Differential outcomes of patients with thyroid FNA diagnoses of AUS/FLUS with and without nuclear atypia: The potential need for separation in the Bethesda System

Differential outcomes of patients with thyroid FNA diagnoses of AUS/FLUS with and without nuclear atypia: The potential need for separation in the Bethesda System:

Abstract

Background: In the current version of The Bethesda System (TBS) for thyroid cytopathology, the atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS) category has an estimated risk of malignancy of 10% to 30%. Diagnostic criteria include presence of nuclear atypia, suggestive of papillary thyroid carcinoma (PTC), as well as other types of atypia, which can be seen with non‐malignant entities. Aim of this study was to investigate differential outcomes of AUS/FLUS, based on specific morphologic criteria, and assess their respective malignancy risks.

Methods: From a total of 1233 patients undergoing thyroid FNAs between 2010 and 2014 at the University of Washington, 119 had AUS/FLUS without nuclear atypia, and 64 with nuclear atypia. Outcomes for patients with and without nuclear atypia (with the exception of 24 patients lost to follow‐up) were evaluated and results were compared.

Results: 16/57 (28.1%) patients with AUS/FLUS and nuclear atypia subsequently had carcinomas on thyroidectomy, statistically higher than the 8/102 patients (7.8%, P = .001) without nuclear atypia. When comparing only patients who underwent surgery (n = 63), again those with AUS/FLUS and nuclear atypia had statistically higher rates of carcinoma (16/31, 51.6%), compared to those without (8/32, 25%; P = .0394). Overall, 24/159 (15.1%) of patients with AUS/FLUS had carcinoma on subsequent histology.

Conclusion: Malignancy rates for AUS/FLUS were in line with TBS estimated risks. However, our data demonstrate that the presence or absence of nuclear atypia is associated with different malignancy rates, suggesting the possibility that the AUS/FLUS category may best be split into two subcategories with different implied risks of malignancy.