Teapot, A Quantitative Approach To Address Intra-Tumor Heterogeneity By Identifying Driver Mutations In Cancer Cell Subpopulations.

e24215 Background: Therapeutic efficacy is often impeded by drug resistance arising from ubiquitous intra-tumor genetic heterogeneity. However, commonly used tests for identification of driver mutations, defined as mutations in cancer genes, report only the presence of driver mutations in a tumor, largely ignoring the mutation’s distribution in cancer cell subpopulations. The most effective cancer treatments today should rely upon a personalized and precision approach that take into account of intra-tumor heterogeneity. Methods: We have developed a novel approach, TEAPOT (Tumor Evolution Assay for Personalized Oncology Therapy), to reconstruct a tumor’s evolutionary history through combination of whole exome sequencing data from a bulk primary tumor and 16-24 microsamples taken from the bulk tumor. The evolutionary history for an individual tumor is expressed as a rooted tree representing the mitotic process of tumor development starting from an ancestral cancer cell. Individual mutations are assigned to the cells in which they first present. The size of each offspring carrying a specific mutation is estimated based on tumor purity, variant allele frequency, and the variant’s copy number. Results: By re-building a tumor’s evolutionary history, TEAPOT depicts the occurrence of driver mutations in the most recent common ancestor cells for a subpopulation. The size of offspring carrying a driver mutation can thus be estimated. Studies in melanoma, ovarian, and lung tumors revealed that driver mutations are often present in minority clones. Most tumors do not have a driver mutation at the ancestor of the entire cancer cell population. More than 50% of tumors have two or more driver mutations and > 50% of driver mutations are present only in a subpopulation with a size of less than 50% of the total cancer cell population. Conclusions: We developed a novel approach, TEAPOT, to define intra-tumor heterogeneity by estimating the intra-tumor prevalence of individual driver mutations. Such information may assist in selection of an effective targeted agent and provide the rationale for cocktail treatments targeting multiple driver mutations simultaneously.