Gamma secretase inhibitors and desmoid fibromatosis: Lessons from a real world, comprehensive genomic study of desmoids and <i>CTNNB1</i> / <i>APC</i> mutated soft tissue tumors.

Abstract

11547 Background: Recently, first-in-class FDA approval was granted in the U.S. for use of the gamma secretase inhibitor (GSI), nirogacestat, for adults with progressive desmoid fibromatosis. In tandem, the unpredictable clinical behavior of desmoids, which ranges from local aggression to regression, raises consideration of whether diagnostic and molecular variability underlie their varying biological potential. With an aim to understand both, and to explore the potential for biomarkers for GSI therapy selection, prediction, and prognosis, we performed a retrospective review of comprehensive genomic profiling and histology of desmoids and other soft tissue tumors harboring CTNNB1 or APC mutations. Methods: Using real-world reference laboratory database of tumors submitted for clinical genomic assessment (Caris Life Sciences), we queried for samples with a diagnosis of desmoid fibromatosis, or for other neoplasms of soft tissue origin harboring CTNNB1 or APC mutations. Samples underwent next-gen sequencing of (whole exome) to identify gene variants/copy number alterations and of RNA (whole transcriptome) for expression and fusion profiling. Findings were correlated with available clinical data and whole slide image histologic review. Results: We identified 74 tumors submitted as desmoid fibromatosis, of which 80% harbored CTNNB1 and 15% harbored APC pathogenic or likely pathogenic variants. CTNNB1 variants included codon 41 (58%), codon 45 (41%), and ubiquitin motif codon 36 (1%), while 91% of APC variants detected were in exon 16. Recurrent co-alterations were rare, involving MUTYH (heterozygous G396D) in 2 samples, and TMB-High (≥10 mutations/Mb) present in 3 . Notably, 4 “desmoids” (5%) lacked characteristic mutations, one of which harbored COL1A1::USP6 fusion, reclassified as nodular fasciitis. Among 76 soft tissue tumors diagnosed as other entities at analysis but found to harbor CTNNB1/APC mutations, 6 (all limited core biopsies), could be confidently reclassified as desmoids. The remaining 70 CTNNB1/APC mutant neoplasms were diverse, including synovial sarcoma (11%) and rhabdomyosarcoma (10%). Conclusions: Correlation of genomics and histopathology may allow identification of other tumor types misclassified as desmoid fibromatosis. Conversely, genomic correlation facilitated recognition of additional desmoids among tumors submitted with other diagnoses. The striking lack of secondary mutations seen in this large cohort with comprehensive DNA sequencing implies that other mechanisms explain and could predict their variable behavior, for which we are exploring paired transcriptome profiling data. Finally, subsets of diverse, other soft tissue neoplasms harbor CTNNB1 or APC mutations, which may have implications for the design of future biomarker-selected Phase II basket trials.