247P Real-world evidence of immune checkpoint inhibitor treatment in lung cancer patients from a Belgian multicenter study

Immune checkpoint inhibitors (ICIs) have shown survival benefits in multiple clinical trials in patients with lung cancer. To gain further insights into real-world treatment patterns of lung cancer patients treated with ICIs, we performed a multicenter study. We analyzed anonymized electronic health records from ten different data sources in three Belgian hospitals in a cohort of lung cancer patients treated with ICIs between March 2017 and August 2022. A total of 597 variables were analyzed, mapped to SNOMED-CT, and OMOP-CDM databases were created. We utilized machine learning and natural language processing (NLP) to process records and algorithm outputs were validated against a physician-generated standard. We detected 731 lung cancer patients with a mean (SD) age of 67 (10) years and predominantly male (67%). Smoking status was: current, 47.7%; former, 25.9%; never, 8.6%; unknown, 17.8%. The most common ICI treatment (monotherapy or with other ICIs or antineoplastic drugs) was pembrolizumab (62.4%), followed by nivolumab (15.4%), atezolizumab (12.4%), and durvalumab (9.4%). Within the patients with available performance status (PS; n = 405, 55.4%), 21.5% had PS 0, 56.3% had PS 1, 10.7% had PS 2, and 1.6% had PS 3-4. Lung cancer was metastatic in 60.6% of patients, non-metastatic in 22.7%, and of unknown status for 16.7%. Preliminary analyses show a median overall survival (OS, 95% confidence intervals) of 30 (26, 35) months across all lung cancers and types of therapies. These preliminary results demonstrate the feasibility of autonomously extracting and locally validating federated hospital databases—a crucial tool for accessing real-world data on patients with lung cancer receiving ICI treatment. Analyses focusing on immune-related adverse events, comorbidities, tumor stage, anatomical pathology, and treatment lines and histology are ongoing.