Prognostic Impact of Direct 131I Therapy After Detection of Biochemical Recurrence in Intermediate or High-Risk Differentiated Thyroid Cancer: A Retrospective Cohort Study

Background: Patients treated for intermediate- or high-risk differentiated thyroid carcinoma (DTC) and Thyroglobulin (TG) elevation during follow-up, require a diagnostic whole-body scan (DWBS) and if positive, 131I treatment. This approach can lead to a delay in treatment and increased costs. The purpose of this study is to compare the oncologic outcomes associated to administration of direct therapy with 131I at first biochemical recurrence. Methods: Retrospective cohort study of patients with intermediate- or high-risk DTC treated with total thyroidectomy, 131I ablation and who developed TG elevation during follow-up, between January 2007 and December 2015. Cohort A included patients who underwent a DWBS with 5 mCi of 131I, and if negative an MRI and/or 18FDG PET-CT prior to the therapeutic dosage, and cohort B included those who only received a therapeutic dosage of 131I, without a DWBS or extensive image studies. Main outcomes were second recurrence (SR) and disease-free survival (DFS). The diagnostic accuracy of DWBS was analyzed. Results: Cohorts A and B had 74 and 41 patients, each. By multivariate analysis, age, differentiation grade, TN classification, ablation dose, and performed DWBS (odds ratio 55.1; 95% CI 11.3–269) were associated with SR (p < 0.0001); age, male gender, ablation dose and performed DWBS (hazard ratio 7.79; 95% CI 3.67–16.5) were independent factors associated with DFS (p < 0.0001). DWBS diagnostic accuracy was 36.48%. Conclusion: 131I treatment in patients with DTC biochemical recurrence and no DWBS or extensive image studies is associated with a significantly lower frequency of SR and an increased DFS. The diagnostic accuracy of DWBS is low, and its clinical efficiency should be defined in prospective phase III studies.


INTRODUCTION
The management of patients with differentiated thyroid carcinoma (DTC) treated with thyroidectomy and 131 I ablation and who during follow-up develop progressive elevation of serum Thyroglobulin (TG), is the administration of 5 milliCuries (mCi) of 131 I to obtain an exploratory diagnostic whole-body scintigraphy scan (DWBS), CT (computed tomography), or MRI (magnetic resonance image), and even a 18 FDG PET-CT or continued follow up of thyroglobulin levels. If image studies are positive, or TG levels continue to increase, patients receive an 131 I therapeutic dose of 50-200 mCi (1,2).
The diagnostic accuracy of a 5 mCi 131 I DWBS in the presence of a detected or suspected biochemical recurrence has not been reported and has been insufficiently investigated in studies yielding mostly inconclusive results (3)(4)(5)(6). This raises the possibility of greater false-negative rates than those previously reported, and the consequent impact on final outcomes. Patients with a progressive increase in TG or anti-TG antibodies levels -even when high-risk factors are present-and a negative DWBS, the so called TENIS syndrome (7), are continuously followed with periodic DWBS until clinical recurrence, a positive DWBS or a greater increase in TG or anti-TG antibodies occurs. Moreover, some studies suggest performance of a 99mTc-3PRGD2 SPECT/CT in these circumstances (8). This policy delays treatment, increases costs, and induces stunning in the recurrent thyroid tissue, with its possible deleterious impact on prognosis.
The administration of 150-200 mCi of 131 I in the presence of a biochemical recurrence without a DWBS, aforementioned image studies or continuous follow-up, is not routinely dispensed due to the assumed low 131 I uptake and the risk of secondary malignancies, bone marrow dysplasia, toxic secondary effects, and endocrine dysfunction (9,10). Therefore, the uptake rate and the oncologic outcomes of this treatment strategy have not been reported, but indirect study conclusions on 131 I administration when TG is elevated and the DWBS is negative (11).
We analyzed the prognostic impact of direct 131 I therapeutic dose in DTC after biochemical recurrence, in a cohort of patients with intermediate-and high-risk DTC, in terms of the frequency of a second recurrence (SR) and disease-free survival (DFS). The diagnostic accuracy of the DWBS was also evaluated.

Patients
Retrospective cohort study of patients with intermediate-or high-risk DTC, treated at our Institution between January 2007 and December 2015. Inclusion criteria: patients older than 18 years with histopathology reports demonstrating DTC (papillary, follicular, or mixed papillary/follicular carcinoma). High-risk was defined as an age above 55 years and one or more of these factors: tumor size over 4 cm, extension to soft tissue, clinical lymph node involvement, or poor differentiation. Patients younger than 55 but with two or more of the aforementioned factors, were considered intermediate-risk (1,2). Patients treated with total thyroidectomy followed by adjuvant 131 I ablation (100-150 mCi) were included if, during follow-up, they developed a biochemical recurrence defined as a Tg elevation ≥2 ng/ml and/or anti-Tg antibodies ≥50 IU/ml, under Thyroid stimulating hormone (TSH) suppression (serum level <0.5 mIU/L), with no clinical or image evidence of disease (4,10). Follow-up visits were scheduled every 4 months during the first 2 years after thyroidectomy, then twice a year, and annually afterwards. Neck ultrasound and chest x-ray were obtained every year, and when the TG value increased.
Cohort A were treated according to ATA guidelines. Patients underwent a DWBS with 5 mCi of 131 I prior to the therapeutic dosage. If uptake was positive, they received a 131 I dosage of 150-200 mCi. A head/neck and chest CT scan and/or magnetic resonance imaging (MRI) was performed if the DWBS was negative, as well as a 18 FDG PET/CT since 2011. If the aforementioned tests were negative for metastases or recurrent disease, follow-up continued until the DWBS was positive or if TG levels continued to increase, and a therapeutic 131 I dosage was administered. If clinical-image (structural) recurrences ensued, these were treated with surgery if feasible, and/or radiation therapy, followed by a 131 I therapeutic dose.
Patients assigned to one of the authors' (JFC) outpatient clinic for follow-up after thyroidectomy and ablative 131 I treatment represent cohort B. To avoid delay in management and according to reports which advice direct administration of 131 I in case of increasing TG levels, these patients received a therapeutic dosage of 131 I without a prior DWBS or extensive image studies, but if this scan was negative an MRI and/or an 18 FDG PET/CT were performed -as in cohort A-to search for structural recurrences elsewhere.
There were no major differences regarding outpatient clinic visits or follow-up tests in cohorts A or B, except regarding diagnostic and therapeutic strategy when a biochemical recurrence occurs.
The study protocol is STROBE-compliant and was accepted by the IRB and Ethics committee (01/2017). Both committees have waived the requirement for written informed consent because of the retrospective nature of the study.

Variables
Clinical, baseline and follow-up blood cytology and biochemistry, ultrasonography and x-ray data were recorded. Serum TG (ELSA-HTG immunoradiometric assay, CISBio International, Codolet, France), and anti-TG antibodies levels (Immulite 2000 Analyzer, Siemens Healthcare GmbH, Erlangen, Germany), as well as serum TSH (radioimmunometric assay, SPAC-5TSHkit, Daiichi, Japan) were measured. DWBS was performed after TSH stimulation after a 20-30 days withdrawal of levothyroxine administration. If TSH levels were ≥30 mIU/L, a 5 mCi 131 I dosage was administered and the DWBS was obtained 48 h later (Symbia T6, Siemens Medical Solutions USA, Inc., Malvern, PA, USA). 131 I treatment was conducted following the same parameters and with the equipment mentioned above; a dose of 150-200 mCi was administered and a scan was obtained 5-10 days later. Scan with planar images was performed in 100% of cases. SPECT-CT was performed in 47 cases (40.86%).

Statistical Analysis
Sample size was calculated assuming a frequency of SR of 90% in cohort A and 70% in cohort B, with 80% power, a probability value of 0.05, and an allocation ratio of 2:1; a total of 132 patients was necessary (88 in cohort A and 44 in cohort B). Student's T or the Chi squared tests were used for comparisons. Factors associated with SR were tested in a logistic regression model (12). SR was defined as the event characterized by an increase in Tg (≥2 ng/ml), increased anti-TG antibodies (≥50 IU/ml) under TSH suppression (serum level ≤0.5 mIU/L), or a clinical or image recurrence detected by neck ultrasound, MRI of the neck and mediastinum or 18 FDG PET/CT. Factors associated with DFS were tested with the Kaplan-Meier method, and the log-rank test was used for comparisons. DFS was calculated from the date of surgery until the SR event or the last visit. Multivariate analysis of prognostic factors associated with DFS was conducted with Cox's model (13). The diagnostic accuracy of DWBS in biochemical recurrences was determined (14). The standard for positivity in DWBS refers to clinical or imaging recurrence during follow-up, a positive follow-up DWBS, or persistent elevation of TG or anti-TG antibodies levels after DWBS, which made surgical resection imperative and/or the final administration of a therapeutic dosage of 131 I. Probability values of 0.05 or lower were considered significant using two-tailed statistics. SPSS software for Mac, version 23 was used (IBM Corp., Armonk, NY, USA).

Recurrences
SR during follow-up and after treatment of a first recurrence, was identified in 77 patients (67%). In cohort A, 68 patients (93.2%) presented a SR and in cohort B, 9 patients (22%) (p < 0.0001).
In cohort A, from 22 cases who had a positive DWBS, four (18.18%) patients finally developed a clinical recurrence which was treated with surgery in three cases-a neck dissection was performed because of neck recurrence-, and one with radiotherapy because of bone metastases.
Two patients received radiotherapy -in each cohort-because of unresectable disease considered by the treating surgeon. Afterwards all of them were subjected to an R0 resection. Three patients received adjuvant radiotherapy because of concerns on microscopic disease because of the locoregional extent of the malignancy.

Survival
Median follow-up of patients in both cohorts was 3.2 years (interquartile range 1.9-5.7). Disease-specific (DSS) and Overall survival (OS) were not calculated because only two patients (2%) in cohort A and one (2.4%) in cohort B have died due to DTC (p = 0.532). The median DFS of both cohorts was 4.66 years (95% CI 3.2-6.04).
The association of relevant factors and DFS by bivariate and multivariate analyses is described in Tables 4, 5, respectively.

DISCUSSION
The incidence of DTC and far advanced cases has increased, especially in low-income societies with access to specialized care, yielding intermediate-or high-risk cohorts (15)(16)(17).
Total thyroidectomy followed by adjuvant 131 I ablation is the current treatment of DTC (18). However, long-term cancer control rates between doses have led to conflicting results. In low-risk cases, results of ablative therapy with 131 I are probably equivalent between doses of 30-50 mCi and doses ≥100 mCi (19). In intermediate-and high-risk patients, higher-doses of 131 I might produce better cancer control (20,21).
There are reports of patients with no clinical or imaging recurrences in spite of persistently elevated TG or anti-TG antibodies, and in the presence of a negative DWBS during follow-up. In some cases, elevated marker levels (5-10%) decrease or disappear, and several authors analyze outcomes with the administration of " 131 I empirical therapy" at dosages ≥100 mCi, and lag spans up to 4 years (22,23). A cohort study of patients with DTC and long-term followup from a single cancer center with homogeneous surgical and radioiodine treatment protocols is presented. Cohort A received 131 I treatment after confirmatory DWBS, Head and neck CT or MRI. From 2011 a 18 FDG PET-CT was performed in these cases-which demonstrated a recurrence-or persistent elevation and increase of TG levels which prompted administration of 131 I therapeutic dose. Cohort B received 131 I treatment but no DWBS or further diagnostic studies or follow-up were performed.
An intermediate-to high-risk category was established in our patients because ≥80% had a T3 or above classification; their mean age was 51 years, lymph nodes were positive in >75% of cases, mean tumor size was 4.5 cm, major extrathyroidal extension was present in >90% of cases, and a small number of patients had aggressive histopathology (1,24).
Mazzaferri reported the limited value of DWBS because of high false-negative rates and recommended a therapeutic dose of 131 I with increasing TG values, in the presence of a negative neck ultrasound (6). However, he considered the detection of recombinant human TSH (rhTSH)-stimulated TG ≥2 ng/ml as indicative of treatment and this policy was used even in low-risk patients. This is controversial because of the slow progression of DTC and the high-cost of rhTSH (22). Others have considered the administration of an 131 I dose when TG increases during follow-up, because of a high DWBS false-negative rate (67%), although they do not elaborate any further (6,25,26). We address these controversies establishing a high efficiency strategy in the presence of biochemical recurrence, consisting in the direct administration of a therapeutic 131 I dosage in high-risk patients. The administration in intermediate-and high-risk cases of a therapeutic 131 I dose ≥100 mCi without the performance of a DWBS, follow up or extensive image studies, in the presence of a biochemical recurrence, was associated with a decrease in SR and increased DFS (4,6).
Although we agree with recommendation 67B from ATA guidelines, sections (i) and (ii) regarding DWBS in patients with abnormal uptake outside the thyroid bed and large thyroid remnants on post-therapy WBS (1), these two situations are probably indicative of metastatic or residual disease and not necessarily represent biochemical recurrence, which is the subject of our study. Regarding section (iii) of same recommendation, anti-TG antibodies elevation in our series was successfully treated with direct 131 I therapy in six cases. Two other cases belonged to cohort A and finally developed clinical recurrence which was treated surgically. The 123 I isotope is not currently available in our country and its high-cost hampers its widespread use.
We agree with recommendation 68B from ATA guidelines which states that 18 FDG PET-CT scan should be considered   in high-risk patients with negative radioiodine scan, and this was included recently-as established in the manuscript-in the diagnostic strategy of cohort A. Even in cohort B cases if a negative uptake ensued after a direct 131 I dose, we performed an MRI and 18 FDG PET-CT of the mediastinum and neck from year 2011. However, a word of caution should be given because the sensitivity and specificity of this tool have been reported as 60 and 80% (27) in patients with stimulated TG ≥30 ng/ml. Aside, a major limit for its diagnostic accuracy is that significant findings are related to a ≥5 mm diameter of the lesion. Moreover, larger lesions could be considered structural recurrences which could be detected with CT and or MRI of head and neck and chest areas. Although our analyses was not designed to evaluate the diagnostic accuracy of 18 FDG PET-CT diagnostic accuracy, sensitivity and specificity were 30.77, 25, and 100%, respectively, in those patients who had this test performed in our study, regarding biochemical recurrences. The differentiation grade was associated with SR by multivariate analysis as has been reported (21,24), and underscores the need to effectively treat patients with biochemical recurrence and high-risk factors, and avoid a DWBS (6). Although, as stated by ATA guidelines, recommendations for 18 FDG PET-CT should be considered in cases with aggressive histology and metastatic disease to identify lesions at risk for rapid progression and evaluate the response after local or systemic therapy for invasive and metastatic disease, same aforementioned 18 FDG PET-CT limitations regarding diagnostic accuracy of the test should be considered regarding this issue.
There is a tendency to administer lower ablation and therapeutic doses of 131 I in DTC patients, even in intermediaterisk cases (1). In our study, higher ablative 131 I doses were associated with decreased SR by multivariate analysis; this underscores the need to define the treatment strategy according to the clinical-histopathological characteristics and risks, especially in advanced stages (21).
RLNL and permanent hypoparathyroidism rates were high. Of note, these cases had more advanced disease and a high surgical risk (28,29).
As in any cohort study, a potential pitfall in this study is the non-random and open nature of patient allocation in each cohort. The actual sample size is adequate because the observed differences between cohorts were greater than expected.
A clear difference exists in the uptake rate of DWBS compared with that after therapeutic doses in the setting of biochemical recurrences; the diagnostic accuracy of DWBS was 36.48%. This was analyzed in spite of the lack of a histopathological gold standard but the follow-up and TG levels are sufficient to establish the diagnostic accuracy values for DWBS (6,26,30). Moreover, no explanation exists for the positive 131 I uptake after a direct therapeutic dose, after a 1-year follow-up with negative TG levels.
There are reports on the so-called "empiric" 131 I dose administration after a negative DWBS, the TENIS syndrome (7), in patients who maintained persistent or elevated values of biochemical markers which made the administration of 131 I therapy imperative or until clinical or image recurrence ensued (4,10). These reports refer a high 131 I uptake rate (≥65%) in the post-therapy scans. However, major drawbacks include the length of time before a therapeutic decision is made, stunning after repeated DWBS, and the high cost and delay of treatment because of the usually performed image studies (CT and 18 FDG PET-CT scans). Also, over 50% of patients were low-risk and early-stage in these series, which decreases the clinical impact of this strategy (4,10). In our study, over 95% of cases who received 131 I therapy had a positive uptake.
Shinohara categorized TG and DWBS. Two groups with negative DWBS were described, one with TG elevation and another with normal TG (21). The first group had higher risk factors and a lower DFS. This suggests the need to take biochemical recurrences into consideration in high-risk patients and use a different strategy, perhaps a direct therapeutic dosage, as we have proposed.
We agree with some authors (1, 10) who consider that a biochemical incomplete response is not uncommon, with final return to normal limits in a majority of cases. Although a special consideration should be given to the fact that this regression refers mostly (10) to stimulated TG levels in low risk patients. Our study refers to a group of patients selected as intermediate-and high-risk with more stringent criteria for risk stratification, and to a non-stimulated cutoff of 2 ng/ml which is highly suspicious for a recurrence as demonstrated in our study. In intermediateand high-risk cases -as established in our study-, spontaneous remission of high TG levels occurred in only 5 cases (6.75%) from patients who had a DWBS, with progressive increase of TG levels in the rest of them.
We agree with the ATA guidelines recommendations 80, 81, and 82 (1), regarding empiric radioiodine therapy in patients with increasing TG and anti-TG levels, although our success rate was higher (probably because of the inclusion in our series of high-and intermediate-risk patients), with 97% positive uptake rate and second recurrence rate of 21% after direct 131 I therapy. A word of caution should be given considering the low sensitivity and negative predictive value of 18 FDG-PETCT already mentioned.
With direct 131 I therapy, the possibility of developing clinical or imaging metastases is decreased, since our strategy has the potential to prevent the appearance of resistant cell strains and therefore avoid stunning-which when unresectable, are frequently reported as resistant to radiotherapy (31,32). Although there is evidence for the use of multitargeted kinase inhibitors in 131 I-resistant cases, resistance and the need to administer these agents should be avoided (31).
There are reports of acute myeloid leukemia in patients with DTC treated with radioiodine (9). However, the total dosages used are usually not described. Furthermore, as stated by the same authors, the recurrence risk if no 131 I therapy is used, is higher than the possibility of developing leukemia in highrisk cases.
In spite of a good prognosis in low-risk patients, some do recur, and early detection is imperative. The administration of 131 I therapy in these patients should be judicious because of quality of life concerns and the risk of secondary malignancies (33). Lower doses in cases of a suppressed TG increase, could be indicated and investigated in the future, given our results.
A greater number of intermediate-and high-risk cases should be recruited to establish the approach reported as standard in the treatment of biochemical recurrences. More sensitive iodine scanners and studies proposing the rational use of 18 FDG-PETCT might confirm our results. Prospective and multiinstitutional studies to confirm these findings with possible impact on survival are warranted (8,34,35).
Intermediate-and high-risk DTC patients with TG elevation during follow-up should probably receive a therapeutic dose of 131 I without a prior DWBS to prevent TENIS syndrome, treatment delays and increased costs. Other image studies like CT and MRI as well as 18 FDG-PETCT and follow up of TG levels could be avoided. This approach improves DFS in patients with biochemical recurrence. DWBS diagnostic accuracy is low and its performance could induce stunning, delay treatment and increase clinical recurrence rates.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/supplementary material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Comité de Investigación y Comité de Etica en Investigación, Instituto Nacional de Cancerología, México. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

ACKNOWLEDGMENTS
We are grateful to Professor Deborah Aleman-Hoey for her kind support with the English-Language review.