Development and preliminary validation of a platform for systematic, unbiased pre-screening of non-small cell lung cancer (NSCLC) across three community oncology practices.

e23230 Background: With a rapidly increasing number of candidate drugs and clinical trials in oncology, efficiency in clinical trial execution has become a primary focus area for innovation. Currently, fewer than 5% of oncology patients enroll in clinical trials, and data suggests a lack of diversity in clinical trial patient populations compared to the broader patient demographic. To address these issues, we developed a systematic unbiased pre-screening platform for all patients. This approach holds the promise of hastening clinical trials by enhancing the identification of eligible patients and concurrently promoting diversity among those invited to participate in these trials. Methods: A framework for use of EMR ingested data and standardized abstraction of select 160 variables was developed for patients with NSCLC. An algorithm based on 7 key eligibility criteria was developed to generate an automated assessment of eligibility for a presumed trial. Abstracted data was compared to a medical oncologist’s assessment (gold standard) and the automated tool was validated by comparing eligibility assessment before each visit with that of a medical oncologist. Results: Three community oncology practices in the US participated. Source data-verified values for histology, ECOG PS, treatment history, line of therapy, and current metastatic status were available for > 90% of patients in real-time. Tumor NGS testing rates for actionable alterations exceeded 75%. From Oct 23, 2023 to Jan 14, 2024, 786 NSCLC pts had MD visits. Preliminary validation of the automated tool in 79 patients revealed concordant classification in 76/79 pts (96.2%). Sensitivity, specificity, PPV and NPV were > 90%. Concordance for all variables was 98% between the abstractor and medical oncologist. Conclusions: A platform for systematic, unbiased pre-screening of patients for clinical trials in NSCLC has been developed to support and facilitate oncology sites without imposing financial or resource burdens or affecting routine clinical workflow. Preliminary validation supports high performance. Updated validation data will be presented in the meeting. The high levels of data completeness and quality, including NGS testing rates that are substantially higher than what has previously been reported in community centers, underscore their suitability as clinical trial sites.