Real-world biomarker testing trends in metastatic non-small cell lung cancer (NSCLC).

e18728 Background: NSCLC is now recognized as not one disease, but many—differentiated by an increasing number of molecular targets. About 50% of patients with advanced nonsquamous NSCLC have a targetable driver mutation. We retrospectively assessed biomarker testing trends in a large community oncology practice to identify evolving real-world testing rates and patterns in the areas of methodology, specimen types, and vendors. Methods: Natural-language processing queries were used to extract data from biomarker testing reports of patients diagnosed with stage IV NSCLC in 2015-2019. Tests were included if done within 1 year of diagnosis and were classified as NGS panels (20+ genes) or targeted single-gene tests. Results: Records of 587 patients were included. Over the 5-year study window, testing rates increased for all 6 mutations analyzed (Table). In 2015, only ALK, ROS1, and EGFR were profiled in our data (n = 20 patients), all using targeted single-gene tests. The use of multigene panels increased steadily over time, constituting the majority by 2019. MET, RET, and NTRK were tested almost exclusively using NGS panels (data beginning in 2016). Tissue-only specimens were used for 93.8% of patients diagnosed in 2015, while tissue plus blood were used in 6.3%. Blood-only samples began appearing in our data in 2016, making up 7% of patients. By 2019, the use of blood tests had increased substantially, to 47% of patients, although tissue-only remained the most common type (54%). While blood testing has become more popular, it is often still performed with tissue tests. The primary testing methods used in 2015-2017 were FISH for ALK and ROS1; NGS for MET, NTRK, and RET; and Sanger or NGS for EGFR. By 2018, NGS had become the predominant test method across all 6 biomarkers. During 2015-2016, several vendors were used for testing, with a very segmented market for ALK and EGFR tests. By 2019, the market had aligned to a few key vendors for the majority of tests. In 2016, the top 5 vendors did 76% of 180 total tests, which increased to 93% of 1880 tests in 2019. Almost all blood testing in our data was performed by a single vendor. Conclusions: Testing rates for ALK, ROS1, and EGFR were more consistent than those for MET, RET, and NTRK throughout the study window, possibly because of earlier availability of tests and treatments for these biomarkers. NGS was more available by the time MET, RET, and NTRK became established as actionable biomarkers, explaining the earlier use of panels. A limitation is that nearly 75% of the data are from the final 2 years. These results, however, represent notable testing trends from a large practice and illustrate the expanding application of molecular testing in community oncology.[Table: see text]