Background: Bethesda category III thyroid cytology (Atypia of Undetermined Significance-AUS) represents a diagnostically diverse category with variable malignancy risk. Cytomorphologic subclassification may improve risk stratification. Methods: We performed a retrospective study of Bethesda category III thyroid nodules from a tertiary-care center. Cases were subclassified according to the predominant cytomorphologic atypia (architectural, nuclear, oncocytic, and cystic/inflammatory). Where available, cytologic findings were correlated with surgical histopathology. Malignancy rates were calculated for both definitively linked cytology–histology cases and all surgically treated nodules. Results: Of 1,872 thyroid fine needle aspiration (FNA) procedures performed between 2013 and 2022, 243 nodules (13.0%) were classified as Bethesda category III. Architectural atypia was the most common subtype (44.4%), followed by nuclear atypia (29.2%), cystic/inflammatory atypia (14.9%), and oncocytic atypia (11.5%). Surgical resection was performed in 186 cases, with definitive cytology–histology correlation established in 136. Malignancy was identified in 41.9% of definitively linked cases. Nuclear atypia demonstrated the highest malignancy rate (68.3%), significantly higher than architectural atypia (32.3%) (OR 4.52; 95% CI 1.94–10.54; p < 0.001). Papillary thyroid carcinoma was the predominant malignant diagnosis. Conclusions: Cytomorphologic subclassification of Bethesda category III (AUS) thyroid nodules identifies nuclear atypia as the strongest predictor of malignancy. Incorporation of cytomorphologic subclassification into routine reporting may improve risk stratification and clinical decision-making in Bethesda category III nodules. Malignancy rates varied substantially according to denominator definition, highlighting the impact of case selection on risk estimation.
Building similarity graph...
Analyzing shared references across papers
Loading...
Rinë Limani
Zgjim Limani
University of Prishtina
Shkëlzen Reçica
University of Prishtina
Acta Cytologica
Building similarity graph...
Analyzing shared references across papers
Loading...
Limani et al. (Mon,) studied this question.
synapsesocial.com/papers/6a1fc4e4dee9eb8c0dce64e2 — DOI: https://doi.org/10.1159/000552791