the neck only in eight patients, the mediastinum only in one patient, and the neck and mediastinum in one patient) and only at distant sites in nine patients. In the other 26 patients, neck and mediastinum involvement was associated with one distant metastatic site in seven patients, with two to four distant metastatic sites in 19.
|
|
Neck recurrence was demonstrated in 29 of the 50 patients (58%) submitted to the three imaging modalities, including neck ultrasonography, CT, and PET. These three imaging modalities were concordant in 34 of 50 patients (showing neck recurrence in 13 and none in 21), two modalities showed neck recurrence in nine patients, and only one imaging modality showed neck recurrence in seven patients, neck US in five patients, neck CT in one, and PET in one patient who also had distant metastases. US demonstrated 97% of the neck recurrences, CT 72%, and PET 55%. US was the most efficient modality (US vs. CT: P = 0.046 and US vs. PET: P < 0.01). There was no statistical difference between CT and PET (P = 0.23). Neck lesions were histologically proven in the nine patients who underwent subsequent surgery. US was positive in all nine of these patients, PET in six, and CT in five.
A total of 125 lesions were found: 124 by US, 76 by CT, and 40 by PET [median SUV maximum 3 (range 1–5)]. On neck US, six patients had more than 10 lymph nodes, and the other patients had one up to eight lymph nodes.
Mediastinum lymph nodes
Mediastinum lymph nodes were demonstrated by spiral CT in 17 of the 55 patients (31%) and PET demonstrated abnormal uptake in the mediastinum in only 11 of these 17 patients (P = 0.04).
CT demonstrated 103 lesions and PET 72 foci of abnormal uptake [median SUV maximum 3 (range 1–6)]. On CT, seven patients had more than 10 lymph nodes and the other patients had one up to six lymph nodes.
Lung metastases
Lung metastases were demonstrated by spiral CT in 19 of the 55 patients (35%) and by PET in only eight of these 19 patients (P = 0.003). Lung metastases and mediastinum lymph nodes were both demonstrated in 11 patients.
Lung metastases were miliary in 15 of 19 patients (79%) (Fig. 2). CT demonstrated 172 lesions and PET 59 foci of abnormal uptake [median SUV maximum 1 (range: 1–4)].
|
Liver metastases were demonstrated in 20 of the 41 patients (49%) submitted to the four imaging modalities, including liver US, CT, MRI, and PET. These four imaging modalities were concordant in 31 of 41 patients (showing lesions in 10 and none in 21). US demonstrated liver metastases in 85% of these patients, CT in 90%, MRI in 100%, and PET in 55%. MRI was the only abnormal imaging modality in two patients with lesions of a few millimeters in diameter. US, CT, and MRI had similar sensitivity, and CT and MRI were more efficient than PET (CT vs. PET: P = 0.023 and MRI vs. PET: P = 0.008) (Fig. 3). The difference between liver US and PET did not reach significance (P = 0.08).
|
Bone metastases
Bone metastases were demonstrated in 25 of 55 patients (45%) by at least one of the three bone imaging procedures, including BS, WB-MRI, and PET. The three procedures were concordant in 47 patients (showing bone metastases in 17 patients and none in 30). BS revealed lesions in 22 of 25 patients and WB-MRI in also 22 patients, but in four of these 22 patients, either BS or WB-MRI was normal. PET demonstrated foci of abnormal bone uptake in 19 of 25 patients. There was no statistical difference in sensitivity among BS, WB-MRI, and PET in this per-patient analysis. BS was the only abnormal technique in one patient demonstrating four lesions in ribs and one in the skull; WB-MRI was the only abnormal procedure in three patients, revealing metastases in the spine and pelvic bones and PET in none.
A total of 308 bone lesions were as found with the combination of these three imaging modalities. BS detected 67% of these lesions, WB-MRI 51%, and PET 30% [median SUV maximum 3 (range 1.5–6)]. The combination of BS and WB-MRI detected all patients with bone metastases and 94% of bone lesions. Four patients had diffuse lesions, whereas the other 21 had one up to 10 lesions. BS was the most efficient technique (P = 0.002 vs. WB-MRI and P < 0.0001 vs. PET). As shown in Table 2, BS was by far the most efficient method to detect lesions with a low mechanical risk located in ribs, sternum, clavicles, and skull (P < 0.001 vs. WB-MRI and PET). BS and WB-MRI detected a similar number of high mechanical risk lesions located in spine, pelvic bones, and long bones and were more efficient than PET (BS or WB-MRI vs. PET: P < 0.01). However, a large number of high mechanical risk lesions were detected only by either BS or WB-MRI, and the combination of these two tools detected significantly more lesions than either procedure alone. Two patients had only low mechanical risk lesions seen on BS or WB-MRI, and all the other patients had spinal, pelvic bones, or femoral lesions.
|
Among the 45 patients with imaged tumor foci, no abnormal FDG uptake was found in 15 patients, and in the other 30 patients, SUV maximum was relatively low: it was equal or less than 2 in six patients, ranged from 3 to 4 in 16 patients, and ranged from 5 to 6 in the other eight. Median SUV (range) was 4 (0 to 6) in patients with progressive disease and 0 (0 to 3) in those with stable disease. The difference was highly significant (P = 0.0009), but there was a large overlap between the two groups of patients. The disease was progressive in eight (38%) of the 21 patients with a SUV equal or lower than 2 and in 18 (75%) of the 24 patients with a SUV higher than 2.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
It clearly appears that FDG PET cannot be used alone for imaging MTC patients because of its low sensitivity. FDG-PET was positive in only 58% of patients in accordance with the 62% previously reported (16). This is mainly due to the low FDG uptake (SUV maximum ranged from 1 to 7) that may be related to the slow progression rate of most MTCs, the nature of MTC lesions (often sclerotic, necrotic or calcified), and their small size. In fact, lung and liver involvement was miliary in many patients and was best detected with spiral CT and MRI, in accordance with previous studies (13, 14, 18, 22). Other studies have reported more promising results of FDG PET, mainly in neck and mediastinum metastases, that may be related to an insufficient sensitivity of the other techniques to which it was compared (12, 13, 15, 17).
Also, SUV cannot be used as a reliable prognostic indicator. Indeed, SUV maximum was lower in patients with stable disease than in those with progressive disease, but there was a large overlap between these two groups of patients.
Neck US is the most sensitive tool to detect lymph node metastases in patients with papillary thyroid carcinoma (38), and according to our data, this also applies to MTC patients. Chest CT can detect lung micronodules as small as 2 mm, and because lung MTC metastases are mainly miliary, CT is superior on any other technique to detect lung lesions, as previously shown (14). Similarly, mediastinum lymph nodes may be benign, but a significant FDG uptake may provide additional information concerning their metastatic origin (28, 39). However, in MTC patients the low accuracy of CT for diagnosis of mediastinum involvement is not improved by FDG-PET because of low FDG uptake. Finally, the association of lung micronodules and mediastinum lymph nodes reinforces the suspicion of tumor involvement. Liver MRI with standardized procedures is the most sensitive imaging modality, as already shown for other endocrine tumors (26). Finally, MRI was found to be the most sensitive tool to detect bone involvement in MTC patients (21). In the present study, bone scintigraphy and WB-MRI were complementary. As previously reported, WB-MRI missed peripheral lesions located in the ribs, skull, and sternum (40, 41). In one of the two patients with only peripheral lesions, bone metastases were seen on bone scintigraphy and not on MRI. This was due to image distortion in peripheral sites. In the other patients with bone metastases, at least one axial site (spine, pelvic bones, and femur) was involved, and a focused MRI on these areas would be sufficient to complement bone scintigraphy (21, 41).
Finally, 10 patients had no demonstrable tumor foci on any imaging modality. These patients had a calcitonin level lower than 816 pg/ml and a CEA level lower than 33 ng/ml, in accordance with previous studies (4, 5, 6, 7). These findings stress the high sensitivity of calcitonin determination (3). Also, we probably underestimated liver involvement because we did not perform any laparoscopy or hepatic angiography, but these techniques are invasive (24, 25). As an alternative, the use of contrast agent injection may improve the sensitivity of liver US because MTC liver metastases are usually highly vascularized (42).
In conclusion, imaging MTC would consist in the combination of neck ultrasonography, chest spiral CT, liver MRI with standardized procedures, and both bone scintigraphy and axial bone MRI (Fig. 3
). Using these imaging modalities we were able to detect 98% of neck recurrences; 100% of mediastinal lymph nodes, lungs, and liver metastases; and 94% of bone metastases. FDG PET has no place in routine imaging of medullary thyroid cancer, and its prognostic value is low.
|
Footnotes |
|---|
Disclosure Statement: The authors indicated no potential conflict of interest.
First Published Online August 28, 2007
Abbreviations: BS, Bone scintigraphy; CEA, carcinoembryonic antigen; CT, computed tomography; FDG, 2-[fluorine-18]fluoro-2-deoxy-D-glucose; MRI, magnetic resonance imaging; MTC, medullary thyroid carcinoma; PET, positron emission tomography; SUV, standardized uptake value; US, ultrasonography; WB, whole body.
Received June 1, 2007.
Accepted August 22, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsMethodsResultsDiscussionReferences |
|---|
lu L, Güler N, Caner B, Uysal U, Elahi N, Halilolu M, Yüksel D, Aras T, Bayhan H, Bekdik C 1996 Comparison of 99mTc(V)-DMSA, 201Tl and 99mTc-MIBI imaging in the follow-up of patients with medullary carcinoma of the thyroid. Eur J Nucl Med 23:1367–1371[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  S. Leboulleux, C. Dromain, A. L. Vataire, D. Malka, A. Auperin, J. Lumbroso, P. Duvillard, D. Elias, D. M. Hartl, T. De Baere, et al. Prediction and Diagnosis of Bone Metastases in Well-Differentiated Gastro-Entero-Pancreatic Endocrine Cancer: A Prospective Comparison of Whole Body Magnetic Resonance Imaging and Somatostatin Receptor Scintigraphy J. Clin. Endocrinol. Metab., August 1, 2008; 93(8): 3021 - 3028. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Laure Giraudet, A. Al Ghulzan, A. Auperin, S. Leboulleux, A. Chehboun, F. Troalen, C. Dromain, J. Lumbroso, E. Baudin, and M. Schlumberger Progression of medullary thyroid carcinoma: assessment with calcitonin and carcinoembryonic antigen doubling times Eur. J. Endocrinol., February 1, 2008; 158(2): 239 - 246. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
