Abstract
Background:
Patients with differentiated thyroid cancer (DTC) who have a suspicious recurrent or persistent disease based on an elevated serum thyroglobulin (Tg) or Tg antibodies (TgAb) are usually referred for empiric radioiodine (131I) administration to localize and treat the disease. The aim of this retrospective monocentric study was to assess the sensitivity of postempiric 131I whole-body scan (WBS) compared to 18-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) in such patients who had an initial normal postablation WBS.
Methods:
Among 47 consecutive patients with DTC who had a normal postablation WBS and were referred for empiric 131I administration, 34 patients (12M, 22F; mean age 53 years) underwent FDG PET/CT and form the basis of this report: 23 patients had persistently elevated serum Tg levels, 10 had elevated Tg levels observed during follow-up after they initially became under 1 ng/mL, and 1 had appearance of TgAb during follow-up. Postempiric 131I WBS and FDG PET/CT were analyzed by independent readers.
Results:
A total of 75 lesions were found in 23 patients, distributed in 36 organs. Lesions were located in the neck (30), lungs (28), mediastinum (11), and bones (6). The sensitivities for the detection of individual lesions and for the diagnosis of metastatic organs were 88% and 97% for PET/CT and 16% and 22% for WBS, respectively (p<0.01). PET/CT was abnormal in 22 patients, among which 5 also had an abnormal postempiric 131I WBS. There was only one patient with an abnormal postempiric 131I WBS and a normal FDG PET/CT. This patient underwent two further 131I administrations, with the last WBS being normal and the last stimulated Tg level being undetectable. Other patients were either treated with surgery, or classified as radioactive iodine refractory and treated with levothyroxine suppressive therapy or tyrosine kinase inhibitors.
Conclusion:
In patients with suspicious recurrence based on the Tg level after a normal postablation WBS, FDG PET/CT is the preferred scintigraphic method to localize disease rather than postempiric 131I WBS. Empiric 131I administration may be used only in patients who do not have a significant FDG uptake.
Introduction
Earlier studies have shown that in such patients, a whole-body scan (WBS) performed with an empiric high activity of 131I (3.7–7.4 GBq; 100–200 mCi) may localize disease that was not detected on a WBS performed with a lower diagnostic activity of 131I (0.074–0.37 GBq; 2–5 mCi), raising the concept for empiric 131I treatment (2 –4). At the time, ablation WBS was not routinely performed or had a low sensitivity due to the presence of large thyroid remnants following a less than total thyroidectomy. Diagnostic WBS was also less sensitive because the gamma-cameras were not adapted for the detection of the 131I gamma emission. Furthermore, neck ultrasonography (US) and neck and chest computed tomography (CT) were not routinely performed, and 18-fluorodesoxyglucose (FDG) positron emission tomography (PET) was not available.
FDG PET has initially been used in thyroid cancer patients with noniodine-avid disease with a sensitivity ranging from 50% to 100% (5,6). FDG PET is also known to be more sensitive in patients with aggressive tumors, which are usually 131I nonavid (7 –9). A recent study showed the absence of 131I uptake after empiric 131I in 14 patients with an elevated Tg level, normal neck US, and normal FDG PET (10). Whether FDG PET is more sensitive than postempiric 131I WBS has not been clearly evaluated. The aim of this single-center retrospective study was to compare empiric 131I WBS and FDG PET/CT in patients with normal postablation WBS and elevated Tg or persistent TgAb.
Methods
Patients
Approval from our Institutional Review Board was obtained for the study. Files of consecutive patients treated with 131I between March 2003 and December 2008 were reviewed. Inclusion criteria were (i) patients with DTC treated with total thyroidectomy and postoperative 131I ablation; (ii) confirmed pathological diagnosis of malignancy by our pathologist (A.A.G.); (iii) normal postablation WBS defined by the absence of nonphysiological iodine uptake outside the thyroid bed; (iv) increasing or persistent elevated serum Tg levels or rise in the serum TgAb titers; (v) empiric 131I administration given in the absence of iodine contamination; and (vi) FDG PET/CT performed within 4 months of empiric 131I treatment.
Radioiodine (131I) empiric administration
Patients were treated following thyroid hormone withdrawal (THW) with an 131I activity of 3.7 GBq (100 mCi) with the serum TSH level above 30 mUI/L. A WBS was performed 3–4 days after 131I administration, using a dual-head gamma camera equipped with high-energy collimators and thick crystals.
FDG PET/CT
Imaging and data acquisitions were performed on an integrated PET/CT Biograph LSO system (Siemens Medical Solutions, Erlangen, Germany). PET/CT scanning was performed after an i.v. injection of 4–5 MBq/kg FDG, followed by a 55–80-minute uptake phase. All patients had fasted for 6 hours. Emission data were acquired for 4 minutes at each bed position from the top of the head to the superior mediastinum with patients' arms along the chest and from the neck to the midthigh with patients' arms above the head. No specific breathing instructions were given. Three-dimensional mode was used for PET image acquisition. PET data were reconstructed on a 128×128 matrix, using an iterative algorithm (FORE and AWOSEM) with two iterations, eight subsets, and a 5-mm FWHM gaussian postfilter. Reconstruction data were acquired with a single slice spiral CT (Somatom Emotion; Siemens Medical Solutions) without an intravenous contrast agent. CT parameters were set to 80 mA and 110 kV, slice thickness of 5 mm, and pitch 1.5. CT data were reconstructed using filtered back projection with a smooth filter on a 512×512 matrix.
Neck US
US was performed on the day of empiric 131I administration with a high-resolution ultrasound system (Aplio ultrasound machine; Toshiba Medical, Puteaux, France) equipped with a high-energy 14-MHz linear probe (PZT; Toshiba), allowing to work in fundamental B-mode (lateral resolution: 0.17 mm; axial resolution: 0.11 mm) and in power Doppler mode (rate of 12 frames/seconds, limit detection of 5 cm/seconds with a pulse repetition frequency of 17 kHz). US examination included the thyroid bed and both central and lateral neck compartments. Neck US was considered abnormal in case of abnormal findings in the thyroid bed or in the case of an abnormal lymph node, based on the following criteria: hyperechoic punctuations, cystic appearance, hypervascularization, a round shape node without hyperechoic hilum, and a short axis >7 mm (11).
Contrast-enhanced neck and chest CT
Neck and chest CT examinations were performed with a Hispeed spiral scanner (GE Medical Systems, Milwaukee, WI). Spiral CT images were obtained before a monophasic injection of 100 mL of monoionic contrast material (Xenetix 300; Guerbet, Roissy, France). Scanning was performed at 120 kV and 270 mA. Contiguously reconstructed sections (pitch of 1:1) were obtained with 5-mm collimation. Contrast-enhanced CT was performed either 8 weeks before or 4 weeks after radioiodine administration.
Image analysis
WBS, FDG PET/CT, and neck and chest CT scan were reviewed, blindly and independently. WBS was classified as abnormal in case of nonphysiological iodine uptake. FDG PET/CT lesions were graded on a scale of 1–4 (1, probably benign; 2, equivocal; 3, probably malignant; and 4, definitely malignant). Only lesions with scores of 3 and 4 were considered to be abnormal findings and called lesions; they consisted on either lesions with FDG uptake or lesions without FDG uptake observed on the CT scan.
The number of involved organs and distinct lesions visualized by WBS, FDG PET/CT, neck US, and neck and chest contrast-enhanced CT were recorded. Central, ipsilateral, and contralateral neck lymph nodes, lung, mediastinum, liver, spine, pelvic bones, ribs, and lower–upper limbs were considered as distinct organs. The total number of lesions was determined by summing the highest number of lesions detected in each organ by either examination with WBS, FDG PET/CT, neck US, or CT scan.
The sensitivities for disease detection of WBS and FDG PET/CT were compared per patient, per involved organ, and per number of detected lesions. The sensitivities for the detection of neck and chest disease were compared for all imaging modalities.
Tg level measurements
From 2003 until 2005, serum Tg was measured using an immunoradiometric assay (SELco® Tg; Medipan Diagnostica, Selchow, Germany). The analytical sensitivity was 0.3 ng/mL, and the Tg level was considered as not accurately measured when the routine recovery test (performed in all serum samples) was <80%. Since 2006, serum Tg was measured using a chemiluminescent immunoenzymatic sandwich assay (Access® Thyroglobulin, automated on UniCel® DxI 800 instruments; Beckman Coulter, Villepinte, France) with an analytical sensitivity of 0.1 ng/mL. The serum Tg level was considered not accurate in the presence of TgAbs at the Access Thyroglobulin Antibody II assay (Beckman Coulter).
Statistics
Quantitative data were expressed as mean and standard deviation and qualitative data were expressed in percentage.
Sensitivities of FDG PET/CT, WBS, neck US, and neck and chest CT, and their 95% confidence intervals (CIs) were calculated and compared using the McNemar test for matched proportions.
All reported p values are two-sided and the significance level was 0.05. Analyses were performed using SAS statistical software (version 9.1; SAS Institute Inc., Cary, NC).
Results
Patients
Records of 47 consecutive patients with DTC were reviewed. FDG PET/CT was not performed in 13 patients. Overall, 34 patients met the inclusion criteria and formed the basis of this report. Their characteristics are reported in Table 1. The histology of the thyroid cancer was papillary in 32 (94%) cases, and follicular in 2 (6%) cases, with aggressive subtypes in 8 (24%) cases (Table 1). Initial treatment consisted of a total thyroidectomy in all cases and neck lymph node dissection in 68% of the patients. Pathologic tumor–node–metastasis (pTNM) staging is detailed in Table 1 (12). Postoperative 131I treatment (3.7 GBq; 100 mCi) was given in all patients after THW. Postablation WBS showed a median residual 131I uptake in the thyroid bed of 0.4% (range: 0.06%–11%; mean: 1.4%) and did not reveal any focus of uptake outside the thyroid bed. Neck US performed at the time ablation was normal in all cases. Ablation was considered successful in 10 cases with a stimulated Tg level under 1 ng/mL in the absence of TgAb and a normal neck US. Ablation was considered incomplete in 24 cases. This was because of the elevated Tg level under levothyroxine treatment in 3 cases (median Tg level: 274 ng/mL; range: 1.5–620), because of the elevated TSH-stimulated Tg level in 20 cases (median Tg level: 9 ng/mL; range: 1.7–290), and because of an undetectable stimulated Tg level, but with TgAb, in 1 case.
Pathologic tumor–node–metastasis (pTNM) staging system (12).
For one patient, none central node dissection was performed, but a metastatic LN close to the thyroid gland was discovered during surgery, and the tumor was classified as N1.
M, male; F, female; PTC, papillary thyroid cancer; LN, lymph node; CND, cervical neck dissection.
Empiric 131I was administered for a persistent elevated serum Tg level in 23 cases, for the appearance of elevated Tg level after initial normalization during follow-up in 10 cases, and for the appearance of TgAb during follow-up in 1 case. The median interval of time between postoperative 131I ablation treatment and empiric 131I treatment was 33 months (range: 6–208 months; mean: 53). Among patients with the newly elevated Tg level, the median interval of time between postoperative 131I ablation treatment and empiric 131I treatment was 80 months (range: 39–208 months; mean: 97). The median Tg level in the absence of TgAb measured after THW on the day of the empiric 131I administration was 47 ng/mL (range: 4–3230; mean: 290).
FDG PET/CT and WBS: imaging results
FDG PET/CT was performed under LT4 treatment in 8 (24%) cases, after rhTSH stimulation in 4 (12%) cases and after thyroid hormone withdrawal on the day of the empiric 131I administration in 22 (65%) cases. The median interval of time between FDG PET/CT and empiric iodine treatment was 5 days (range: 0–120; mean: 30), and FDG PET/CT was performed before empiric iodine was administered in 25 of the cases. Median Tg levels measured after LT4 withdrawal when empiric iodine was given was 30 ng/mL (range: 4–209) in patients with normal FDG PET/CT uptake, and 86 ng/mL (range: 6–3230) in those with an abnormal FDG PET/CT.
Per patients analysis
There were 23 (68%) patients in whom the WBS, FDG PET/CT, neck US, or neck and chest CT localized at least one lesion. WBS and FDG PET/CT were concordant in 16 patients, being both abnormal in 5 patients and both normal in 11 patients (Table 2). WBS and FDG PET/CT were discordant in 18 patients: FDG PET/CT was abnormal in 17 patients with normal WBS and normal in 1 patient with an abnormal WBS. The sensitivity of FDG PET/CT for disease localization was 65% [CI 49%–81%], whereas the sensitivity of WBS for disease localization was 18% [CI 5%–31%] (p<0.01).
WBS, whole-body scan; FDG PET/CT, 18-fluorodesoxyglucose positron emission tomography/computed tomography; NA, not applicable.
Per organ analysis
The total number of metastatic organs was 36 in these 23 patients (Table 2). FDG PET/CT detected lesions in 35 organs and WBS in 8 organs. The sensitivity of FDG PET/CT for the diagnosis of metastatic organ was 97% [CI 92%–100%] and the sensitivity of WBS was 22% [CI 9%–36%], the difference being statistically significant (p<0.01).
Per-lesion analysis
The total number of lesions detected was 75 in these 23 patients (Tables 2 and 3). There were 30 lesions detected in the neck, 28 in the lungs, 11 in the mediastinum, and 6 in the bones. FDG PET/CT detected 66 of these lesions, among which 7 also disclosed 131I uptake. The WBS detected 12 lesions, among which 5 did not disclose any FDG uptake. The sensitivity of FDG PET/CT for the diagnosis of lesions was 88% [CI 81%–95%] and the sensitivity of WBS was 16% [CI 8%–24%], the difference being statistically significant (p<0.01). Of note, 16 of the 28 lung lesions did not show any FDG uptake and were diagnosed on the CT part of the FDG PET/CT.
No., number; US, ultrasonography.
There were only five lesions with 131I uptake that did not show any FDG uptake consisting of four rib lesions in one patient and one lung lesion in one patient who had other lung lesions with FDG uptake.
Among the five patients with both abnormal WBS and FDG PET/CT, the number of lesions detected on the FDG PET/CT was similar to the number of lesions detected on the WBS in two cases, lower in one case, and higher in two cases (Table 4).
PTC, papillary thyroid cancer; Tg, thyroglobulin; TKI, tyrosine kinase inhibitors; EBR, external beam radiation; FTC, follicular thyroid cancer.
Neck US and neck and chest CT: imaging results
Neck US detected 18 of the 30 lesions located in the neck, neck CT detected 14 lesions, and FDG PET/CT detected 27 neck lesions (Table 3). The sensitivities for the localization of local recurrence were 60% [CI 43%–73%] for neck US, 47% [CI 29%–65%] for neck CT, and 90% [CI 79%–100%] for FDG PET/CT (neck US vs. FDG PET/CT: p=0.02; neck CT vs. FDG PET/CT: p<0.01). There were three neck lesions detected with neck US that did not show any FDG uptake.
Chest CT detected 24 of the 28 lung (86%) lesions, 8 of the 11 (73%) mediastinum lymph nodes, the spine lesion, and 1 of the 5 (20%) rib lesions. The chest CT scan did not detect any lesion that was not detected with the FDG PET/CT, neck US, or WBS.
Follow-up
Following empiric 131I administration and FDG PET/CT, nine patients underwent neck surgery for a local recurrence. After a median follow-up of 2.7 years (range: 0–8.2; mean: 2.8), 3 patients died from tumor progression, 12 patients had metastatic disease; the other 19 patients had normal imaging with a detectable Tg level (n=13), or with undetectable Tg level in the absence of TgAb under LT4 treatment (n=3), or with undetectable Tg level and elevated TgAb (n=1), and with pending postoperative Tg level (n=2). All three patients who died during follow-up had abnormal FDG PET/CT at the time of the study.
The clinical characteristics of the patients with abnormal postempiric 131I WBS are reported in Table 4. The patient with 131I uptake in the ribs and a normal FDG PET/CT underwent two other 131I treatments, with last post-therapeutic WBS being normal, and the last stimulated serum Tg level being undetectable.
Discussion
Most patients with DTC are cured after initial treatment, and empiric 131I administration is rarely indicated. Cases where persistent and/or recurrent disease is suspected for an increasing serum Tg level or for a persistent detectable Tg level account for <20% of the patients after 131I ablation, even using ultrasensitive Tg assays (13 –16). In such cases, the Tg level will most frequently decrease spontaneously and recurrent disease is found in no more than one-third of the cases (14,16 –18). Our results are concordant with the rarity of the situation with only 47 patients referred to our center for empiric 131I administration during a 5-year period of time.
The limits of this study include its retrospective design, a low number of patients, and the absence of pathology for all lesions. Due to this retrospective design, the major bias of this study is the selection of older patients with fairly aggressive histologies (about one-fourth of the patients); these are known to have low radioiodine uptake and high FDG uptake (8,9,19).
However, our results clearly show that in patients with normal postablation WBS, the WBS performed after the administration of an empiric 131I given for an elevated Tg level is not sensitive enough to localize recurrent and/or persistent disease, and its sensitivity is only 18% compared to 65% for FDG PET/CT. We only found one patient with a normal FDG PET/CT who had an abnormal postempiric 131I WBS showing rib metastases. Since all patients had normal postablation WBS with a median thyroid bed uptake of 0.4% and postempiric 131I WBS that did not demonstrate any significant uptake in the neck, the absence of systematic single-photon emission computed tomography (SPECT)/CT was not, to our point of view, a limitation in the study (20,21).
The FDG PET/CTs were performed in three-fourths of the cases under optimal conditions, that is after TSH stimulation, and its sensitivity was in the range of the sensitivity usually reported in patients with DTC (5,6,22,23). As previously reported, we found that Tg levels that are related to tumor burden, were higher in patients with abnormal FDG PET/CT compared to patients with normal FDG PET/CT. Recommendations to perform FDG PET/CT usually include a Tg level above 10 ng/mL, a threshold that may be lower in case of aggressive subtypes of thyroid cancer or with an old age. This threshold is indeed similar to the threshold recommended to consider empiric iodine treatment.
FDG PET/CT not only localized the disease in a greater number of patients, it also detected many more lesions than postempiric 131I WBS, with a sensitivity of 88% for FDG PET/CT and only 16% for postempiric 131I WBS. As expected, the lesions were located in the neck, lung, and bones (24). Surprisingly, the sensitivity of neck US for the diagnosis of neck recurrence was lower in this study than usually reported (25,26). This might be partially explained by the fact that all patients enrolled in the study had a normal neck US at the time of ablation. Furthermore, postablation WBS is currently highly informative because of postoperative small thyroid remnants following total thyroidectomy performed nowadays. Interestingly, we did not show an added value of the diagnostic chest CT scan to the FDG PET/CT, even though the chest CT portion of the FDG PET/CT is performed while the patients is breathing normally, whereas diagnostic chest CT is usually performed after a full inspiration and is usually considered more sensitive for the diagnostic of small lung lesions (27).
Among the six patients with abnormal postempiric 131I WBS, five also had abnormal FDG PET/CT. These patients disclosed lesions both with and without 131I uptake and can therefore be considered as refractory to 131I treatment (28). Four of them have been treated with tyrosine kinase inhibitors, and the remaining one is under TSH suppression therapy and is followed with regular scans with a slow progression of the metastases. One patient with 131I uptake and no FDG uptake received two additional 131I administrations, with the last WBS being normal and the last stimulated Tg level being undetectable. Among the 28 patients with normal postempiric 131I WBS, 17 (61%) had abnormal FDG uptake, and treatment consisted in surgery when feasible, levothyroxine suppressive therapy, or tyrosine kinase inhibitors. None received any other administration of 131I. We therefore did not find any benefit in giving empiric 131I in patients with FDG uptake, because the lesions were either surgically removed or because of the coexistence of lesions with and without 131I uptake, making these patients refractory to 131I treatment.
The results of this study are in contrast with the high usefulness of empiric radioiodine administration that were reported 25 years ago. This may be attributed to a different selection of patients, to a more extensive surgical treatment that permitted the postsurgical 131I WBS to be highly sensitive, and to more accurate performances of gamma cameras (2 –4).
In conclusion, in patients with a suspicion of residual disease based on the serum Tg level after a total thyroidectomy and a normal postablation WBS, the preferred scintigraphic method to localize disease should be a FDG PET/CT rather than a postempiric 131I WBS. Our results suggest a selective use of empiric 131I administrations in patients who do not show FDG uptake.
Footnotes
Acknowledgments
The authors are indebted to Catherine Martin for secretarial assistance. This research did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.
Disclosure Statement
The authors declare that no competing financial interests exist.