New molecular insights into the pathogenesis of lipoblastomas. Clinicopathologic, immunohistochemical, and molecular analysis in pediatric cases

Lipoblastomas can occasionally require further molecular confirmation when occurring outside of the usual age groups or demonstrating unusual morphology. We reviewed 28 lipoblastomas with 16 controls. Lipoblastomas were subdivided into myxoid (n = 7), classic (n = 9), or lipoma-like (n = 12) subtypes. PLAG1 immunohistochemistry, PLAG1 fluorescence in situ hybridization (FISH), and targeted RNA sequencing were performed on formalin-fixed paraffin-embedded tissue. Karyotypes were available in a subset of lipoblastomas (n = 9). Gene rearrangements were identified in 17/25 (68%) lipoblastomas, including PLAG1 (15/25, 60%) and HMGA2 (2/25, 8%). Five novel fusion partners (DDX6, KLF10, and KANSL1L with PLAG1 and EP400 and FGD6 with HMGA2) were found. PLAG1 immunohistochemistry was positive (nuclear, moderate/strong) in myxoid and classic subtypes lipoblastomas with preferential expression in mesenchymal cells within myxoid stroma and fibrous septa and negative in all controls. When comparing PLAG1 immunohistochemistry with molecular testing (FISH and/or RNA sequencing and/or karyotype), concordant results were noted in 13/25 (52%) cases, increasing to 15/25 (60%) after slight adjustment of the PLAG1 FISH positive threshold. In myxoid and classic lipoblastomas, PLAG1 immunohistochemistry seems to be a better surrogate marker for PLAG1 rearrangement, as compared with lipoma-like subtypes. In lipoma-like subtypes, targeted RNA sequencing appears to detect PLAG1 fusions better than FISH and immunohistochemistry. The preferential expression of PLAG1 in the mesenchymal and fibroblast-like cells deserves further investigation as the putative cell of origin in lipoblastoma.