"Suspicious" salivary gland FNA: Risk of malignancy and interinstitutional variability

BACKGROUNDFine-needle aspiration (FNA) cytology is well accepted as a safe, reliable, minimally invasive, and cost-effective method for the diagnosis of salivary gland lesions. Salivary gland neoplasms are often difficult to diagnose because of morphologic heterogeneity and a variety of epithelial metaplastic changes. Hence, a number of salivary gland FNA specimens yield indeterminate results. For indeterminate FNA specimens, the suspicious-for-malignancy (SFM) category is used when a specific neoplasm falls short in quantity or quality for the criteria for malignancy. Therefore, the findings are not sufficient for a conclusive diagnosis of malignancy.METHODSThis study was designed to evaluate the risk of malignancy (ROM) for the SFM group at 5 tertiary medical centers worldwide with the aforementioned criteria. Among 12,606 salivary gland FNA cases between 1997 and 2014, 276 (2.2%) were reported to be SFN. Specifically, 114 suspicious cases (41%) had histological follow-up.RESULTSHistological follow-up of the 114 suspicious cases showed 95 malignant tumors indicating a risk of malignancy (ROM) of 83.3%. The ROM varied between 74% and 88% for the 5 participating institutions, and a Fisher's exact test with significance set to p<.05 showed no significant difference in ROM among the institutions (p = .78).CONCLUSIONSOverall, 83.3% of SFM salivary gland FNA specimens turned out to be malignant; there was no significant interinstitutional variability in the ROMs. The SFM category for salivary gland FNA is very homogeneous, and the ROMs are quite similar worldwide. Cancer Cytopathol 2018;126:94-100. (c) 2017 American Cancer Society.The suspicious-for-malignancy category in salivary gland cytology is used for lesions when a specific neoplasm falls short in quantity or quality for the criteria for malignancy. In this study, 83.3% of 114 suspicious-for-malignancy salivary gland fine-needle aspiration specimens are found to be malignant, and there is no significant interinstitutional variability in the risk of malignancy.