Association between Refractoriness to 131I Therapy for Differentiated Thyroid Carcinoma and 18F-FDG Accumulation in Lung Metastasis

Background: The purpose of this study was to retrospectively investigate the association between 2-[F-18]-fluoro-2-deoxy-D-glucose (18F-FDG) accumulation in lung metastasis (LM) before 131I therapy and refractoriness to 131I therapy for differentiated thyroid carcinoma (DTC) patients. Methods and Findings: Sixty-one DTC patients with LM who underwent Positron emission tomography/computed tomography using 18F-FDG (18F-FDG PET/ CT) before an initial 131I therapy were retrospectively evaluated. Maximum of standardized uptake value (SUVmax) in LM with the highest 18F-FDG accumulation was measured in each patient. The SUVmax was compared between patients with and without 131I-positive LM, and between patients with and without an increased level of thyroglobulin (tumor marker) 12 ± 2 months after 131I therapy using the Wilcoxon test. Discussion: Predictability for the patients with an increased thyroglobulin level was also analyzed by receiver-operating-characteristic (ROC) analysis. SUVmax of LM was significantly greater for patients without 131I-positive LM than for those with 131I-positive LM (5.9 ± 6.0 vs. 1.9 ± 2.0, p<0.01) and was significantly greater for patients with an increased level of TG after 131I therapy than for those without (7.0 ± 4.9 vs. 1.2 ± 1.0, p<0.01). All 11 of the 49 patients with SUVmax > 3.8 showed an increased TG level after 131I therapy. Use of the optimal cutoff threshold for SUVmax of 1.6 differentiated patients with an increased level of TG from those without at a sensitivity of 74.2%, a specificity of 94.4%, an accuracy of 81.6% and an AUC of 0.91. Conclusion: 18F-FDG accumulation in LM from DTC can be one of predictors for refractoriness to 131I therapy.


Introduction
The most common disease of malignant endocrine tumors is thyroid cancer, which is still increasing in incidence [1,2]. Differentiated thyroid carcinoma (DTC) including papillary and follicular thyroid carcinoma accounts for more than 90% of all thyroid cancers. DTC generally has a favorable prognosis, with 10-year overall and cause-specific survival rates of 76.8% and 84.9%, respectively [3]. However, in DTC patients with distant without 131 I accumulation have a poor prognosis [9,10]. However, according to some recent articles, molecular-targeted therapy with sorafenib significantly improved progression-free survival in patients with progressive radioactive iodine-refractory DTC [11,12]. Therefore, establishment of the indications for the molecular-targeted therapy would be of great clinical benefit, especially in DTC patients with distant metastasis.
Positron emission tomography/computed tomography using 2-[F-18]-fluoro-2-deoxy-D-glucose ( 18 F-FDG PET/CT) is known to be a useful modality for the detection of iodine-negative DTC lesions [10,13,14], since Feine et al. first formally proposed the "flip-flop pattern", an inverse correlation between 18 F-FDG and 131 I uptake in the metastatic lesions from DTC patients [15]. As the mechanism, it has been reported that DTC cells show glucose transporter 1 (GLUT1) upregulation and reduced expression of the sodium-iodide symporter during the dedifferentiation process [16]. Although it has been reported that residual lymph node metastasis from DTC showed significantly higher 18 F-FDG uptake in lesions without 131 I uptake than those with 131 I uptake [17,18], among DTC patients with lung metastasis (LM), the relationship between the resistance to 131 I therapy and 18 F-FDG accumulation in LM remains unclear. The purpose of this study was to retrospectively clarify the association between 18 F-FDG accumulation in LM before 131 I therapy and resistance to 131 I therapy among DTC patients.

Patients
This retrospective study was approved by our institutional review board, and the written informed consent was from all patients.
Two-hundred sixty-three consecutive patients with DTC who were treated with 131 I therapy between October 2012 and September 2016 at Kyushu University Hospital after near-total or total thyroidectomy were retrospectively analyzed. DTC patients with LM histopathologically diagnosed as either papillary or follicular carcinoma were included. The definition of LM was determined by at least one of the following criteria: (1) 131 I accumulation in the lung field higher than the surrounding tissue identified on 131 I SPECT/CT (32 patients), and (2) multiple pulmonary nodules in the bilateral lung, which showed progressive increase in size on follow-up CT (observation period 35 ± 9 months: 29 patients). Patients who had a past history of any other malignant disease, who had distant metastasis in organs other than the lung, who had a low thyroid-stimulating hormone (TSH) level (< 30 U/mL), or who had a high blood glucose level (> 150 mg/dL) were excluded from this study. Consequently, a total of 61 patients (41 females and 20 males) were included in our study. All patients underwent thyroid hormone withdrawal for at least 4 weeks before 131 I therapy for the purpose of TSH stimulation, and all patients were prescribed a low-iodine diet for 2 weeks in preparation for 131 I administration. The patient characteristics are indicated in Table 1.
All patients underwent a whole-body 131 I scan (WBS) and SPECT/ CT with a hybrid camera combining a dual-head c-camera with a 6-slice spiral CT within the same gantry (Symbia T6: Siemens, Hoffman Estates, IL). On WBS, anterior images were acquired at a speed of 10 cm/min with high-energy parallel-hole collimators, a 256 × 1024 matrix, and a 364-keV photopeak with 15% windows. SPECT images were acquired in a step-and-shoot mode, with 40 projections (a duration of 45 s at each projection), a noncircular orbit over 360°, high-energy parallel-hole collimators, a 128 × 128 matrix, and a 364-keV photopeak with 15% windows. Then, 3D ordered-subset expectation-maximization iterative reconstruction was performed, with 4 iterations and 8 subsets. SPECT images were subjected to CT-based attenuation correction without scattered correction. The CT scan parameters were 130 keV, 30 mAs or less (for minimization of radiation exposure), a 512 × 512 matrix, and a 2 × 2.5 mm collimation.

F-FDG PET/CT
Each 18 F-FDG PET/CT acquisition was performed under the TSH-stimulated state. In each patient, 185 MBq of 18 F-FDG was intravenously administered after at least 4 hr of fasting. Scans were conducted from the middle of the thigh to the top of the skull 60 min after 18 F-FDG administration. 18 F-FDG PET/ CT images were obtained using an integrated PET/CT scanner Discovery STE (GE Medical Systems, Milwaukee, WI). The PET scanner comprises 24 ring detectors consisting of 560 BGO crystals (4.7 × 6.3 × 30 mm). All emission scans were performed in 3-dimensional mode with 128 × 128 matrices (5.47 × 5.47 × 3.27 mm), and the acquisition time per bed position was 3 min. The PET images were reconstructed using the ordered-subset expectation-maximization method (VUE Point Plus) with 2 full iterations of 28 subsets, and the full-width at half maximum was 5.2 mm. A low-dose 16-slice CT (tube voltage, 120 kV; effective tube current, 30-250 mA) from the vertex to the proximal thigh was performed for attenuation correction, and for determining the precise anatomic location before acquisition of the PET image. The CT scan was reconstructed by filtered back-projection into 512 × 512 pixel images with a slice thickness of 5 mm to match the PET scan. The PET/CT fusion images were made by GENIE-Xeleris software using a dedicated work station, Xeleris (GE Medical Systems, Milwaukee, WI).

Diagnostic CT protocol
A diagnostic chest CT covering the upper mediastinum to the upper abdomen was performed with a 64-MDCT (multi detector-

Archives in Cancer Research
ISSN 2254-6081 row CT) scanner (Aquilion 64; Toshiba Medical Systems, Tokyo, Japan) after 131 I scintigraphy, using the following parameters: tube voltage 120 kV, effective tube current 300 mA, collimation 0.5 mm, pitch 27.0. The MDCT scan was reconstructed by filtered back projection into 512 × 512 pixel images with a slice thickness of 3 mm. In all patients, 18 F-FDG PET/CT, diagnostic CT, and 131 I scintigraphy were performed within one week.

Data Analysis
All patients in this study underwent patient-based analysis by LM. After determination of the LM with highest 18 F-FDG accumulation in PET images, the highest pixel value in the LM on 18 F-FDG PET/ CT was determined as the maximum standardized uptake value (SUV max ). On 131 I scintigraphy, 131 I accumulation higher than the background in at least one LM was defined as 131 I-positive LM and that as low as background in all LMs as 131 I-negative LM by visual evaluation. SUV max was compared between patients with and without 131 I-positive LMs, and between those with and without an increased thyroglobulin (TG) level 12 ± 2 months after initial 131 I therapy. The predictability for 131 I accumulation in LM or an increased level of TG after 131 I therapy and the correlation of SUV max in LM with a TG level before or after 131 I therapy were also analyzed.

Statistical Analysis
Comparisons of SUV max between patients with and without 131 I-positive LM, and between those with and without an increased level of TG after 131 I therapy were analyzed by the Wilcoxon test. Analysis of the predictability for 131 I accumulation in LM or an increased TG level after 131 I therapy was performed by receiveroperating-characteristic (ROC) analysis. The correlation of SUV max in LM with TG level before or after 131 I therapy was analyzed by Pearson's correlation analysis. The tests were performed using JMP® (version 9.0.2; SAS Institute, Cary, North Carolina) statistical software. A p value less than 0.05 was considered statistically significant.

Comparison of 18 FDG accumulation between patients with 131 I-positive and 131 I-negative LM
The SUV max in LMs of all 61 patients ranged from 0.5 to 26 (mean ± SD; 3.8 ± 4.8). In 3 of 61 patients, LMs were detected only with 131 I scintigraphy, not with CT images. For the other 58 patients, the greatest short-axis diameters in the largest LM nodules ranged from 3 to 21 mm (mean ± SD; 7 ± 4 mm). Of the 61 patients, 32 patients had 131 I-positive LM and 29 had no 31 I-positive LM.

Comparison of 18 FDG accumulation between patients with and without an increased level of TG after 131 I therapy
Forty-nine of 61 patients without a high level of anti-TG antibody (> 45 IU/mL) were analyzed. The TG levels before and after 131 I therapy of the 49 patients ranged from 15.2 to 2590 (mean ± SD; 461 ± 531) and from 0.9 to 1850 (mean ± SD; 379 ± 450), respectively. Of the 49 patients, 18 patients had an increased TG level and 31 patients did not. 18 F-FDG accumulation in LM as indicated by SUV max was significantly higher in patients whose TG levels increased than in those whose TG levels didn't increase [7.0 ± 4.9 vs. 1.2 ± 1.0 (mean ± SD), p<0.01] shown in Figure 2. All 11 patients with SUV max greater than 3.8 showed an increased TG level after 131 I therapy. I-positive LM consisted of 4 patients with an increased TG level and 23 patients without, and there was no significant difference between patients with and without an increased TG level (3.0 ± 2.1 vs. 1.1 ± 0.9, p=0.08). On the other hand, of the 22 patients without 131 I-positive LM, 14 patients whose TG levels increased showed significantly higher SUV max than 8 patients whose TG

Correlation of 18 F-FDG accumulation in LM with the TG levels before or after 131 I therapy
Among the 49 patients, there was no significant correlation between the SUV max in LM and the TG levels before 131 I therapy (r =0. 21, p=0.14). On the other hand, the TG levels after 131 I therapy Figure 2 Box-and-whisker plot of SUV max in LM in patients with and without an increased level of TG after 131 I therapy.
The boxes represent the 25-75% range with bisecting lines showing the median value, and the horizontal lines represent the 10-90% range. Of 49 DTC patients without a high level of anti-TG antibody (upper left), SUV max of LM was significantly greater for patients whose TG levels after 131 I therapy did not increase than for those whose levels did increase. Of 27 patients with 131 I-positive LM (lower left), there was no significant difference in SUV max of LM between patients with and without an increased level of TG after 131 I therapy. Of 22 patients without 131 I-positive LM (lower right), SUV max of LM was significantly greater for patients whose TG levels after 131 I therapy did not increase than for those whose levels did increase.

Archives in Cancer Research ISSN 2254-6081
was significantly higher than those without. Therefore, we hypothesize that 131 I-negative LM includes two types of clinically aggressive and unaggressive features. For DTC patients with an aggressive type of LM, alternative treatment such as moleculartargeted therapy is needed instead of 131 I therapy. It is proposed that 18 F-FDG -PET/CT is effective in the selection of DTC patients with the aggressive type of 131 I-negative LM and can determine the indications for molecular-targeted therapy, which has a beneficial effect on tumor progression in patients with radioactive iodine-refractory metastatic DTC.
Additionally, SUV max of LM in DTC patients before 131 I therapy had a significant correlation not with the TG levels before 131 I therapy, but with that after 131 I therapy. These data show that 18 F-FDG accumulation in LM before 131 I therapy expresses post-therapeutic resistance to 131 I therapy, although, interestingly, it had no significant relation to the progression of DTC before 131 I therapy.
Recently, an association has been found histopathologically between high expression of GLUT1 in thyroid cancer stem cells showed a significant correlation with the SUV max in LM (r =0.69, p<0.01) represented by Figure 5.

Discussion
Our results demonstrated that SUV max of LM on 18 F-FDG PET/CT had a high predictive value for 131 I accumulation in LM from DTC and the response to 131 I therapy. That suggests that high 18 F-FDG accumulation in LM is associated with a shift to dedifferentiation of LM and poor clinical outcome in DTC patients with LM. Hong et al previously reported that SUV max greater than 3.6 in distant metastases from DTC was significantly predictive of reduced disease-specific survival in multivariate analysis [19]. All of the patients with SUV max greater than 3.8 for LM in our study showed an increased TG level after 131 I therapy. The previous article and our study indicate that SUV max of LM in DTC patients is an important index to predict a therapeutic effect for 131 I therapy.
Especially in patients without 131 I-positive LM, SUV max of LM in DTC patients with an increased level of TG after 131 I therapy

Figure 5
Scatterplots of SUV max in LM and pre-therapeutic TG levels (left) and of SUV max in LM and post-therapeutic TG levels (right) in the 49 patients without a high level of anti-thyroglobulin antibody. Best-fit lines are shown. There was no significant correlation between SUV max in LM and the TG level before 131 I therapy (r = 0.21, p = 0.14). The TG level 12 ± 2 months after 131 I therapy had a significant correlation with SUV max in LM (r = 0.69, p<0.01).
and high resistance to 131 I therapy [20]. Therefore, 18 F-FDG PET/ CT can be a helpful tool to assess the potential of DTC patients with LM to benefit from 131 I therapy and can contribute to the clinical management and determination of the best therapeutic strategy post-resection.
In our study, 18 F-FDG PET/CT was performed on the condition of sufficient TSH stimulation, because 18 F-FDG PET under TSH stimulation improves the detection of DTC metastases [21]. Moreover, our judgment of 131 I accumulation in LM was performed using 131 I SPECT/CT as post-therapy 131 I scintigraphy in all patients based on the findings of recent articles, reporting that SPECT/ CT improves detection and localization of 131 I accumulation in distant metastases in comparison with whole-body scintigraphy [22,23]. Thus, the present study has shown that the evaluation of 18 F-FDG and 131 I accumulation in LM from DTC is feasible and effective.
Our study had several limitations. First, diagnoses of LM were not always made using a histopathologic procedure but were also sometimes made at clinical follow-up. Although 131 I-avid LM uptake usually indicates a metastatic lesion from DTC, 131 I-nonavid LM judged at the time of clinical follow-up might be revealed not to be a metastatic lesion at a later date. Because DTC is a slowgrowing neoplasm, a long follow-up period is needed to make a more accurate clinical diagnosis. Second, we lack sufficient data to prove that the outcomes of patients without increased TG level were better than those of patients with increased TG level.
Further studies are needed to evaluate the outcomes of DTC patients, who generally having long disease courses.

Conclusion
In conclusion, 18 F-FDG accumulation in LM was related to the lack of 131 I accumulation and also associated with poor response to 131 I therapy especially in patients without 131 I accumulation. High

18
F-FDG accumulation in LM from DTC was associated with poor treatment outcome after 131 I therapy. By contrast, even though low 18 F-FDG accumulation in the LM indicates the poor likelihood of 131 I accumulation, it also indicates a more stable status of DTC tumor activity. 18 F-FDG PET/CT can predict refractoriness to 131 I therapy and determine the indication of molecular-targeted therapy in DTC patients with LM.