Patients with PTC who underwent TT followed by RRA between 2000 and 2012 were eligible for enrollment in the current study. The definition of low-risk PTC was determined according to the ATA guideline as follows: without local or distant metastasis; no extrathyroidal extension; negative residual margin; no radioactive iodine (RAI) avid metastatic foci outside the thyroid bed on the first post-surgical RAI scan; no lymphovascular invasion; and clinical N0 or <5 pathologic N1 micrometastases (<0.2 cm in the largest dimension) [1]. Of the 1,446 patients with low-risk PTC, those with papillary microcarcinoma measuring <1 cm were excluded. Finally, 526 patients were enrolled in this study. The patients were divided into the early RRA group (<90 days; n=451) and the delayed RRA group (≥90 days; n=75) based on the interval between TT and the first RRA [2]. Written informed consent by the patients was waived due to a retrospective nature of our study and the study protocol was approved by the Institutional Review Board of Asan Medical Center (study number: 2019-1562), Seoul, South Korea.

All patients underwent TT with or without central neck dissection, followed by RRA. The RRA therapy and I-131 therapeutic doses were decided according to the protocol of Asan Medical Center [2]. All patients underwent the thyroid hormone withdrawal method for RRA. The I-131 doses were determined based on clinicopathological parameters. In brief, PTCs with a primary tumor size of <4 cm that were pathologically defined as N0 were treated with 30 mCi of I-131. PTCs with a primary tumor size of <4 cm that were pathologically defined as N1a were treated with 80 mCi of I-131. All PTCs with a primary tumor size of >4 cm with or without cervical LN metastasis were treated with 150 mCi of I-131.

After RRA, all patients were administered levothyroxine for suppression of thyroid-stimulating hormones. Patients were followed up regularly (every 6 to 24 months) with physical examination, neck ultrasonography (US), and laboratory tests including thyroid function, serum Tg, and anti-thyroglobulin antibody (TgAb) tests. During the first 12 to 24 months of initial therapy, neck US and/or diagnostic whole-body scan (DxWBS) were performed with measurement of stimulated Tg (sTg) and TgAb, and the response to the initial therapy was assessed [17–20]. If suspected signs for recurrence were present in routine follow-up, further evaluation, such as additional imaging or biopsy, was performed. Additional diagnostic imaging studies, such as computed tomography (CT) or whole-body fluorodeoxyglucose positron emission tomography/CT, were performed in some patients, as required.

The study outcomes were ongoing risk stratification (ORS; response to therapy) and DFS between the two groups based on the interval of RRA. ORS was categorized into excellent, indeterminate, biochemical incomplete, structural incomplete response based on the response to the initial therapy within the first 2 years of follow-up [1,2,21,22]. DFS was calculated from the date of the initial operation to the date of confirmation of structural recurrence by cytopathology. Disease-specific survival was not included in the study outcome because the mortality rate was very low in the low-risk PTC cases.

Rstudio version 1.1.463 (RStudio: Integrated Development for R, RStudio Inc., Boston, MA, USA; www.rstudio.com) was used to perform statistical analyses. Continuous variables are presented as median with interquartile range (IQR), and categorical variables are presented as frequencies with percentages. PSM was used to adjust baseline characteristics related to disease recurrence. The propensity score was defined as the probability of being assigned to the early and delayed RRA group based on age, sex, tumor size, T stage, LN metastasis, initial ablation dose, and RAI avidity. After PSM, a standardized mean difference was used to measure balancing covariates between the groups; if a variable was balanced, then the standardized mean difference was <0.1. The chi-square test was used to determine associations between categorical variables. DFS curves were plotted using the Kaplan-Meier method, and the log-rank test was used to determine their significance. The Cox proportional hazards model was used to evaluate the risk factors for recurrent disease and was presented as hazard ratio and 95% confidence interval. P values of <0.05 were considered statistically significant.

The baseline clinical and pathological characteristics of the 526 patients with low-risk PTC is shown in Table 1 (left panel). The median interval between RRA and the initial surgery was 45 days (IQR, 40 to 56) and 127 days (IQR, 110 to 137.5) in the early RRA and delayed RRA groups, respectively (P<0.001). The median age of the patients was 47.2 years, and of the 526 patients, 93 (17.7%) were males. The median primary tumor size was 1.4 cm, and approximately 85% patients had T1 staging according to the 8th edition of the American Joint Committee on Cancer (AJCC) tumor node metastasis (TNM) staging system. Details of AJCC TNM staging system in each group are shown in Supplemental Table S1. There was no significant difference in age, sex, primary tumor size, and T staging between the early and delayed RRA groups. Cervical LN metastasis was confirmed in 80 patients (12%), and patients in the delayed RRA group had significantly more metastasis than those in the early RRA group (64% vs. 7%). In 80 patients with LN metastasis, the number of metastatic LN was as below; 1 (n=46, 57.5%), 2 (n=24, 30.0%), 3 (n=5, 6.2%), and 4 (n=5, 6.2%). The median size of metastatic LN was 0.1 cm (minimum to maximum range, 0.01 to 0.15). Further, patients in the delayed RRA group received a significantly higher RRA dose than those in the early RRA group (P<0.001). Therefore, we evaluated these two groups after PSM, and 57 patients were included in each group. There was no significant difference in LN metastasis and RRA dose between the two groups after PSM (Table 1, right panel).

In the first 12 to 24 months after initial therapy, DxWBS was performed in 431 (95.6%) and 72 (96%) patients in the early RRA and delayed RRA groups, respectively. There was no abnormal iodine uptake outside the operative bed and no evidence of cervical LN metastasis or distant metastasis. In total, 56 (13%) and four patients (5.6%) had focal iodine uptake limited in the operative bed in the early and delayed RRA groups, respectively. There was no significant difference in the proportion of patients with iodine uptake between the early and delayed RRA groups before and after PSM.

We evaluated the initial treatment response by comparing ORS between the early and delayed RRA groups (Table 2). A majority of the patients (77.6%) showed an excellent response, and approximately 20% showed an indeterminate response. Approximately 1% patients showed a biochemical or structural incomplete response. There was no significant difference in the initial treatment response between the early and delayed RRA groups before and after PSM.

The median follow-up period was 9.1 years (IQR, 6.8 to 11.8) after initial therapy, and the structural recurrence rate was 1.9% (10/ 526). The recurrence rate was 1.6% (7/451) and 4% (3/75) in the early and delayed RRA groups, respectively, but there was no significant difference (P=0.242). There was no significant difference in DFS between the two groups before and after PSM (P=0.181 and P=0.094, respectively) (Fig. 1). All cases of recurrent disease were locoregional recurrence, and all patients with recurrent disease underwent additional surgery, such as lateral neck dissection or selective LN excision (Table 3). After additional surgical treatment, all patients were disease free during the follow-up period, and there were no cases of disease-specific death.

Table 4 shows the clinical and pathological parameters associated with recurrent disease in patients with low-risk PTC. In multivariate analyses using the Cox-proportional hazard model, male sex and primary tumor size (>2 cm) were independent risk factors of recurrent PTC. Delayed RRA was not a significant risk factor in the univariate and multivariate analyses. High RRA dose (5.55 GBq, 150 mCi) was a significant risk factor in the univariate analysis, but was not significant in the multivariate analysis.