Real-World Immuno-Oncology Treatment Patterns And Outcomes In Us Patients With Metastatic Non-Small-Cell Lung Cancer.

e20537 Background: Three immune-oncology (IO) agents are approved by the FDA to treat metastatic NCSLC (mNSCLC). This retrospective observational study aimed to observe treatment patterns and real-world outcomes of mNSCLC patients treated in the community setting, predominantly in the southern region of the U.S. Methods: Electronic health records and charts were reviewed for patients with a diagnosis of mNSCLC initiating an IO between September 1, 2015 to September 30, 2017. Patient data were examined from the date of IO therapy initiation until death, loss to follow-up, or study end date (December 31, 2017), whichever was earliest. This study was hypothesis generating thus no statistical tests were performed. Results: Of the 2,864 patients who met initial study criteria, 750 were randomly chosen for full chart review and 577 met inclusion/exclusion criteria among whom 571 had metastatic NSCLC prior to initiating IO therapy. 464, 97, and 16 patients were initiated on nivolumab, pembrolizumab, and atezolizumab, respectively. Average age was 69.7 years and 65% of patients initiated treatment in 2015 or 2016. At diagnosis, 36% of patients had an Eastern Cooperative Oncology Group performance status of 0 versus 16% at time of therapy initiation. Duration of IO therapy was 5.15 months among all metastatic patients (5.37 months for 1L only) and 4.13. months among those who completed therapy during the study period. Median real world progression free survival (rwPFS) was 5.53 (4.31. 7.20) months for 1L patients and 3.32(2.76, 4.01) months for 2L+. Real world overall survival (rwOS) was 19.04 (13.03, NR) months for 1L patients and 12.01 (8.82, 15.92) months for 2L+. Conclusions: Median rwPFS and rwOS were slightly higher than published clinical trials, yet patients were older and with poorer performance status compared to clinical trial enrollees. Despite the differences in patient characteristics, results suggest that clinical real world outcomes for this population fall within the range of median OS and PFS values observed in immune checkpoint inhibitor trials. Future research should explore the use of EHR, chart and other data sources to confirm benefit and risk in the real world.