The Expanding Family of SMARCB1(INI1)-deficient Neoplasia: Implications of Phenotypic, Biological, and Molecular Heterogeneity

Since the description of atypical teratoid/rhabdoid tumors of the central nervous system and renal/extrarenal malignant rhabdoid tumors in children, the clinicopathologic spectrum of neoplasms having in common a highly variable rhabdoid cell component (0% to 100%) and consistent loss of nuclear SMARCB1 (INI1) expression has been steadily expanding to include cribriform neuroepithelial tumor of the ventricle, renal medullary carcinoma and a subset of collecting duct carcinoma, epithelioid sarcoma, subsets of miscellaneous benign and malignant soft tissue tumors, and rare rhabdoid carcinoma variants of gastroenteropancreatic, sinonasal, and genitourinary tract origin. Although a majority of SMARCB1-deficient neoplasms arise de novo, the origin of SMARCB1-deficient neoplasia in the background of a phenotypically or genetically definable differentiated SMARCB1-intact parent neoplasm has been convincingly demonstrated, highlighting the rare occurrence of rhabdoid tumors as double-hit neoplasia. As a group, SMARCB1-deficient neoplasms occur over a wide age range (0 to 80 y), may be devoid of rhabdoid cells or display uniform rhabdoid morphology, and follow a clinical course that varies from benign to highly aggressive causing death within a few months irrespective of aggressive multimodality therapy. Generally applicable criteria that would permit easy recognition of these uncommon neoplasms do not exist. Diagnosis is based on site-specific and entity-specific sets of clinicopathologic, immunophenotypic, and/or molecular criteria. SMARCB1 immunohistochemistry has emerged as a valuable tool in confirming or screening for SMARCB1-deficient neoplasms. This review summarizes the different phenotypic and topographic subgroups of SMARCB1-deficient neoplasms including sporadic and familial, benign and malignant, and rhabdoid and nonrhabdoid variants, highlighting their phenotypic heterogeneity and molecular complexity.