Abstract 3409: MSK-IMPACT Heme: Validation and clinical experience of a comprehensive molecular profiling platform for hematologic malignancies

Background: As the repertoire of molecular targeted therapies for hematologic malignancies continues to expand, so too does the opportunity for molecular profiling to inform treatment decisions. While mutations in certain genes, such as JAK2, MPL, MYD88 and BRAF have diagnostic utility, others such as FLT3, NPM1, IDH1, IDH2, DNMT3A, KIT and CEBPA have prognostic value. Here, we present the development and clinical experience of MSK-IMPACT Heme (Integrated Mutation Profiling of Actionable Cancer Targets for Hematologic malignancies), a comprehensive molecular profiling platform, utilizing hybridization capture and high coverage next generation sequencing of paired tumor and normal tissues. Methods: We designed custom DNA probes corresponding to all exons of 400 key oncogenes and tumor suppressor genes implicated in hematologic malignancies, including all genes that are targetable by approved and experimental therapies being investigated in clinical trials at our institution. The accuracy, precision, and sensitivity of MSK-IMPACT Heme was assessed on a validation set of 113 unique tumor samples with known SNVs and indels previously confirmed by orthogonal methods. We implemented a custom analysis pipeline to integrate the analysis of any number of normal samples with a given tumor and provide a reliable assessment of somatic alterations, even in post-transplant chimeric patients. The selection of matched nail, saliva, and/or blood tissue was determined at the time of test initiation as indicated by patient diagnosis and transplant history. The ability to detect somatic copy number alterations was demonstrated with samples previously characterized by SNP array platforms. Results: We sequenced 821 tumor samples, from 759 patients that represented over 50 tumor types to a mean depth of 758X. 429 patients were male (56.5%) and 20 cases were post allogeneic stem cell transplantation. The most common tumor types sequenced were Follicular lymphoma (11.9%), DLBCL (11.3%), and AML (11.0%). We identified 4,935 mutations from 732 samples. The most commonly altered genes were TP53, KMT2D, and CREBBP. Implementation of the MSK-IMPACT Heme workflow enabled the characterization of complex tumor specimens, including sorted cells and tumor samples from post-transplant chimeric patients. The joint utilization of matched patient and donor normal tissues enabled differentiation between somatic alterations and both host and donor derived common polymorphisms. Conclusions: The MS