Tumor size and molecular risk group are associated with differentiated thyroid cancer recurrence.

Abstract

The threshold at which active surveillance can be considered is variable, with some algorithms proposing nonoperative treatment for differentiated thyroid carcinomas ≤2 cm and lobectomy alone for lesions 2.1-4 cm. To inform both decision for and extent of initial surgery, we aim to evaluate whether molecular results can complement tumor size to identify differentiated thyroid carcinomas associated with disease recurrence. Patients from 2007-2013 and 2017-2021 who had initial thyroidectomy (differentiated thyroid carcinoma size 1-4 cm, clinical N0M0) were included. When available, molecular testing results were categorized into 3 previously described molecular risk groups (low, intermediate, and high). Primary outcome was structural recurrence. Recurrence was diagnosed in 3.8% of 1,739 patients with differentiated thyroid carcinomas. Preoperative variables including size (1-2 cm vs 2.1-4 cm, P = .43), age >55 years (P = .92), and male sex (P = .31) were not associated with recurrence. Molecular testing results were available for 1,020, and after excluding molecular risk group high-risk differentiated thyroid carcinoma, structural recurrences were associated with molecular risk group intermediate risk (7.2% vs molecular risk group low, 0.7%, P < .001), and most likely in differentiated thyroid carcinoma, which were both 2.1-4 cm and molecular risk group intermediate risk (11.3% vs size 1-2 cm 5.8%, P = .04). Overall, structural recurrences for differentiated thyroid carcinomas ≤4 cm were low (<5%) and molecuar testing was the only preoperative variable associated with recurrence. However, when molecular risk group intermediate risk was present, larger tumor size (2.1-4 cm) had a 2-fold greater risk of recurrence compared with tumors 1-2 cm, and size may still be helpful to guide management. When considering de-escalated treatment for the proposed guidelines with a cutoff of 2 cm, initial decision-making may be further optimized with identification of preoperative molecular risk groups.