Evidence for overestimation of the prevalence of malignancy in indeterminate thyroid nodules classified as Bethesda category III

Background. Several recent analyses of indeterminate thyroid nodules classified as Bethesda III (follicular lesion of undetermined significance) have reported considerably greater rates of malignancy than those initially reported by the Bethesda System for Reporting Cytopathology (BSRTC). These values, however, may be overestimates owing to several sources of bias, such as referral, selection, and publication biases. Our aim was to analyze the prevalence of malignancy in Bethesda III and IV thyroid nodules in a comprehensive health system less prone to institutional referral bias, excluding incidental carcinomas, and we examine the literature for publication bias. Methods. We performed a retrospective analysis with pathologic re-review of 119 patients with Bethesda III/IV cytology undergoing surgery in a comprehensive health system by examining patient and nodule characteristics. A review of the literature was performed and analyzed for publication bias. Results. The malignancy rate in resected thyroid nodules was 13% (6/48) for Bethesda III and 28% (20/71) for Bethesda IV. There were 9 of 119 patients (8%) with incidental microcarcinomas. Age <30 years was associated with an increased risk of malignancy (odds ratio, 25.8; P = .005). Sex, nodule size, and ultrasonographic features were not associated with risk of malignancy. Analysis of the literature was indicative of publication bias for Bethesda III cohorts, with reported rates positively skewed (P = .039). Conclusion. In a comprehensive health system, the rate of malignancy in Bethesda III nodules was similar to the range reported by the BSRTC. Recent reports of greater rates of malignancy may be attributable to institutional referral patterns, operative selection, inclusion of incidental microcarcinomas, and publication bias.