e18087 Background: Differentiated thyroid cancer (DTC) has a good prognosis; however, about 30% of patients will experience a recurrence after treatment and become refractory to radioiodine therapy. Some studies, including randomized controlled trials (RCTs), have shown the efficacy of kinase inhibitors in these patients, but whether overall response rate (ORR) can predict progression-free survival (PFS) and/or overall survival (OS) has not been well established in the literature. We examined the correlations between ORR and PFS/OS with kinase inhibitors in radioiodine (RAI)-refractory DTC. Methods: PubMed, SCOPUS, Embase, and ClinicalTrials.gov were queried for studies that used kinase inhibitors in RAI-refractory DTC. Eligibility included a clinical diagnosis of RAI-refractory DTC, the use of kinase inhibitors, and articles published in English. ORR, PFS, and OS data were manually retrieved from the studies, and correlative analyses using Spearman's rank correlation coefficient ( rs) were conducted to determine their relationships. Results: Two hundred ninety articles were identified from the literature search; after removing duplicates and screening for relevance and eligibility, 24 studies (6 RCTs and 18 observational), consisting of 1,791 participants, were included for analysis. The kinase inhibitors used were: sorafenib, lenvatinib, donafenib, apatinib, sunitinib, cabozantinib, and vemurafenib. There was no significant correlation between ORR and PFS or OS, and between PFS and OS. However, a sub-analysis of the 6 RCTs (n = 410) alone showed a significantly strong positive correlation between ORR and PFS ( rs = 0.92, p = 0.003) with a 20% increase in ORR corresponding to a prolonged PFS by about five months. No significant correlation was found between any two endpoints in the observational studies. Conclusions: Overall response rate is a predictor of PFS and can be used as a surrogate endpoint for the latter (PFS) in RCTs with kinase inhibitors in RAI-refractory DTC, especially when the time to reach PFS data may be protracted, costly, and delayed treatment. The lack of correlations between ORR and PFS or OS in the observational studies may be due to different populations with varying responses, less controlled settings, time to reach OS data, and probably smaller sample sizes.