Abstract
AIMS: Since the six-tier Milan salivary gland classification has been introduced, there are very limited studies in literature reporting the risk stratification of the Milan classification. METHODS: We retrospectively classified a total of 285 salivary gland cytology cases into Milan reporting categories; there were 23 (8.1%) nondiagnostic, 48 (16.8%) nonneoplastic, 19 (6.7%) atypia of undetermined significance (AUS), 138 (48.4%) benign neoplasm, 13 (4.6%) neoplasm of uncertain malignant potential (NUMP), 8 (2.8%) suspicious for malignancy, and 36 (12.6%) malignant. Almost 110 cases (38.6%) had surgical follow-up resections. RESULTS: The overall risk for malignancy (ROM) was 12.5% for AUS, 3.2% for benign neoplasm, 72.7% for NUMP, and 100% for the suspicious for malignancy and malignant. The ROM for nondiagnostic and nonneoplastic categories was not representative due to limited follow-up resections. The salivary cytology had sensitivity, specificity, positive predictive value, and negative predictive value of 93.0%, 100%, 100%, and 46.2% for neoplasm and 82.3%, 95.8%, 90.3%, and 92.0% for malignant. CONCLUSION: Our study supports the adaptation of the six-tier Milan classification for reporting salivary gland cytology, as well as emphasizing the utility of the NUMP category.