Abstract Introduction: Early cancer detection with liquid biopsies has the potential to reduce mortality rates by improving screening strategies, adoption, and adherence. Towards this, a deeper understanding of the cancer genomic landscape among patients of different ethnicities is critical to ensure comparable performance across a broad population. In this study, we sought to understand how the mutational landscape of patients with colorectal cancer (CRC) varies across ethnicities. Methods: A total of 20,845 de-identified patients with stage I-IV CRC, who had undergone commercial personalized and tumor-informed ctDNA testing (SignateraTM) as of April 14, 2022, were eligible for inclusion in the study. We performed ancestry inference using Principal Component Analysis on single nucleotide polymorphisms (SNPs) and validated the predictions in a subset of 2,032 patients with whole exome sequencing data available along with self-reported ethnicity. Individuals were assigned to one of four major ethnic groups: African (AFR), European (EUR), East Asian (EAS) and South Asian (SAS). Results: Our cohort consisted of 69.1% EUR, 9.6% AFR, 20% EAS, and 1.3% SAS patients. On comparing the somatic driver mutations across ancestry groups, we observed APC mutations in 71% of EUR, 79% of AFR, 80% of EAS and 69% of SAS (p<0.0001), KRAS mutations in 39% of EUR, 50% of AFR, 40% of EAS and 38% of SAS (p<0.0001), and TP53 mutations in 64% in EUR, 64% in AFR, 74% in EAS and 69% in SAS (p<0.0001) patients, respectively. The microsatellite instability (MSI) rates similarly varied across ancestral groups, with EUR ancestry (12.3%) having a significantly higher rate of MSI than AFR (8.9%) and EAS (7.5%) (p<0.0001). Comparing the mutational patterns in left vs. right colon across 712 patients with available tumor location information, we found 64% of tumors in EAS to be left sided vs. 49% in EUR (p=0.0002). Left-sided tumors were enriched for MSI and RAS mutations and right-sided tumors were enriched for APC and TP53 mutations, confirming known differences in left- vs. right-sided colorectal tumors. Among left sided tumors, we also observed TP53 mutations in 79% of EAS as compared to 64% of EUR (p=0.001), APC mutations in 84% of EAS as compared to 70% of EUR (p=0.002) and RNF43 mutations in 1% of EAS as compared to 7% in EUR (p=0.002) suggesting differences in mechanisms of WNT pathway inactivation across ethnicities. Finally, we looked at the predicted performance of a cancer screening panel which is based on large and heterogeneous data and maximizes recurrent mutation coverage while minimizing panel size. Despite the differences in individual driver gene mutation frequency, we found that the mean number of mutations covered by the panel in AFR, EAS and SAS individuals relative to EUR was 1.07, 1.1 and 1.02, respectively. Conclusions: Our results demonstrate that the expected performance of a theoretical cancer screening panel was similar across the different ancestry groups despite differences in their somatic landscape. Citation Format: Preethi Srinivasan, Boris Gutman, Vasily Aushev, Shruti Sharma, Taly Arbel, Hsin-Ta Wu, Meenakshi Malhotra, Raheleh Salari, Bernhard Zimmermann, Ryan Swenerton, Trupti Kawli, Breeana Mitchell, Matthew Rabinowitz, Alexey Aleshin, Johannes G. Reiter. Mutational heterogeneity of colorectal cancers across ancestries and its implications for cancer early detection. [abstract]. In: Proceedings of the AACR Special Conference: Precision Prevention, Early Detection, and Interception of Cancer; 2022 Nov 17-19; Austin, TX. Philadelphia (PA): AACR; Can Prev Res 2023;16(1 Suppl): Abstract nr P040.
Read full abstract