Multi‐institutional validation of a modified scheme for subcategorizing salivary gland neoplasm of uncertain malignant potential ( <scp>SUMP</scp> )

The salivary gland neoplasm of uncertain malignant potential (SUMP) category in the Milan System is diagnostically challenging. This study aims to validate a modified scheme for subcategorizing SUMP in a large multi-institutional cohort.Retrospective review of salivary gland fine-needle aspirations (FNAs) from 10 institutions were classified based on the Milan System. Cases diagnosed as SUMP with available cytology slides and surgical follow-up were retrieved for review and subcategorized based on a modified scheme as follows: basaloid SUMP (B1: absent/scant nonfibrillary matrix; B2: presence of nonfibrillary/mixed-type matrix), oncocytic/oncocytoid SUMP (O1: with mucinous background; O2: without mucinous background), and SUMP not otherwise specified (NOS).A total of 742 (7.5%) cases from 9938 consecutive salivary gland FNAs were classified as SUMP. Among them, 525 (70.8%) had surgical follow-up and 329 (62.7%) were available for review. The overall risk of malignancy (ROM) of SUMP was 40.4%. There were 156 cases (47.4%) subcategorized as basaloid SUMP with a ROM of 36.5%, 101 (30.7%) as oncocytic/oncocytoid SUMP with a ROM of 52.5%, and 72 (21.9%) as SUMP NOS with a ROM of 31.9%. The ROM of oncocytic/oncocytoid SUMP was significantly higher than basaloid SUMP (P = .0142) and SUMP NOS (P = .0084). No significant differences in ROM were noted between B1 and B2 (36.7% vs 36.4%, P = 1.0000) and O1 and O2 (65.2% vs 48.7%, P = .2349).The ROM of oncocytic/oncocytoid SUMP was 52.5% and significantly higher than that of basaloid SUMP (36.5%, P = .0142) and SUMP NOS (31.9%, P = .0084), whereas no significant differences in ROM were noted for cases with different types of extracellular matrix or background material.