Evaluation of criteria for distinguishing uterine smooth muscle tumors of uncertain malignant potential (STUMP) from mimics and improved prediction of patient outcome: A multi-institutional study of 51 cases

Abstract

Disclosures: Philip Ip: None; Zehra Ordulu: None; Diana Lim: None; Robert Soslow: Speaker, Ebix/Oakstone; Robert Young: None; Liwei Jia: None; Sarah Chiang: None; Esther Oliva: None Background: The diagnostic criteria for uterine smooth muscle tumor of uncertain malignant potential (STUMP) have varied. In addition to the Bell criteria, there have been other proposed parameters for typical spindle cell tumors. Design: Uterine STUMPs/low-grade leiomyosarcomas seen in consultation (n=72) and from our respective tertiary hospitals (n=28) were reviewed in 2 multi-head microscopy sessions when ≥4 of us were present at any time. After consensus review, 49 cases were excluded (18 diffusely epithelioid and 4 diffusely myxoid tumors, 2 inflammatory myofibroblastic tumors, 3 high-grade leiomyosarcomas, 13 benign variants and 9 with other unique morphologies). The remaining 51 cases (37 were consults) were evaluated using the Bell criteria (atypia, mitotic count and tumor cell necrosis) and classified into atypical leiomyomas with low-risk of recurrence (ALLRR), atypical leiomyomas with limited experience (ALLE) and smooth muscle tumors of low-malignant potential (SMTLMP). These were further assessed for presence of long sweeping fascicles (defined as broom-like, long and occupying >75% of the field diameter of a x10 objective, hypercellular of at least moderate degree, nuclei with a clonal, uniform appearance and with a palisading/herringbone arrangement), focal myxoid or epithelioid areas, infiltrative margins, vascular space intrusion/invasion (VSI) and atypical mitoses. 50 Results: The median age of patients was 46 years. The median size of tumors was 7.5 cm. Follow-up was obtained in 33/51 patients (64.7%). After a median disease-free period of 84 months, 15 patients (45.5%) had a recurrence. Histologically, there were 8 ALLRR, 6 ALLE, and 1 SMTLMP. Among these, 5 had an infiltrative border, 3 had VSI and 1 had focal epithelioid differentiation. All 15 tumors had extensive long sweeping fascicles. In tumors without recurrence, 8 were ALLRR, 5 were ALLE, 2 were SMTLMP. Two others had focal myxoid areas, and 1 had focal atypia and indeterminate necrosis. Among tumors without a recurrence, 6 had an infiltrative border, 2 had VSI and 1 with atypical mitoses. None of these 18 tumors had long sweeping fascicles. As controls, 51 other benign leiomyomas were reviewed but none contained fascicles which met our criteria of long sweeping fascicles. Conclusions: The only parameter in our study which was associated with a subsequent recurrence was long sweeping fascicles. Whether the other parameters in combination can predict outcome in cases of problematic spindle cell smooth muscle tumors remains to be evaluated.