Abstract 1695: Linking clinical molecular profiles of tumors to the electronic medical record

Abstract By linking “real-word” patient data mined from electronic medical records to the results of clinical genomic testing, we can analyze the impact of the rapid expansion in tumor molecular profiling on the management of cancer patients over the past several years. Theoretically, the widespread adoption of comprehensive molecular characterization of patient tumors using next-generation sequencing panels provides new opportunities for advances in precision oncology. For patients, identification of specific alterations can potentially lead to new therapeutic options from an ever-widening array of targeted agents. For example, the recent FDA approval in 2017 of pembrolizumab for patients with unresectable or metastatic, microsatellite instability-high (MSI-H) or mismatch repair deficient (dMMR) solid tumors is the first example of an FDA approval agnostic with regards to the tumor site of origin. The availability of this novel treatment option has likely led to more frequent use of tumor molecular profiling in the clinical setting, and in the future, we can expect similar examples of treatment strategies that favor specific genomic perturbations over the tumor site of origin. However, despite their ever-expanding use in the clinical setting, the overall impact of genomic testing remains unclear, and evidence demonstrating improvement in patient outcomes is lacking. In this study, we present the development of a database that combines clinical genomic results from our internal next-generation sequencing panel, as well as commercially available external clinical genomic panels, with a variety of data points from our institution’s electronic medical record (EMR). This data set allows us to measure the utility of genomic tumor testing for the management of cancer patients. We analyze the magnitude of positive impact for various tumor types, identify barriers to efficient use of genomic tumor testing, and assess for disparities in the utilization of testing. Patient survival data relative to the clinical action of obtaining tumor molecular profiling can also be measured. We find that a small fraction of patients dies before results return and any action can be taken, especially in cases of pancreatic cancer, suggesting that molecular profiling should be considered earlier in the course of disease for certain cancer types. Combining information on patient outcomes from “real-world” EMR data with clinical genomic testing provides the additional opportunity of identifying exceptional responders or non-responders with greater efficiency, which can lead to the discovery of new therapeutic targets and treatments as well as the identification of novel biomarkers for response or resistance to therapy. Citation Format: Debajyoti Datta, Atul Butte, Theodore Goldstein. Linking clinical molecular profiles of tumors to the electronic medical record [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1695.