Threshold-dependent diagnostic performance of fine-needle aspiration cytology in parotid tumors: a Milan system-based analysis.

OBJECTIVE: Fine-needle aspiration cytology (FNAC) is routinely used in the preoperative evaluation of parotid gland tumors, yet its diagnostic performance varies across categories of the Milan System for Reporting Salivary Gland Cytopathology (MSRSGC). This study evaluated category-specific malignancy risk and the clinical impact of different diagnostic thresholds. METHODS: We retrospectively analyzed 277 patients who underwent parotidectomy with available preoperative FNAC. Cytologic diagnoses were classified according to the MSRSGC and correlated with final histopathology. Diagnostic performance was assessed using two thresholds: Milan V-VI and Milan III-VI. Trends in malignancy risk across Milan categories were examined, and false-negative and false-positive cases were reviewed by histopathological subtype. Overall discriminatory performance was evaluated using receiver operating characteristic analysis. RESULTS: Final histopathology revealed benign disease in 85.2% and malignancy in 14.8% of cases. Malignancy risk increased significantly across higher Milan categories (p = 0.006). Using Milan V-VI as the threshold yielded high specificity (97.0%) and negative predictive value (88.8%) but low sensitivity (29.3%). Expanding the threshold to Milan III-VI increased sensitivity (78.0%) but markedly reduced specificity (21.2%). Discriminatory performance was poor (AUC 0.622). False-negative results were mainly low-grade malignancies, particularly low-grade mucoepidermoid carcinoma, whereas false-positive results were largely benign oncocytic or hypercellular lesions. CONCLUSIONS: FNAC provides useful preoperative risk stratification for parotid gland tumors within the Milan framework, but its diagnostic value is highly threshold-dependent and limited by tumor biology.