Diagnostic accuracy of fine-needle aspiration cytology of salivary gland lesions: A 6-year retrospective review

Introduction: The aim of this study was to evaluate the diagnostic accuracy of salivary gland fine-needle aspiration (FNA) in comparison to histologic examination and to recognize possible pitfalls in diagnosis. Materials and methods: The diagnoses and demographics of all cases of salivary gland FNAs with concurrent or subsequent histologic correlation at our institution over a 6-year period (2006-2011) were retrospectively reviewed and compared for discrepancies. Discrepancies were categorized as either major or minor and due to sampling or interpretive variance. Results: Overall, the following values were calculated: sensitivity 80.6%, specificity 97.5%, positive predictive value 92.6%, negative predictive value 92.8%, accuracy 92.7%, and concordance rate 90.9%. In addition, concordance rates were calculated for the 2 most common diagnoses: pleomorphic adenoma (97.1%, n = 35) and Warthin tumor (88.9%, n = 9). Five major and 5 minor discrepancies were found. Most of the major discrepancies and all of the minor discrepancies were due to sampling and interpretive variances, respectively. Sampling issues occurred in FNAs with and without ultrasound guidance. The interpretive variance included interpretative discrepancies in monomorphic cellular lesions, abundant inflammation and reactive atypia, cystic changes, abundant matrix deposition or fibrosis, and difficulty in diagnosing mucoepidermoid carcinoma or lymphoma on cytology. Conclusions: FNA of salivary gland lesions is a procedure with high specificity, positive predictive value, negative predictive value, accuracy, and concordance with histologic examination; however, discrepancies do exist. Recognizing potential pitfalls is key to avoiding discrepancies.