AI-Driven Precision Targeting Beyond KRAS: Daraxonrasib, Pan-RAS(ON) Inhibition, and Emerging Predictive Biomarkers in RAS-Driven Cancers

Abstract

The high level of therapeutic resistance, pathway activation and variation in Rat Sarcoma (RAS) mutations makes the treatment of RAS-driven cancers a significant challenge. While targeted therapy through allele-selective Kirsten Rat Sarcoma (KRAS) inhibitors has made a significant leap, the limited number of mutations covered by these inhibitors and the potential for reactivation of the pathway indicate the need for more comprehensive approaches. This review focuses on daraxonrasib (RMC-6236), an orally available, noncovalent panRAS(ON) inhibitor that targets active GTP-bound RAS through a cyclophilin A–mediated ternary complex. It is active against a wide range of RAS variants, offering a possible strategy to break through the problem of being blocked by selective inhibition of individual mutations. A review of emerging pan-RAS inhibitors, their therapeutic potential and issues of toxicity and acquired resistance are also discussed. Concurrently, the role of artificial intelligence (AI) in the precision oncology is explored, with special attention on AI for better diagnosis, prognosis, and prediction of treatment response. New predictive biomarkers such as RNA signatures or multi-omics could be used to improve patient stratification and to identify those most likely to respond to RAS intervention. The combination of pan-RAS(ON) inhibition and the discovery of biomarkers using AI may therefore contribute to the development of personalized therapy for RAS driven malignancies. In summary, these results suggest that pan-RAS(ON) inhibition, especially with the use of AI-powered biomarker strategies, is a promising approach for extending RAS targeting, better patient selection, and more targeted treatment of RASdependent malignancies.