The impact of post-radioiodine therapy SPECT/CT on early risk stratification in differentiated thyroid cancer; a bi-institutional study

Objective SPECT/CT has numerous advantages over planar and traditional SPECT images. The aim of this study was to evaluate the role of post-radioiodine therapy SPECT/CT of patients with differentiated thyroid cancer (DTC) in early risk classification and in prediction of late prognosis. Patients and methods 323 consecutive patients were investigated after their first radioiodine treatment (1100–3700 MBq). Both whole body scan and SPECT/CT images of the head, neck, chest and abdomen regions were taken 4–6 days after radioiodine therapy. Patients were re-evaluated 9–12 months later as well as at the end of follow up (median 37 months). Results Post-radioiodine therapy SPECT/CT showed metastases in 22% of patients. Lymph node, lung and bone metastases were detected in 61, 13 and 5 patients, respectively, resulting in early reclassification of 115 cases (36%). No evidence of disease was found in 251 cases at 9–12 months after radioiodine treatment and 269 patients at the end of follow-up. To predict residual disease at the end of follow-up, the sensitivities, specificities and diagnostic accuracies of the current risk classification systems and SPECT/CT were: ATA: 77%, 47% and 53%; ETA: 70%, 62% and 64%; SPECT/CT: 61%, 88% and 83%, respectively. There was no difference between cohorts of the two institutions when data were analyzed separately. Conclusions Based on our bi-institutional experience, the accuracy of post-radioiodine SPECT/CT outweighs that of the currently used ATA and ETA risk classification systems in the prediction of long-term outcome of DTC.


INTRODUCTION
The worldwide incidence of thyroid cancer has continuously increased during the last few decades. This rise can be attributed to the increased diagnosis of occult cancers through the use of neck ultrasound and other techniques of diagnostic neck imaging [1][2][3][4][5]. In Hungary, approximately 800 new cases were registered by the National Cancer Registry and Biostatistics Center in 2014 (http://www.onkol.hu/hu/nemzeti_rakregiszter, accessed

Clinical Research Paper
in February 07, 2017). Although the use of improved techniques leads to earlier and more accurate diagnosis, it may result in overdiagnosis and overtreatment; there is an urgent need to better distinguish the high-risk patients requiring therapy from those who do not need radioiodine after surgery, or may not need treatment at al. [3].
Patients with differentiated thyroid cancer (DTC) usually have a favorable prognosis with high cure rates; however, lifelong follow-up is required as potentially curable local recurrences and distant metastases may occur even decades later [6][7][8]. The conventional and effective treatment consists of surgical management followed by radioiodine (RAI) ablation of thyroid remnants and thyroid-stimulating hormone (TSH) suppressive therapy [6,7]. Recently, the universal use of remnant ablation after surgery has been debated and mainly restricted to advanced disease [8]. However, radioiodine therapy has additional benefits, e.g. the destruction of undetected residual tumor foci and the ablation of normal thyroid tissue which facilitates the detection of recurrent disease during follow up. The information obtained through the posttherapeutic 131 I whole-body scan (WBS) or single photon emission computed tomography/computed tomography (SPECT/ CT) may reveal previously undiagnosed tumor foci [9][10][11][12]. Planar WBS is routinely performed after radioiodine treatment; however, SPECT/CT is proven to be more accurate in the evaluation of residual disease [13][14][15][16]. Hybrid systems − integrating a SPECT camera with a computed tomography (CT) scanner in one gantry − have been in use since 2001, and in the last 5 years the application of SPECT/CT imaging system is gaining more importance [17]. In comparison to WBS, SPECT/CT adds simultaneous 3D anatomic mapping to functional imaging [18]. The recognition of artifacts is easier and metastatic foci without radioiodine uptake are also detected [19]. SPECT/CT significantly ameliorates the diagnosis and staging, and differentiates between benign and malignant foci of radioiodine accumulation [20].
Postoperative and follow-up management of patients with DTC highly depends on risk classification. Different risk stratification systems are used by the American Thyroid Association (ATA, 2009(ATA, , 2015 and the European Thyroid Association (ETA, 2006). Risk stratification systems incorporate data from cancer related factors, clinical features, results of first WBS after radioiodine therapy and serum thyroglobulin (Tg) level. The evaluation of response to initial therapy during the follow-up is especially important; risk categories may change during the course of disease. The reclassification of patients based on post-radioiodine therapy imaging influences the management of the disease and the intensity of follow-up [21].
The aim of this study was to evaluate the clinical utility of combined imaging with SPECT/CT after radioiodine treatment of patients with DTC in early risk classification and in prediction of late prognosis. These results have been compared to the currently used ATA and ETA risk stratification systems.

Characteristics of the patients
After their first radioiodine treatment, 323 consecutive DTC patients (181 at the University of Pecs and 142 at the University of Debrecen) were investigated. Demographic data are summarized in Table 1. The female to male ratio was 246 to 77. The median age at diagnosis was 46 (range 13 to 86) years. All patients were diagnosed with DTC; papillary and follicular histotypes were identified in 249 and 74 cases, respectively. Among the papillary tumors, classical papillary variant was the most common subtype (75.1%), while the incidence rate of follicular (20.9%), sclerosing (2.4%) and tall cell variants (1.6%) was lower. Of the follicular tumors, classical variant was the most common (81.1%); Hürthle-cell, trabecular and insular variants were diagnosed in 16.2% 1.3% and 1.3% of cases, respectively. Histology detected lymph node involvement in 95 cases, distant metastases were known in 12 patients. Thyroglobulin antibody (TgAb) positivity was found in 88 patients.
The study protocol was approved by the Institutional Ethics Committees of the University of Pecs and the University of Debrecen.

RAI ablation
Based on the surgical and pathological status of the tumor, patients were classified into risk groups according to the European consensus guideline [7]. Patients with low risk for recurrence, younger than 45 years and without aggressive histology were treated with 1100 MBq, while other patients received 3700 MBq doses. In order to reach effective thyroid ablation, two methods of preparation were available: thyroid hormone withdrawal or administration of recombinant human thyrotropin (rhTSH, 34 patients). Total body retention and external radiation dose were measured before discharge of the patient.

Post-radioiodine therapy imaging with WBS and SPECT/CT
Both planar whole body scan (WBS) and SPECT/ CT from the neck and chest were carried out in all patients 4-6 days after oral administration of 1100-3700 MBq radioiodine. Additional SPECT/CT scans of the abdomen and pelvis were acquired if suspicious isotope accumulations were detected on the WBS. The WBS examination consisted of anterior and posterior whole-body images acquired at 6 cm/min using a DHV SPECT/CT equipment (Mediso, Budapest, Hungary). The SPECT/CT unit consisted of dual head SPECT, 50 sec/frame, 64 frames, and a low dose, 16 slices spiral CT, 120 KeV, 50 mAs [22]. The examination was carried out with HEGP collimator. Evaluation of WBS and SPECT/CT images were performed by two independent nuclear medicine specialists and a radiologist; in case of dissent opinion, consensus was achieved. The CERTOP 01-13044/6/11-07755 Quality Management System was applied, which satisfied the requirements of the Hungarian Health Care Standards Guide (MEES) 1.0. Identical protocols were used at both university centers.

Risk classification systems and SPECT/CT based upgrading and downgrading rules
The risks of recurrence were calculated separately according to both the ATA 2009 and ETA 2006 guidelines.
According to the ATA risk stratification system, patients can be categorized as (1) low risk, if there are no local or distant metastases, all macroscopic tumor has been removed, no tumor invasion of local regional tissues, no aggressive histology or vascular invasion, no RAI uptake outside the thyroid bed on the first posttreatment WBS are present; (2) intermediate risk, if there are microscopic invasion of the tumor into the perithyroidal tissue, or cervical lymph node metastases are present, or there is RAI uptake outside the thyroid bed on the first posttreatment WBS, or aggressive histology or vascular invasion, and (3) high risk, if there is macroscopic tumor invasion, or incomplete tumor resection, or distant metastases, or thyroglobulinemia out of proportion to what is seen on the posttreatment scan [6].
According to the ETA risk classification the risk is (1) very low if the tumor is unifocal T1 (≤ 1 cm) N0M0 and there is no extension beyond the thyroid capsule; (2) low, if the tumor is T1 (> 1 cm) N0M0, or T2N0M0, or  (3) high if the tumor is any T3; any T4; any T with N1 or M1 [7]. The risk of recurrence was reevaluated based on SPECT/CT results. Patients without RAI uptake outside the thyroid bed, except those with aggressive histology, were downgraded, while those with detected tumor foci were upgraded according to the ATA classification.
The SPECT/CT results served as a basis for separating patients into two groups: those with or without residual tumor.
Serum Tg, TgAb, neck ultrasound and other imaging modalities were used during reclassification of patients at 9-12 months after the RAI treatment. No evidence of tumor was established with negative neck ultrasound, onthyroxine Tg < 0.2 ng/ml or stimulated Tg < 2.0 ng/ml, negative TgAb or significant decrease in TgAb titer. Incomplete biochemical response was determined if Tg was measurable or TgAb titer did not decrease without morphological abnormality. Structural disease was diagnosed with positive imaging findings.

SPECT/CT after the first 131 I treatment
No evidence of tumor was detected by SPECT/CT in 78.3% of cases (Table 2). Local residual tumor was observed in 6 patients (1.8%), lymph node metastases were detected in 61 cases (18.8%), lung and bone metastases were found in 13 (4.0%) and 5 (1.5%) patients, respectively. In the ATA low risk category (n = 138), 91% of patients were tumor-free; lymph node, lung and bone metastases were detected in 10, 2 and 1 cases, respectively. In the ATA intermediate category (n = 159), no evidence of tumor was established in 75%. Lymph node, lung, bone and other metastases were diagnosed in 35, 3, 1 and 1 cases. Posttherapeutic SPECT/CT detected residual disease in every forth patient. ATA high risk patients (n = 26) were tumor-free only in 18%. Typical images of a patient with lymph node and pulmonary metastases are shown in Figure 1. Non-radioiodine avid lesions with suspected malignancy were detected in 8 cases (2.5%); these cases were further investigated by positron emission tomography/computed tomography (PET/CT), CT with contrast material or magnetic resonance imaging (MRI). Changes in clinical staging were not so profound (Cohen's kappa: 0.894), since the stage of young patients did not change even if they had lymph node metastases (Table 3B). However, 18 patients were upgraded, and 14 of them were classified to stage IV category, increasing the number of patients in stage IV by 25.9% (p < 0.001).

Follow-up
Follow-up data were available in 315 cases; the median follow-up time was 37 months (range: 9-98 months). One patient died within one year and seven patients were lost for follow-up. Patients with confirmed residual tumor were treated by repeated surgery,

Comparison of the diagnostic value of the currently used risk stratification systems and SPECT/CT
Sensitivity, specificity, positive and negative predictive values and diagnostic accuracy of risk classification systems and SPECT/CT based on follow-up data at 9-12 months after radioiodine therapy are presented in Table 4A. All methods had acceptable sensitivity and NPV to predict the presence of DTC; however, the sensitivity of SPECT/CT compared to the ATA system was significantly lower (61% to 77%, p = 0.021). The ATA classification had the lowest specificity (47%) and diagnostic accuracy (53%) compared to the other systems tested (p < 0.001). The modification of ATA classification based on SPECT/CT findings significantly improved the specificity (73%) and diagnostic accuracy (72%) of this method (both p < 0.001). The results of SPECT/CT alone, without any other data, had the highest specificity (88%) and diagnostic accuracy (83%, p < 0.001).
The usefulness of risk classification systems and SPECT/CT to predict the presence of thyroid cancer at the end of follow-up is shown on Table 4B. The reclassification of patients at one year was included in the analysis. No significant differences in sensitivities were found except in case of reclassification at one year, which was 100%. Specificity of the individual parameters differed significantly, the highest value was also found in case of one-year reclassification (93%, p < 0.01).
Reclassification of patients at one year resulted in excellent diagnostic accuracy (94%). The specificity and the diagnostic accuracy of SPECT/CT alone were also high (87% and 84%), being significantly better (p < 0.01) than the values of the ATA and ETA risk stratification systems (ATA: 47% and 51%, ETA: 61% and 63%, respectively). The completion of ATA classification by SPECT/CT results provided better specificity (73%) and diagnostic accuracy (74%) than the ATA classification (p < 0.001). The diagnostic accuracy provided by the SPECT/CT to predict the presence or relapse of DTC at the end of follow-up was similar to the result of the oneyear reclassification (p = 0.59). However, SPECT/CT results are obtained one year earlier. Diagnostic accuracies of different risk stratifications according to disease stages were also calculated ( Table 5). The diagnostic accuracies of SPECT/CT at the end of follow-up in stage I, II, III and IV were 84.6%, 89.3%, 94.3% and 71.1%, respectively; these values were significantly higher than the diagnostic values of ATA and ETA risk stratifications in every stage.
The role of SPECT/CT in predicting the disease outcome was further investigated by binary logistic regression analysis; age, TNM stage, clinical staging, histology, ATA, ETA risk classification and SPECT/CT were included to the model. The age, T, M stage and the SPECT/CT result proved to be the independent predictors of the outcome at one year. These determining factors were completed by ETA risk at the end of follow-up. SPECT/CT results were the strongest predictors in both models (p < 0.001).

DISCUSSION
The postoperative management of DTC is based on the risk stratification of patients. However, different risk classification systems are used in the US, in Europe, and in other parts of the world [6,7]. The risk classification mainly rests on the pathological results and surgical findings. The ATA risk classification contains the results of WBS after RAI; however, performing WBS is not obligatory. In the last few years several articles have been published evaluating the advantages of additional SPECT/ CT over WBS alone in the management of DTC patients [11,17]. Investigating 148 consecutive patients, SPECT/ CT significantly reduced the number of equivocal findings on WBS and simultaneously was more accurate in the characterization of focal iodine accumulation in one fifth of patients [23]. The important diagnostic impact and the superiority of SPECT/CT over planar scintigraphy in cases of inconclusive lesions were also highlighted by others [24][25][26][27][28][29][30]. Despite of the obvious advantages of the hybrid imaging method, it is not a routine procedure in the world.
In this study, the role of SPECT/CT was evaluated in early risk classification of patients with DTC and in prediction of long-term prognosis compared to the risk of relapse determined by ATA and ETA risk classifications. To our best knowledge, so far our study has had the largest number of DTC patients with the longest follow-up time investigated by SPECT/CT. Moreover, this is the first study where the diagnostic value of combined imaging with additional SPECT/ CT to predict the long-term outcome of DTC was compared to the usefulness of ATA and ETA risk stratifications.
Residual tumor was detected by post-radioiodine SPECT/CT in 22% of patients and this was unexpected in the majority of cases. The results of SPECT/CT basically modified the management in a considerable ratio of patients. The information about the lack of residual disease was equally important. The ratio of reclassified cases by SPECT/CT was high (36%). The majority of reclassifications moved the patients towards lower risk   *No significant differences in sensitivities were found except in case of one-year reclassification (p < 0.01). **Specificities of the individual parameters differed significantly, the one-year reclassification had the highest value (p < 0.01).
The specificity of SPECT/CT was also significantly better than the values of the ATA and ETA risk classifications (p < 0.001). ***Diagnostic accuracy of one-year reclassification was excellent but not significantly better than that of SPECT/CT (p = 0.59). Both method provided better prediction than ATA, ETA and ATA after SPECT/CT classifications (p < 0.01).
categories. This reclassification influences the treatment and follow-up e.g. the TSH target values and the frequency of follow-up visits. The detection of non-RAI avid lesions by SPECT/CT has also crucial importance as the loss of RAI accumulating capability means that this tumor will be resistant to RAI treatment and other treatment options are required e.g. irradiation or sorafenib treatment.
In prognostic models of disease outcome evaluated by binary logistic regression analysis, age, T, M stage and SPECT/CT results were found as independent predictors; The result of SPECT/CT was the strongest determining factor both at one-year evaluation and at the end of follow-up.
We tested two different applications of the postradioiodine therapy SPECT/CT. Using the ATA risk categories, a large proportion of patients had to be reclassified based on the SPECT/CT results. Further, when post-radioiodine therapy SPECT/CT was used as the sole predictor of outcome, its specificity and diagnostic accuracy was significantly higher than any of the other currently used risk stratification systems. Using the SPECT/CT results alone, its sensitivity in predicting residual disease at one-year was lower than that of the ATA classification without SPECT/CT data; however, this difference disappeared by the end of follow-up. The lower sensitivity may be explained by the fact that very small metastatic foci are below the detection limit of SPECT/CT.
The response to the initial therapy is essential in determining long-term outcome. It has also been proven in our investigation that reclassification of patients at one-year based on the residual disease has the highest sensitivity, specificity and diagnostic accuracy predicting long-term outcome.
It is worth to mention that the ratio of follicular cancer (with potentially poor prognosis) was relatively high in our patients' cohorts, probably due to marginal iodine deficiency in Hungary. The ratio of TgAb positive patients was also higher than expected [31,32]. In TgAb positive cases, Tg cannot be used as a tumor marker for the follow-up. Therefore, the role of imaging methods in the TgAb positive patient population is even more important.
In our study, the residual disease was responsible for the biochemically or structurally incomplete response in the vast majority of patients and not a relapsing tumor was detected. It is possible that previous methods e.g. earlier Tg assays were not enough sensitive to detect the residual disease, however the follow-up time in this study is not enough long to withdraw final conclusion.

CONCLUSIONS
In conclusion, SPECT/CT after radioiodine treatment is a useful tool in the early classification of DTC patients and largely influences treatment strategy. ATA and ETA risk classification systems are sensitive and have high negative predictive values, but are less specific when compared to post-radioiodine therapy SPECT/CT. Due to its better diagnostic accuracy, post-radioiodine therapy SPECT/CT can greatly facilitate staging, risk classification and management of DTC. We suggest that post-radioiodine therapy SPECT/CT should be included in the risk classification of patients with DTC.

Informed consent
For this type of study formal consent is not required. The collection and analysis of patients' data were approved by the Institutional Research Ethical Committees of both Universities.
Author contributions E Mezosi, E V. Nagy, Sz Szujo, L Bajnok, K Zambo conceived of the study, analyzed data, and wrote the manuscript. E Mezosi, P Kenyeres performed the statistical analysis. L Sira, B Bodis, F Gyory, O Nemes, K Rucz, Zs Valkusz and K Sepp collected data and assisted in patient recruitment. E Schmidt, Zs Szabo, S Szekeres, Risk stratification of American Thyroid Association (ATA risk), Risk stratification of European Thyroid Association (ETA risk), Risk stratification of American Thyroid Association after SPECT/CT (ATA after SPECT/CT) and SPECT/CT alone (SPECT/CT).