The germline and somatic landscape of uterine leiomyosarcoma.

e17619 Background: Uterine leiomyosarcoma (LMS) is a rare, aggressive cancer arising from uterine soft tissue. Surgical cytoreduction is the cornerstone of management in amenable disease, but the recurrence rate is 50-70%. Treatment involves cytotoxic chemotherapy with or without radiation therapy. LMS is genetically heterogeneous and demonstrates chromosomal fusions and other genomic changes, which may make them susceptible to novel targeted therapies. Our aim was to characterize the genetic landscape LMS using a national database to identify changes that may enable use of alternative targeted therapies. Methods: We queried the Myriad Collaborative Research Registry, a national database that includes over 1.1 million patients who have undergone germline and/or somatic tumor testing, which categorizes variants according to the 4-tiered system of the American Society of Clinical Oncology: I are variants of strong clinical significance, II are variants of potential clinical significance, III are variants of unknown clinical significance, and IV are benign or likely benign variants. IA variants are those for which there exists an FDA-approved therapy. IB indicates a variant for which treatment is suggested by well-powered studies and consensus in the field. IIC variants are those for which there are FDA-approved therapies for different tumor types or investigational therapies backed by small clinical studies and some consensus. IID variants are those for which there are pre-clinical trials or case reports. Tumor mutational burden (TMB), PD-L1 status, and microsatellite instability (MSI) were also examined. Results: Among 37 patients, 158 somatic variants were identified: 10 IA, 130 IIC, and 18 IID. IA included mutations in BRCA2 , MSH2 , PMS2 , and MSH6 . IIC included mutations in the PTEN pathway, SWI/SNF complex genes, and TP53 . High TMB (≥10 mutations/Mb) was found in 8 patients. 8 patients were PD-L1 positive. 6 patients were MSI-H. Of those with high TMB, 4 were PD-L1 positive and 6 were MSI-H. Of the patients with high TMB who were PD-L1 positive, 2 were MSI-H. PD-L1 status was not available in 7 patients. 1 patient was found to have at least 20 mutations, most significantly, 2 in each of BRCA2 and MSH2 . This individual was found to have high TMB and was MSI-H. 11 germline mutations were identified in 6 individuals, including 5 deleterious mutations in MUTYH . Conclusions: The rarity, aggressiveness, and genetic heterogeneity of LMS make it difficult to study and determine standard-of-care on a large-scale. Our data suggest that routine genetic testing may provide potential novel treatment routes. For example, TMB-H, PD-L1 positive, or MSI-H statuses suggest a possible role for checkpoint inhibitors. Additionally, mutations in double strand DNA repair genes and SWI/SNF genes provide other novel treatment routes. These data suggest that targeted therapies directed towards tumor genotype should be considered in addition to standard-of-care regimens.