The study was carried out at the Oncological Endocrinology Sector of INCA, through a retrospective analysis of medical records of patients with DTC at high risk of recurrence, with the progression of structural disease after RAI. To identify the patients, an active search was carried out in the historical records of the Nuclear Medicine and Endocrinology services of INCA, including patients who underwent RAI between 08/01/1997 and 06/05/2019. This study was approved by the local Research Ethics Committees.

A total of 81 patients were included in the study. Inclusion criteria were: age over 18 years, histologically confirmed diagnosis of DTC after total thyroidectomy or near-total thyroidectomy, clinical staging of ATA high risk of recurrence; at least one RAI activity at the INCA, minimum follow-up of 12 months and presence of progression of structural disease after RAI anytime during follow-up.

Regarding the method of RAI preparation, patients were divided into three groups. Group 1 was composed of patients who performed RAI with rhTSH only. Group 2 was composed of patients who performed more than 1 RAI, and whose preparation was sometimes with rhTSH and sometimes with THW. Group 3 was composed of patients who performed RAI and whose preparation was only endogenous hypothyroidism after THW for 3–4 weeks. All patients treated with RAI at our center were instructed in a low-iodine diet.

As for the clinical characteristics of the patients, the variables studied were age at diagnosis, gender, the histological subtype of DTC, tumor size, and TNM staging, according to American Joint Committee on Cancer (AJCC) 8th edition (2018). The diagnosis of distant metastasis (DM) was classified as distant metastasis at the time of initial disease staging, distant metastasis identified during longitudinal follow-up, or no distant metastasis. The distribution of distant metastases (lung and/or bone) was also evaluated.

Regarding the presentation of distant metastasis, it was considered as initial distant metastasis if the presence of metastasis was found in imaging or histopathological exam before or at the time of diagnosis of DTC; evidence of RAI avid distant metastasis (DM) found on posttherapy whole-body scanning (RxWBS); or high stimulated thyroglobulin (Tg) observed in the first RAI, suggestive of distant metastasis and confirmed in the first imaging exam performed after RAI. Follow-up metastasis was considered when the patient had a normal imaging exam after the first RAI and the subsequent appearance of a metastatic lesion. The absence of distant metastasis was considered in those patients with no diagnosis of distant metastasis by functional examination (diagnostic RAI whole-body scanning, PET/CT, bone scintigraphy) or anatomical imaging (CT or MRI).

As for treatment, the following variables were evaluated: initial activity of the first RAI, posttherapy whole-body scanning (RxWBS) after the first RAI, RAI preparation, cumulative radioiodine activity, additional therapies. Additional therapies were those performed during follow-up, such as surgical intervention for patients with persistent/recurrent loco-regional disease, external radiotherapy, multikinase inhibitors (MKI), metastasectomy, and bisphosphonates (zolendronic acid).

And finally, variables were evaluated regarding clinical outcomes: dynamic risk stratification at the end of follow-up, according to the ATA guideline 2015; structural disease status at the end of follow-up, through the evaluation of the last Tg and antithyroglobulin and available imaging tests; progression-free survival; overall survival (at death); follow-up time; and death rate.

Due to the profile of the patients included in the study, none of them had an absence of disease at the end of follow-up. Regarding the status of the structural disease at the end of follow-up, stable disease was considered when there was no structural progression in the last year of follow-up, and disease progression when there was an increase or appearance of a new structural lesion in the last year of follow-up, as well as the need for additional therapy during this period or death.

Two different analyses were performed: Comparing the three proposed groups: THW only, rhTSH only, and a mixed group; and a second analysis comparing only patients prepared exclusively with one modality.

Median, with minimum and maximum, was used to express continuous variables. Percentages and frequencies were used for categorical variables. Categorical comparisons were analyzed using the chi-square test and continuous variables were compared using the median test and Kruskal–Wallis test. Kaplan Meier and log-rank were used for the overall survival and progression-free survival analyses. P < 0.05 were considered statistically significant. Analyses were performed using SPSS for Windows software (version 20.0; SPSS, Inc., Chicago, IL).

Eighty-one patients were included in the study. Clinical characteristics at the time of diagnosis and clinical outcomes at the end of follow-up are described in Tables 1, 2.

The median age was 63 years (range 24–83) and 61.7% were women. The most frequent histological subtype was papillary thyroid carcinoma (PTC), present in 36 patients (44.5%); followed by follicular thyroid carcinoma (CFT) in 18 patients (22.2%); papillary thyroid carcinoma, follicular variant (PTCFV) in 10 patients (12.3%); Hürthle cell carcinoma in six patients (7.4%); poorly differentiated carcinoma in six patients (7.4%); and PTC, other variants in five patients (6.2%).

The median tumor size was 4.7 cm (range 0.4–12), and the initial staging T4 was the most frequent, present in 24 patients (29.7%). In total 21 patients (25.9%) have been classified N1b lymph node metastasis and 12 patients (14.8%) N1a.

Regarding the diagnosis of distant metastasis, 81.5% were diagnosed metastasis at initial presentation, 9.9% distant metastasis during follow-up, 8.6% no distant metastasis, and were considered to have locally advanced cancer. The most common site of metastasis at initial presentation was lung, present in 45 patients (55.6%), followed by lung and bone in 11 (13.6%), and bone in 10 (12.3%).

The median activity of the first RAI was 200 mCi (range 100–300) and 29 patients (35.7%) had RAI avid distant metastasis (DM) found on posttherapy whole-body scanning (RxWBS). Regarding RAI preparation, 21 patients (25.9%) were in group 1, 11 patients (13.6%) were in group 2, and 49 patients (60.5%) were in group 3.

The median follow-up time was 83 months (range 13–281); 80 patients (98.8%) had an incomplete structural response at the end of follow-up and 1 patient had an indeterminate response. At the last evaluation, 70 patients (86.4%) had progressive structural disease or death and 11 patients (13.6%) had stable disease. Of the 81 patients, 42 (51.9%) died and 39 (48.1%) did not.

As described above, 21 patients (25.9%) were prepared for RAI only with rhTSH (group 1), 11 patients (13.6%) performed more than one RAI and were prepared sometimes with THW and sometimes with rh-TSH (group 2), and 49 patients (60.5%) performed RAI only after THW (group 3). Of the patients in group 2, 10 performed the first RAI after THW, and the others RAIs after rhTSH. Only one patient in group 2 had the first RAI after rhTSH, and the others RAIs after THW.

The clinical characteristics based on the method of preparation for RAI treatment are given in Table 3. When evaluating the clinical characteristics of the three groups according to age, sex, histology, tumor size, TNM staging, diagnosis and distribution of distant metastasis, activity, and RxWBS of the first RAI, there was no statistical difference according to the method of preparation for RAI.

The frequency of distant metastasis at initial presentation was 90.5% in group 1, 90.9% in group 2, and 75.5% in group 3. The most common site of metastasis at initial presentation in all 3 groups was the lung, present in 15 patients in group 1 (71.4%), 6 patients (54.5%) in group 2, and 24 patients (49%) in group 3.

The median activity of the first RAI was 150 mCi, 175 mCi, and 200 mCi in groups 1, 2, and 3, respectively. Regarding RAI cumulative administered activity, there was a statistical difference between the 3 groups (p 0.003), with the median in group 1 being 200 mCi (range 150–850), group 2 400 mCi (range 200–850), and group 3, 250 mCi (range 100–850). In group 2, all patients performed more than one RAI (100%), while only 23.8 and 42.8% performed more than one RAI in groups 1 and 3, respectively. When performing the paired comparison of the median accumulated radioiodine activity, there was no statistical difference between groups 1 and 3 (p 0.389) and 2 and 3 (p 0.062), but there was a statistical difference between groups 1 and 2 (p 0.002).

Additional therapies were performed in 17 patients in group 1 (81%), 10 patients in group 2 (90.9%), and 42 patients in group 3 (85.7%), and there was no statistical difference between the three groups (p 0.74).

Comparing THW only with rhTSH only, also no differences among the groups were seen (Table 4).

The clinical outcomes at the end of follow-up based on the method of preparation for RAI treatment are given in Tables 5, 6. There was no statistical difference in progression-free survival, dynamic risk stratification, structural disease status, and death between the 3 groups. All 100% of patients in group 1, 100% of group 2, and 98% of group 3 had an incomplete structural response (p 0.71). At the last evaluation, 85.7% of patients in group 1, 100% of group 2, and 83.7% of group 3 had progressing structural disease (p 0.36). The frequency of deaths was 47.6% of group 1, 36.4% of group 2, and 57.1% of group 3 (p 0.41).

There was a statistical difference (p 0.014) among the groups when evaluating the median follow-up time: 61 months (range 19–149), 125 months (range 29–281), and 88 months (range 13–241) in groups 1, 2, and 3, respectively. When performing the paired comparison between groups of the median follow-up time, there was no statistical difference between groups 1 and 3 (p 0.08), and 2 and 3 (p 0.28), but there was a statistical difference between groups 1 and 2 (p 0.002).

The median progression-free survival after RAI1 from groups 1, 2, and 3 was 21 (range 7–68), 55 (range 23–225), and 38 (range 2–121) months, respectively. There was no statistical difference in median progression-free survival after RAI1 between the groups (p 0.10).

Overall survival time was evaluated in patients who died (42 patients). The median overall survival was 73 months (CI: 56.3–89.6), being 50 months in group 1 (CI: 25.2–74.7), 125 months in group 2 (CI: 31.9–218.1), and 68 months in group 3 (CI: 54.1–81.8). There was no statistical difference between the groups (p 0.06). The overall survival time curve in patients who died is shown in Figures 1A, 2A.

Progression-free survival after first RAI was evaluated in patients with progressing structural disease at the end of follow-up, and patients with stable disease were excluded. The median progression-free survival in these patients was 24 months (CI: 15.5–32.4) in group 1, 55 months in group 2 (CI: 37.7–72.2), and 40 months in group 3 (CI: 24.9–55). There was no statistical difference between the groups (p 0.6). The progression-free survival curve is shown in Figures 1B, 2B.

No adverse effect due to the use of TSHrh was reported nor major adverse effects due to THW.

Comparing groups 1 and 3, those with only one method of preparation, it was also possible to show no difference in outcomes (Table 6).