Abstract B65: Measuring intratumoral heterogeneity with a mutant-allele frequency score across distinct cancer types

Intratumoral heterogeneity (ITH) is a phenomenon where a tumor, over space and time, branches into a set of genetically distinct clones, which have unique somatic branch mutations, although sharing several somatic trunk mutations. Clonal populations expand or regress in response to selective pressures exerted by the tumor microenvironment. Therefore, ITH has already been linked to several cancer features such as the development of metastases and resistance against a broad range of therapies. Through whole exome sequencing, we can access the number of reads covering the alleles of somatic mutations in tumor samples. Thus, mutant-allele frequency (MAF), i.e., the proportion of reads reporting the mutant allele over all reads mapped to the locus, is informative about the number of clones within a tumor bearing this genetic variant. However, MAF estimations can be impaired by tumor purity, copy number alteration, or sampling approaches. It is proposed that the analysis of multiple samples of a tumor improve accuracy in quantifying clone populations, but single sample is cheaper and more feasible for the clinicians. Mutant-Allele Tumor Heterogeneity (MATH) is a method capable of trustworthily scoring ITH in single samples, as it is based on the range of the MAF distribution of a tumor. The higher the MATH score, the more heterogeneous is a tumor. The goal of this project is to evaluate the ITH levels of 6,390 tumor samples of 20 distinct tumor types, for which genomic data is publicly available through GDAC Firehose repository. We accessed the MAF files and filtered somatic variants: (i) selected only single nucleotide variants, (ii) mutant alleles reported by at least three reads, (iii) total coverage equal to or higher than 30 reads. We estimated MATH scores using R (v3.2.4). The distribution of MATH scores across distinct tumor types exhibited a wide variability (ranging from 12 to 128). Uterine corpus endometrial carcinoma had the lowest MATH scores, whereas prostate adenocarcinoma and testicular germ cell tumors presented the highest MATH scores. The wide ITH variability among nine distinct tumor types has already been assessed in previous studies by using another scoring methodology. High MATH score was significantly associated with decreased overall survival in head and neck squamous cell carcinoma and predicted the development of metastasis in individuals with colorectal cancer. We will further explore the correlation between MATH scores and clinical data to all the 20 tumors here described by exploring the corresponding public clinical data from the cBioPortal for Cancer Genomics. Should MATH score prove to be an optimal scoring method for single biopsies, capable of predicting patient9s prognosis, drug response, or disease outcome, it will be a valuable tool for clinicians for a broad spectrum of cancer treatments. Citation Format: Cibele Masotti, Filipe F. Santos, Anamaria A. Camargo, Pedro A.F. Galante. Measuring intratumoral heterogeneity with a mutant-allele frequency score across distinct cancer types [abstract]. In: Proceedings of the AACR International Conference held in cooperation with the Latin American Cooperative Oncology Group (LACOG) on Translational Cancer Medicine; May 4-6, 2017; Sao Paulo, Brazil. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(1_Suppl):Abstract nr B65.