Somatic and germline heterogeneity across colorectal cancer patient characteristics: Implications for early cancer detection

A deeper understanding of both germline and somatic genomic landscapes across a broad population is critical to ensuring optimal performance of early cancer detection assays for all patients. Here, we investigated the germline and somatic mutational landscape in patients with colorectal cancer (CRC). A total of 2,303 de-identified patients with stage I-IV CRC who had commercial, personalized, tumor-informed ctDNA testing (SignateraTM) ordered were eligible for the study. Ancestry was inferred based on DNA extracted from blood samples using Principal Component Analysis on single nucleotide polymorphisms and assigned to African (AFR), East Asian (EAS), European (EUR) and Southeast Asian (not included in ancestry-specific analyses because of sample size). Proportions of patients with somatic driver mutations and pathogenic germline variants (PGVs) in CRC-associated and genes were calculated overall and across various patient characteristics including age of disease onset (< 45 years: 15.8%, 45-65 years: 49.5%, < 65 years: 34.7%), sex (female, 47.3%), disease stage (I: 5.8%, II: 29.8%, III: 41.1%, IV: 18.0%), microsatellite stability vs instability (MSS: 83.8%, MSI: 16.2%). Among all patients, the most common somatic driver mutations were in APC (75.5%), TP53 (65.3%), TTN (48.6%), and KRAS (41.9%). PGVs in CRC-associated genes were observed in 7.6% of patients (high-penetrance 4.3%, moderate-penetrance 0%, low-penetrance 2.7%, uncertain-penetrance 0.6%). Lynch syndrome (LS) genes were responsible for most high-penetrance findings (78.8%). PGV rates in CRC-associated genes decreased with older age ( 65 years: 5.1%) and was higher in patients with MSI compared to MSS (21.5% vs. 5.2%). PGV rates were similar regardless of patient sex (male, 8.2%; female, 7.5%) and CRC stage (I: 6.0%, II: 9.3%, III: 7.4%, IV: 8.0%). Genetic ancestry was associated with differences in the genomic landscape of CRC. MSI rates were highest in EAS patients (20.6%), followed by EUR (16.4%) and AFR (9.9%) patients. The most common somatic driver mutations were similar to the overall cohort: APC (AFR: 83.4%, EAS: 76.7%, EUR: 74.3%), TP53 (AFR: 68.4%, EAS: 70.4%, EUR: 64.2%), TTN (AFR: 47.8%, EAS: 48.4%, EUR: 48.9%) and KRAS (AFR: 51.4%, EAS: 35.0%, EUR: 41.2%). PGV rates were similar across ancestries (AFR: 8.3%, EAS: 7.6%, EUR: 7.6%), but high-penetrance genes were most common in EAS (6.7%) compared to AFR (5.5%) and EUR (3.7%) patients. LS genes were enriched in EAS patients compared to EUR patients (EAS 6.7%, EUR 2.8%, p MUTYH), respectively. These results demonstrate that integrated analysis of germline and somatic findings across various patient characteristics is critical for an optimal approach for all patients and for developing risk stratification strategies pertaining to early cancer detection.