(P 177) GENOMIC LANDSCAPE OF RARE EWSR1 FUSIONS

Objective: The advancement of molecular profiling in cancer has identified gene fusions as an oncogenic driver across malignancies. Fusions involving Ewing’s sarcoma RNA Binding Protein 1 (EWSR1) have been identified as drivers of a diverse set of malignancies including Ewing’s sarcoma, neuroectodermal tumors, and various other tumors. EWSR1 has a large spectrum of fusion partners resulting in malignancies with different phenotypes and sub-types partially related to the EWSR1 fusion partner. While many malignancies harboring the common EWSR1 fusion partners have been genomically characterized, the genomes of rare EWSR1 fusion tumors are yet to be explored. Here, we present the largest assessment of rare EWSR1 fusion partner genes to date and an analysis of genomic variants driving the EWSR1 fusion.

Methods: Clinical and genomic profiling data from EWSR1 fusion positive tumors were collected using Foundation Medicine Incorporated’s (FMI) hybrid-capture-based next-generation sequencing platform. For this study, we defined rare EWSR1 fusion partners as prevalent in < 1% of the dataset. Fusion partners were further classified by transcription factor (TF) status and class as determined using TF Class. All data was analyzed in Rv.4.1.1 or GraphPad Prismv.9.2.0 and p values ≤ 0.05 were considered statistically significant. Chi-squared tests and t-tests were used as appropriate.

Results: A total of 1,140 unique tumors exhibiting an EWSR1 gene fusion were collected from FMI representing 74 unique fusion partners. Most tumors harbored EWSR1 fusion with previously characterized fusion partners including EWSR1-FLI1/ERG (n=447, 39.2%), EWSR1-ATF1 (n=159, 13.9%), EWSR1-WT1 (n=151, 13.2%) and EWSR1-NR4A3 (88, 7.7%). After selecting for tumors harboring EWSR1 fusions with rare fusion partners (prevalence ≤1%), we identified one hundred twenty-seven tumors representing sixty-four rare EWSR1 fusions. Forty-three non-recurrent fusion partners were identified, six fusion partners were identified in two tumors each, and six fusion partners were identified in three tumors each. Twelve rare fusion partners were members of the ETS family of TFs (ETS1, ETV1, ETV4, FEV) and thirteen were CREB-related TFs (CREB3L2-4). Additional rare fusion partners were classified as other transcription factors (Other (TF) – Rare; n=53) and other non-transcription factors (Other (non-TF) – Rare; n=49). Recurrent rare fusions were found in CREB3L2 (n=11), DDIT3 (n=9), TFCP2 (n=8), and ZNF444 (n=6). Clinical characteristics (including sex, age at sequencing, and disease type, and disease location) between common and rare fusion partners of the same TF class were not significantly different. The genomic landscape of rare and common ETS partners, as well as rare and common CREB partners, were similar. Notably, tumors harboring rare, non-TF fusion partners were genomically distinct and demonstrated a high prevalence of potentially targetable genomic variants.

Conclusion: This study of rare EWSR1 fusion partners demonstrates the similarity between rare and common fusion partners in the ETS and CREB partner classes. TF and non-TF fusion partners may represent unique molecular entities. Further research is required to explore the use of target-specific treatment for these rare malignancies.