A Comprehensive Clinicopathologic and Molecular Reappraisal of GLI1-altered Mesenchymal Tumors with Pooled Outcome Analysis Showing Poor Survival in GLI1- amplified Versus GLI1-rearranged Tumors

Abstract

GLI1-altered mesenchymal tumor is a recently described distinct pathologic entity with an established risk of malignancy, being defined molecularly by either GLI1 gene fusions or amplifications. The clinicopathologic overlap of tumors driven by the 2 seemingly distinct mechanisms of GLI1 activation is still emerging. Herein, we report the largest series of molecularly confirmed GLI1-altered mesenchymal neoplasms to date, including 23 GLI1-amplified and 15 GLI1-rearranged new cases, and perform a comparative clinicopathologic, genomic, and survival investigation. GLI1-rearranged tumors occurred in younger patients (42 vs. 52 y) and were larger compared with GLI1-amplified tumors (5.6 cm vs. 1.5 cm, respectively). Histologic features were overall similar between the 2 groups, showing a multinodular pattern and a nested architecture of epithelioid, and less commonly spindle cells, surrounded by a rich capillary network. A distinct whorling pattern was noted among 3 GLI1-amplified tumors. Scattered pleomorphic giant cells were rarely seen in both groups. The immunoprofile showed consistent expression of CD56, with variable S100, CD10 and SMA expression. Genomically, both groups had overall low mutation burdens, with rare TP53 mutations seen only in GLI1-amplified tumors. GLI1-amplified mesenchymal tumors exhibit mostly a single amplicon at the 12q13-15 locus, compared with dedifferentiated liposarcoma, which showed a 2-peak amplification centered around CDK4 (12q14.1) and MDM2 (12q15). GLI1-amplified tumors had a significantly higher GLI1 mRNA expression compared with GLI1-rearranged tumors. Survival pooled analysis of current and published cases (n=83) showed a worse overall survival in GLI1-amplified patients, with 16% succumbing to disease compared with 1.7% in the GLI1-rearranged group. Despite comparable progression rates, GLI1-amplified tumors had a shorter median progression-free survival compared with GLI1-rearranged tumors (25 mo vs. 77 mo). Univariate analysis showed that traditional histologic predictors of malignancy (mitotic count ≥4/10 high-power fields, presence of necrosis, and tumor size ≥5 cm) are associated with worse prognosis among GLI1-altered mesenchymal tumors.