Impact of RAS mutation subtype on clinical outcome-a cross-entity comparison of patients with advanced non-small cell lung cancer and colorectal cancer

Abstract

Mutated RAS onco-proteins are key drivers across many cancers. The distribution of somatic RAS mutations varies between cancer entities. Retrospective analyses have associated some RAS mutations with distinct clinical outcomes. However, the clinical impact of the full spectrum of RAS mutations in their disease contextuality remains to be defined. To improve upon this situation, we studied genomically and clinically annotated, prospectively recruited cohorts of patients with RAS-mutated metastatic lung cancer and colorectal cancer. Mutational spectra were compared with predictions derived from analyzing the mutagenic impact at the genome level for each entity. Interestingly, we found concordance of predicted signatures with those actually observed in our patients. Thus, composition of the functionally active RAS mutational subtypes is primarily determined by the mutagenic context. Most RAS mutations seemed dominant oncogenic drivers with entity-dependent clinical outcomes. RAS comutations were enriched in tumors harboring class 2/3 BRAF mutations, highlighting the functional dependency of some mutated BRAF isoforms on RAS. With our dataset, we established a probabilistic model for cross-entity comparison of the prognostic impact of specific RAS mutational subtypes. The resulting prognostic clusters showed largely consistent clinical categorizations in both entities. This suggests mutant subtype-specific functional properties leading to similar clinical effects. A notable exception is KRAS G12C, which imparted an adverse prognosis only in colorectal cancer. Our findings provide a framework for risk stratification of specific RAS mutations across several cancer entities, which is required to guide the analysis of clinical findings in patients treated with direct RAS inhibitors or agents targeting downstream pathways.