Early-Onset Colorectal Cancer Somatic Gene Mutations by Population Subgroups

In the United States, there are striking discrepancies between the epidemiology of early-onset colorectal cancer and late-onset colorectal cancer. According to the Surveillance, Epidemiology, and End Results (SEER) database (RRID:SCR_006902), the incidence and mortality of late-onset colorectal cancer have been decreasing over the past 20 years, whereas early-onset colorectal cancer has seen a dramatic increase in cases and deaths. Early-onset colorectal cancer occurs in individuals under the age of 50 and more frequently in the rectum and distal colon, accounting for about 70% of early-onset cases (1). Diagnosis is typically at an advanced clinical stage (stages III and IV), but it is not clear whether this is due to delayed diagnosis or more aggressive tumor biology (1). Previous research has identified a range of factors, including environment, lifestyle, socioeconomics, and genetics, that can influence the etiology and outcomes of the disease. These factors contribute to differences in the burden of disease among difference population subgroups (2). For example, Black, Asian, and Hispanic patients with early-onset colorectal cancer have lower relative 5-year survival rates compared with white patients, and survival rates have only increased over time for white patients (3). Additionally, males have a 12% to 18% increased hazard of disease-specific death compared with females, and there are also differences in the location of tumors in the colorectum by sex, suggesting different biological mechanisms of tumor development (4, 5). However, few studies have examined the molecular mechanisms underlying these disparities by race/ethnicity and sex in early-onset colorectal cancer. Given that these disparities are multifactorial, it is important to understand the molecular mechanisms that could influence this complex interplay.

In a major contribution to this research gap, Holowatyj and colleagues examined somatic cancer gene mutation patterns and tumor mutational burden (TMB) by race/ethnicity and sex, using the American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) database (6). The patient cohort consisted of 6,903 patients with colorectal cancer presenting with adenocarcinomas [5,856 non-Hispanic white (NHW), 535 non-Hispanic Black (NHB), and 512 Asian/Pacific Islander (API) patients], of which 29.2% were defined as early-onset cases. It should be noted that the race/ethnicity information used in this study was self-identified by the study participants. The study used data from the targeted sequencing of 76 candidate genes.

The study mainly focused on patients with nonhypermutated colorectal cancer, which comprised ∼90% of the cohort. For the remaining hypermutated cases, there was a limited sample size to stratify the cases by race/ethnicity. However, examination of the age of onset and sex differences among these hypermutated colorectal cancers showed that early-onset patients had a higher TMB than late-onset patients, and females had a lower TMB than males in both early-onset and late-onset cases. As hypermutated colorectal cancers display a unique biological landscape, it would be valuable to conduct a similar study on hypermutated early-onset colorectal cancers in a larger patient cohort. Further, hypermutation status is defined based on a small number of colorectal cancer–associated genes and the observed inflection point in the TMB distribution across all tumors. It would be informative to examine the sensitivity of the results in the context of the definition of hypermutation, as it has been suggested that different thresholds should be used depending on various factors such as sequencing strategy (7).

Based on the 6,250 patients with nonhypermutated colorectal cancer, Holowatyj and colleagues showed a lower TMB among early-onset compared with late-onset cases, opposite to that seen for hypermutated cases, and significantly higher odds of having TP53, LRP1B, and TCF7L2 nonsilent somatic mutations (after adjustment for multiple comparisons using FDR < 0.05). Examination by race/ethnicity showed that NHB patients with early-onset colorectal cancers, but not API patients, had a higher TMB than NHW patients. Furthermore, comparing early-onset with late-onset cases showed differences in nonsilent somatic mutations within each race/ethnicity; however, most of these were not significant after FDR adjustment. NHW patients had higher odds of TP53, LRP1B, and TCF7L2 mutations in early-onset versus late-onset cases, similar to the results of the analysis before stratification for race/ethnicity, after adjusting for FDR. Across these early-onset cases, differences in mutation frequencies between racial/ethnic groups were seen for FLT4, FBXW7, RNF43, LRP1B, APC, PIK3CA, and ATRX. Cochran's Q-test for heterogeneity showed that APC, FLT4, and FAT1 were significantly associated with nonsilent somatic mutations between early-onset and late-onset nonhypermutated colorectal cancer cases in at least one racial/ethnic group. Moreover, the mutation frequency of APC was significantly different among the three racial/ethnic groups. It is worth noting that the sample sizes of nonhypermutated early-onset patients were 157 and 164 for NHB and API, respectively, whereas there were 1,511 NHW patients. Data with highly imbalanced distributions of such demographic features are common in studies that examine racial/ethnic differences in colorectal cancer, which can introduce bias. Therefore, future studies that aim to recruit larger sample sizes from diverse populations would be valuable to further examine these race/ethnicity-specific tumor mutation signatures.

Holowatyj and colleagues also observed heterogeneous tumor mutation patterns by sex in their analysis of 1,832 patients with early-onset nonhypermutated colorectal cancer. They found a trend toward lower TMB among males compared with females, opposite to that which was seen for the hypermutated early-onset cases. In addition, female patients with early-onset colorectal cancer had significantly elevated odds of presenting with a TP53 mutation compared with late-onset cases. More genes showed similar heterogeneous mutational patterns between males and females, particularly for the effect of EP300, BRAF, WRN, KRAS, AXIN2, and SMAD2 between early-onset and late-onset nonhypermutated colorectal cancer cases; however, they were not statistically significant after FDR adjustment. The prevalence of KRAS mutations in early-onset colorectal cancers has been inconsistently reported; however, in this study, despite the loss of statistical significance after multiple comparison adjustments, males with early-onset colorectal cancer had lower odds of presenting with KRAS mutations compared with late-onset colorectal cancer, whereas no such pattern was found in females. This sex-specific difference might provide an explanation for the inconsistency of evidence on KRAS mutational patterns between early-onset and late-onset colorectal cancer. An important future research area would be to examine the interactions between specific somatic mutations and sex hormone levels among patients with early-onset colorectal cancer to determine how these sex-specific factors play a role in colorectal cancer development. In addition, further examination of patient subgroups by both sex and race/ethnicity would be of importance. It has been reported that females with early-onset colorectal cancer have the highest proportion of their tumors occurring in the rectum; however, NHB female patients have the highest proportion occurring in the proximal colon. Males also have the highest proportion of their tumors occurring in the rectum for early-onset cases, but NHB male patients have an equal proportion occurring in both the rectum and proximal regions, suggesting again different tumor biology or risk factors by sex and race/ethnicity (5).

The clinical-grade sequencing data in the GENIE database allowed the authors to analyze a larger spectrum of genes compared with previous studies. However, there are some intrinsic limitations. As acknowledged by the authors, the sample sizes for NHB and API cases were too small to analyze nonhypermutated cases and draw statistically significant conclusions after multiple comparisons for some of the results. Thus, a larger group of patients is necessary for external validation. In addition, as an alternative to the self-identified race/ethnicity data used by Holowatyj and colleagues, a recent study on racial/ethnic differences in colorectal cancer estimated genetic ancestry from genomic data instead, which may be less biased (8). Self-identified race/ethnicity, as pointed out by Holowatyj and colleagues, is a social construct and may be correlated with other social determinants that contribute to different colorectal cancer outcomes. Moreover, clinical and demographic information such as clinical stage, medical history, dietary patterns, environmental exposures, and socioeconomic status could offer better insight into potential confounders.

Overall, Holowatyj and colleagues highlighted somatic gene mutation patterns that differ by race/ethnicity and sex for patients with early-onset colorectal cancer. This raises the important question of how these underlying differences in tumor biology could affect clinical outcomes and may provide potential clues for more personalized therapeutic strategies by considering population diversity. Importantly, the data provide an important step forward in understanding the more complex etiology of early-onset colorectal cancer, which likely occurs through a complex interplay of genetic, environmental, and socioeconomic influences.