Performance of Two-Tiered Subclassification of Atypia of Undetermined Significance in Thyroid Fine-Needle Aspiration Without Routine Molecular Testing.

Background: The third edition of the Bethesda system for reporting thyroid cytopathology recommends a simplified two-tiered subclassification of atypia of undetermined significance (AUS), dividing cases into AUS with nuclear atypia (AUS-Nuclear) and other atypia (AUS-Other). This study aims to evaluate the performance of these subcategories in estimating the risk of malignancy (ROM) in a setting without routine molecular testing.

Methods: A retrospective review was conducted on consecutive thyroid fine-needle aspiration (FNA) cases diagnosed as AUS between 2018 and 2023. Surgical pathology reports were matched with the FNA-targeted nodules to enable cyto-histologic correlation. ROM and risk of neoplasm (RON) were calculated for all AUS cases and for each subcategory.

Results: Among 16,030 thyroid FNA cases, 617 (3.8%) were diagnosed as AUS. Histologic follow-up was available in 190 cases. Final diagnoses included non-neoplastic lesions (42.6%), benign neoplasms (17.4%), low-risk neoplasms (5.3%), and malignant neoplasms (34.7%). The ROM for AUS-Nuclear was significantly higher at 54.7%, compared to 14.7% for AUS-Other (p < 0.001). Similarly, the RON was significantly higher in the AUS-Nuclear group (67.4%) than in the AUS-Other group (47.4%) (p = 0.005).

Conclusion: AUS-Nuclear carries a substantially higher ROM than AUS-Other, with a ROM (54.7%) comparable to the reported positive predictive values of molecular assays such as Afirma GSC (47%, 95% CI: 36%-58%) and ThyroSeq v3 (66%, 95% CI: 56%-75%). These findings support the clinical utility of the two-tiered AUS subclassification in enhancing risk stratification, particularly in settings where molecular testing is not routinely available.