Relapse-free survival after adjuvant radioactive iodine therapy in patients with differentiated thyroid carcinoma with a microscopically positive tumor margin

Objective In patients with differentiated thyroid carcinoma (DTC) with a microscopically positive tumor margin, the optimal first dose of I-131 during adjuvant radioactive iodine (RAI) therapy to improve relapse-free survival (RFS) is unclear. Due to the limited number of hospital beds, 1110 MBq of I-131 is administered to such patients in Japan. This study primarily aimed to retrospectively determine the difference in RFS between high-dose (3700 MBq) and low-dose (1110 MBq) adjuvant RAI therapies in DTC patients with a microscopically positive tumor margin. The secondary aim was to investigate the background factors affecting RFS. Methods Forty-eight consecutive patients (15 men and 33 women) who underwent total thyroidectomy between April 2007 and December 2017 for DTC without gross residual tumors and distant metastasis, and who were diagnosed with a positive margin histopathologically, followed by RAI therapy, were enrolled. We retrospectively investigated initial I-131 dose, classifying it into high-dose and low-dose groups. The primary endpoint was RFS. Relapse was considered in the following cases: patients with visible recurrent tumor and/or metastasis on neck ultrasound and/or computed tomography, and patients without visible tumor, but with clearly increasing thyroglobulin levels on follow-up every 3–6 months. RFS outcomes were analyzed using the Kaplan–Meier method. Univariate analysis was performed using the log-rank test, and multivariate analysis was undertaken using the Cox proportional hazard model. Results There were nine cases of recurrence (52.9%) in the low-dose group and seven in the high-dose group (22.5%) during follow-up. The estimated median RFS was 69.4 months in the low-dose group and 120.7 months in the high-dose group. High-dose administration was associated with improved RFS, as demonstrated by Kaplan–Meier survival curves (log-rank test [ P = 0.009]). Patient factors associated with worse RFS included low-dose administration (hazard ratio [HR] = 91.9; 95% confidence interval [CI] = 7.66–1102.79); P