Utility of tumor genomic profiling in guiding targeted therapy and referral for clinical trials in the community practice.

e18381 Background: Next-generation sequencing (NGS), which enables tumor genomic profiling with limited tumor specimen for the guidance of targeted therapy, has been widely used in oncology practice. It also reports mutations of which no approved therapy is available, but potential therapies approved in a different tumor type may be reasonable to try, based on molecular mechanisms or limited data, either in clinical trials or off trial treatment. The utility of this information in the community practice is unclear. Methods: We performed a retrospective analysis on the usage of NGS result on patients with lung or colorectal cancers treated between November 2011 and February 2018. Patients were identified from Foundation Medicine Company database and were linked to records at Maimonides. We evaluated if NGS was used to guide FDA-approved targeted therapy, potentially-useful targeted therapy not approved for the cancer type, or referral to a clinical trial. Results: 177 patients (lung ca = 119, colon ca = 58) were included. NGS identified 34 (28.6%) lung ca patients with driver mutations, (21 EGFR, 6 RET, 3 ALK, 2 MET amplification, 1 ROS1 and 1NTRK mutation), who were all given FDA-approved therapy. 70 (58.8%) patients had at least 1 target with an FDA-approved therapy for a different cancer; 89 (74.9%) had a mutation being studied in a clinical trial, and 48 patients were eligible for the NCI-MATCH trial. None received non-FDA-approved drugs and none was referred to clinical trials. In the colon cancer cohort, NGS identified alterations in KRAS (27), BRAF (5), ERBB2 mutation (2), ERBB2 amplification (1), MLH1 (1), MSH2 (1) and BRCA 2(1). Among them, one patient received BRAF inhibitors. 45 and 24 patients were eligible for phase I/II trials and NCI-MATCH trial respectively, and none was referred. Conclusions: NGS has a high efficiency of detecting driver mutations in lung cancer; but only reveals low frequencies of alterations otherwise not tested in colon cancer. The approach of prescribing un-approved targeted treatment based on theoretical mechanism of action was very uncommon, and the referral to clinical trials was rare in this community practice, both of which decreasing the utility value of NGS.