Genetic Ancestry Differences in Tumor Mutation Between Early and Average Onset Colorectal Cancer

Authors Brooke Rhead, David Hein, Yannick Pouliot, Justin Guinney, Francisco M. De La Vega, Nina Sanford

The incidence and mortality of early onset (EO) colorectal cancer (CRC), defined as CRC diagnosed prior to age 50, are rising when compared to the declining rates of average onset (AO) CRC. Epidemiologic trends for CRC differ by race/ethnicity, which could be connected to underlying differences in tumor mutation and gene expression. Previous research has used self-reported race/ethnicity categories, which are often missing or uninformative for understanding biological underpinnings of such differences, particularly in highly admixed groups such as Black and Hispanic. Hence, we compared tumor mutation profiles of EO and AO
CRC – within a de-identified dataset of 1,643 and 3,175 patients, respectively – in the context of genetic ancestry groups.

Ancestry informative markers were utilized to estimate ancestry proportion likelihoods for the five continental groups defined in the 1000 Genomes Project: Africa, Americas, East Asia, Europe, and South Asia. Recognizing the complexity of ancestry and race relationships, we imputed several race/ethnicity categories using admixture thresholds based on literature and comparisons with available metadata. Similar numbers of Black and Asian patients were observed in both EO and AO groups (13% and 5-6%), however, more Hispanic/Latino patients were found in the EO subgroup (16%) compared to AO (10%). We assessed the association of genetic ancestry proportions (per 20% increase in each ancestry proportion) with the presence of protein altering somatic mutations or copy number changes in 22 CRC driver genes.

Across all ages of onset, African ancestry was associated with higher odds of somatic mutations in APC (OR=1.11) and KRAS (OR=1.14) and lower odds of BRAF mutations (OR=0.84). East Asian ancestry was associated with higher odds of mutations in TP53 (OR=1.15). Furthermore, we found a novel association between Native American ancestry with higher odds of somatic mutations in the MLH1 gene (OR=1.61), unexplained by misclassified germline variants in this gene (i.e. Lynch syndrome), as we limited our analysis to cases with matched tumor/normal sequencing. When testing for interactions between the effects of onset age (EO vs. AO) and genetic ancestry on gene mutations, we observed a significant interaction (p=0.031) between African ancestry and onset age on APC: African ancestry was associated with mutations in this gene in AO (OR=1.16) but not EO (OR=0.98). When compared to race/ethnicity classes, genetic ancestry provides a more quantitative and precise profile of shared genetic background that can underlie biological race differences in cancer etiology and outcomes in EO and AO CRC.