Analysis of genomic alterations and treatment landscape in patients with advanced tumors using real-world data for precision oncology

6584 Background: Analysis of real-world data is critical for clear understanding of treatment effectiveness. Here, we explore the clinical utility of next-generation sequencing (NGS) and the impact of mutational landscapes on targeted therapy and outcomes in almost 10,000 patients in real-world setting. Methods: We constructed a data analysis platform that integrated NGS results with machine and manually curated electronic medical records data from a single large healthcare system. We assessed eligibility for a targeted therapy based on the constellation of mutations detected by NGS, investigated whether a matched therapy was given and how this affected patient outcome. Results: The study comprised of 9848 patients with NGS results across 90 cancer types including lung (2208;22%), colorectal (1495;15%), multiple myeloma (1153;12%), breast (783;8%), and prostate (480;5%) and other remaining cancers (3,729; 38%) from 2011 to 2021. Overall, 95% of patients had a positive NGS result, of whom 6364 (65%) had a targetable mutation in 34 genes per any NCCN guidelines in any cancer type. Of these, 2930 (46%) were late-stage (3-4). 820 patients across 31 cancer types received any of 54 targeted therapies based on an NGS result; 605/820 (74%) were stage 3-4; overall 605/2930 (21%) of all tested stage 3-4 patients received targeted therapy. Overall stage 4 patients with a targetable mutation who were treated with a targeted therapy had better 5-year overall survival (OS) than patients who did not (HR = 0.71, 95% CI 0.57-0.89, p = 0.003). We further analyzed important biomarkers in specific cancers. Stage 4 lung cancer patients carrying EGFR, ALK, ROS1 or MET alteration and receiving targeted therapy had better 5-year OS (HR 0.56, 95% CI 0.40-0.80, p = 0.001). We observed that multiple myeloma patients carrying KRAS and NRAS mutation received off-label Trametinib, however it did not improve their OS (HR 1.28, 95% CI 0.76-2.15, p = 0.4). Moreover, there are 665 patients with tumor mutational burden (TMB) across 64 cancer types. TMB has been categorized into high (55, 8%), intermediate (177, 27%) and low (433, 65%). OS was significantly longer in TMB-low patients compared to TMB-high and TMB-intermediate patients across all cancers (HR 0.63, 95% CI 0.43-0.92, p = 0.016). There are 5051 patients with microsatellite instability status across 72 cancers categorized into high (828, 16%), and low (4223, 84%) with no significant survival difference between groups. Conclusions: We show that real-world data offers important insights into clinical practice and patient outcome, with respect to genomic alterations and associated targeted therapies. Moreover, we have built an analytics platform which allows users to ask relevant clinical questions around outcome of patients treated with targeted therapy by genomics biomarker and potential to identify off-label use and clinical response in any cancer type.