ENLIGHT: Pancancer response prediction to targeted and immunotherapies via tumor transcriptomics.

e13556 Background: Precision oncology is gradually advancing into mainstream clinical practice. Despite the significant recent growth in the number of approved biomarkers for immune and targeted therapies, demonstrating significant survival benefits, eligibility and response rates remain limited in many cases, calling for better predictive biomarkers. Methods: We developed ENLIGHT - a transcriptomics-based computational platform that identifies and utilizes clinically relevant genetic interactions (GIs) to predict a patient’s response to cancer treatments. ENLIGHT markedly extends and improves upon SELECT, a previous GI-based framework for obtaining transcriptomics-based response biomarkers. In this study, we focus on three key translational aspects: (i) the number of drugs for which predictions can be obtained (ii) defining a biomarker-based test for Personalized Medicine (PM) that would identify favorable treatments for an individual; and (iii) Improving clinical trial design by excluding a sub-population of patients likely not to respond to the treatment. A key translational feature of the approach is that it does not require training on treatment response data. Thus, in addition to its ability to predict patients' response to approved and well-studied therapies, it can predict the response to new, unexplored treatments. Results: We first tested Enlight in the PM scenario, analyzing 21 patient cohorts from diverse indications, treated with a variety of targeted and immunotherapies. The ENLIGHT treatment matching score is associated with better response with an aggregate Odds Ratio (OR) of 2.59 (95% confidence interval [1.85, 3.55], p= 3.41 e-8). Applied to the WINTHER trial data, encompassing multiple indications and individualized treatments, ENLIGHT recommendations achieved a highly remarkable OR of 11.15 ([2.3, 54.5], p = 8 e-04), demonstrating ENLIGHT’s strong predictive power across a broad spectrum of treatments and cancer types. Second, using ENLIGHT to exclude patients from clinical trials increases the trial response rate, achieving more than 90% of the response rate attainable under an optimal exclusion strategy. Conclusions: ENLIGHT is a powerful transcriptomics-based precision oncology pipeline developed by Pangea Biomed that broadly predicts response to both extant and novel targeted and immune therapies in translationally oriented, clinical terms, going beyond case-specific biomarkers.