Abstract LB291: Uncovering gene fusions with 3D genomics: from clinical validation to actionable insights for undiagnosable solid tumors

Abstract Identifying gene fusions in tumor biopsies is critical for understanding disease etiology; however, clinical NGS panels often fail to yield clear genetic drivers. A key challenge is that RNA-seq does not perform well in FFPE tissue blocks due to RNA degradation and/or RNA panel design, and cannot detect breakpoints outside of the gene body, which are established clinical biomarkers with mechanistic significance and clinical utility in solid hematological cancers. We developed a novel DNA-based partner-agnostic approach for identifying fusions from FFPE tumors using 3D genomics based on Arima-HiC technology and NGS. We profiled 184 FFPE tumors across various tumor types. This cohort includes 33 tumors with known gene fusions detected by RNA panels for clinical concordance analysis. It also includes 151 FFPE tumors, including 62 CNS tumors, 59 gynecological sarcomas, and 22 solid heme tumors, with no detectable genetic drivers from prior DNA and RNA panel CLIA assays. For clinical concordance, Arima-HiC technology detected 33/33 fusions previously detected by RNA panels. For clinical validation and utility studies in our driver-negative cohort, Arima-HiC detected 1 or more fusions in 72% (109/151) of tumors. To attribute clinical significance, we compared the genes implicated in our fusion calls with NCCN and WHO guidelines, and OncoKB, and assigned which tumors had an FDA-approved therapeutic level biomarker (“Tier 1”), a biomarker targeted in an ongoing clinical trial (“Tier 2”), or a diagnostic/prognostic biomarker (“Tier 3”). Arima-HiC analysis found 33.8% (51/151) of tumors had Tier 1 biomarkers, 4.0% (6/151) had Tier 2 biomarkers, and a further 14.6% (22/151) had Tier 3 biomarkers, indicating an overall yield of clinically actionable biomarkers at 52.3%. Several cases with Tier 1-3 fusions underwent confirmatory immunohistochemistry testing for oncoprotein expression, and 91.6% (11/12) showed diffuse or focally positive staining. To highlight examples from prospectively analyzed cases from this cohort, we identified a novel PD-L1 rearrangement in a pediatric glioma tumor that was not detected by DNA or RNA panels, which we confirmed by PD-L1 IHC, and the patient was administered pembrolizumab after tumor recurrence and has since exhibited stable disease. We also identified a MYBL1 fusion in a glioma that spared the patient unnecessary chemotherapy post-resection. Our findings provide evidence of clinical concordance, validation, and utility, and underscore the need for 3D genome profiling to increase diagnostic yield by finding clinically actionable fusions in FFPE solid tumors. Citation Format: Anthony Schmitt, Kristin Sikkink, Kristyn Galbraith, Misha Movahed-Ezazi, George Jour, Matija Snuderl. Uncovering gene fusions with 3D genomics: from clinical validation to actionable insights for undiagnosable solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB291.