Fine-needle aspiration (FNA) cytology of salivary glands is a well-established technique that plays a critical role in the preoperative diagnosis of lesions. A risk stratification of FNA diagnostic categories has been recently proposed to be useful in reporting. The aims of this study were to evaluate the diagnostic accuracy of salivary gland FNA and to apply the proposed Milan system for reporting salivary gland lesions.
A retrospective audit of FNA specimens of salivary gland lesions reported from 2014 to 2016 was performed. A correlation with the follow-up histopathology, wherever it was available, was performed. The aspirates were then categorized according to the Milan system as follows: nondiagnostic, nonneoplastic, atypical, benign neoplasm, neoplasm of uncertain malignant potential (NUMP), suspicious for malignancy, or positive for malignancy. Furthermore, the risk of malignancy and the risk of high-grade malignancy were calculated for all diagnostic categories.
A total of 631 salivary gland aspirates were evaluated: 2.2% were nondiagnostic, 55.8% indicated nonneoplastic lesions, and 40.4% indicated neoplastic lesions. Histopathology was available for 94 cases (14.9%). FNA had a sensitivity of 79.4% and a specificity of 98.3% with an overall diagnostic accuracy of 91.4% for differentiating malignant tumors from benign tumors. The overall risk of malignancy was 17.4% for the nonneoplastic category, 100% for the atypical category, 7.3% for the benign neoplasm category, 50% for the NUMP category, and 96% for the positive-for-malignancy category.
Salivary gland FNA continues to have high diagnostic accuracy and is thus helpful for guiding management. Neoplasms with classic cytomorphology are easily diagnosed; however, in difficult cases showing overlapping features, the use of the Milan system could be beneficial. Cancer Cytopathol 2017;125:767-75. © 2017 American Cancer Society.