An Updated Conceptual Framework for Myoepithelial Tumors of Soft tissues and Bone: Toward a Molecularly Informed Classification

Abstract

Myoepithelial tumors (MET) of soft tissue and bone comprise a rare group of neoplasms unified by partially overlapping morphology and so called myoepithelial immunophenotype. Historically, MET have long posed diagnostic and prognostic challenges. Grading and risk stratification have relied largely on the presence of cytologic atypia. Recent molecular and epigenetic studies have fundamentally revised this concept, demonstrating that MET represent a biologically heterogeneous family rather than a single disease entity. Soft tissue (mostly deep-seated) and osseous MET frequently harbor recurrent gene fusions, most commonly involving FET family genes (EWSR1 or less often FUS) with partners such as POU5F1, PBX1, PBX3, KLF15, KLF17 , and ZNF444 , and more rarely non-FET fusions including SS18::POU5F1 . These fusion types correlate with reproducible clinicopathologic patterns and, in emerging outcome datasets, with subtype-specific differences in behavior. In contrast, superficially located adnexal tumors with ductal differentiation - representing true cutaneous mixed tumors/myoepitheliomas - typically lack EWSR1/FUS rearrangements and instead show PLAG1 rearrangements, supporting a bona fide myoepithelial origin and close relationship to PLAG1 -driven salivary gland counterparts. Additional complexity arises from SMARCB1-deficient, fusion-negative tumors and a small subset lacking identifiable recurrent drivers, as well as substantial overlap in morphology and immunophenotype with multiple MET mimics, contributing to diagnostic misclassification when using morphology and immunohistochemistry alone. To address these issues, we synthesize clinicopathologic, molecular, methylomic and pooled outcome data across major MET subgroups from recent multi-institutional cohorts, highlighting pronounced epigenetic and clinical heterogeneity and providing practical diagnostic guidance for surgical pathologists. We propose a molecularly informed classification framework that improves diagnostic precision, clarifies terminology - particularly distinguishing PLAG1 -rearranged cutaneous salivary-gland analogs from fusion-associated soft tissue/bone sarcomas with myoepithelial-like phenotype - and lays a foundation for refined prognostic stratification and future therapeutic studies.