Algorithmic Approach to Diagnosis of Salivary Gland Neoplasms Based on Morphology and Select Immunostains

Salivary gland tumors present diagnostic challenges even to the experienced surgical pathologist due to overlap in both histologic and immunohistochemical features. Morphology, immunoprofile, and genetic testing (in some cases) can be used judiciously in an algorithmic approach to narrow the differential diagnosis on a salivary biopsy or resection specimen. The purpose of this review is to present an objective algorithmic approach to assessing salivary gland specimens that utilizes histomorphology, a limited panel of immunohistochemical stains, and, occasionally, genetics to aid in diagnosis. The diagnostic groups include: biphasic neoplasms, monophasic glandular neoplasms, monophasic squamoid neoplasms, and oncocytic neoplasms. This review is not exhaustive but covers the most frequently encountered and clinically relevant neoplasms. The suggested initial immunopanel for most tumors includes p40, CD117, and S100, with second-round immunostains depending on the specific differential diagnosis (beta-catenin, mammaglobin, DOG1, etc). One risks getting caught in the weeds of ever-expanding salivary neoplasia, and it is important to remember that, in order for clinicians to make treatment decisions, a specific diagnosis is not always required. Important information includes whether a tumor is low-grade (including benign), high-grade, possibly adenoid cystic, or eligible for targeted therapy. If possible, correlation of salivary biopsies to their ultimate resection specimen will help ensure diagnostic accuracy and promote active learning.