Identification of Subsets of Genetic Alterations in KRAS-mutant Lung Cancer using Association Rule Mining

Background Lung cancer is the leading cause of all cancer death accounting for 1 out of 4 cancer-related death in both men and women. KRAS mutations occur in ~ 25% of patients with lung cancer, and the presence of these mutations is associated with poor prognosis. Efforts to directly target KRAS or associated downstream MAPK or the PI3K/AKT/mTOR pathways have seen little or no benefits. One probable reason for the lack of progress in targeting KRAS-mutant tumors is the co-occurrence of other cell survival pathways and mechanisms. Method and results To identify other potential cell survival pathways in subsets of KRAS-mutant tumors, I performed unsupervised machine learning on somatic mutations in metastatic lung cancer from 725 patient samples. I identified 67 other genes that were mutated in at least 10% of the samples with KRAS alterations. This gene list was enriched with genes involved in the MAPK, AKT and STAT3 pathways, cell-cell adhesion, DNA repair, chromatin remodeling, and the Wnt/beta-catenin pathway. I also identified 160 overlapping subsets of 3 or more genes that code for oncogenic or oncosuppressive proteins that were mutated in at least 10% of KRAS-mutant tumors. Conclusions In this study, I identified genes that are co-mutated in KRAS-mutant lung cancer. I also identify subpopulations of KRAS-mutant lung cancer based on the set of genes that were also altered in the tumor samples. The design of research models that captures these subsets of KRAS-mutant tumors would enhance our understanding of the disease and facilitate personalized treatment for lung cancer patients with KRAS alterations.