Performance of individual and joint risk stratification by an environmental risk score and a genetic risk score in a colorectal cancer screening setting

Abstract

Early detection of colorectal neoplasms can reduce the disease burden of colorectal cancer by timely intervention of individuals at high risk. Our aim was to evaluate a joint environmental‐genetic risk score as a risk stratification tool for early detection of advanced colorectal neoplasm (ACRN). Known environmental risk factors and high‐risk genetic loci were summarized into risk scores for ACRN in 1014 eligible participants of a screening study. The performances of single and joint environmental‐genetic scores were evaluated with estimates and 95% confidence intervals (CI) of the absolute risk, relative risk and predictive ability using the area under the curve (AUC). Individuals with higher environmental risk scores showed increasing ACRN risk, with 3.1‐fold for intermediate risk and 4.8‐fold for very high risk, compared to the very low environmental risk group. Similarly, individuals with higher genetic risk scores showed increasing ACRN risk, with 2.2‐fold for intermediate risk and 3.5‐fold for very high risk, compared to the lowest genetic risk group. Moreover, the joint environmental–genetic score improved the ACRN risk stratification and showed higher predictive values (AUC = 0.64; 95%CI = 0.60–0.67) with substantial difference (p = 0.0002) compared to the single environmental score (0.58; 0.55–0.62). The integration of environmental and genetic factors looks promising for improving targeting individuals at high‐risk of colorectal neoplasm. Applications in practical screening programs require optimization with additional genetic and other biomarkers involved in colorectal carcinogenesis.

Adding genetic scores to risk models in colorectal cancer

“…The model combining all scores estimated CRC risk with a discriminatory accuracy value of 0.63 for men and 0.62 for women, higher in both genders when comparing to those models based only in family history or E-score and G-score separately. In line with these results, Balavarca et al [10] after evaluate environmental factors and GRS for advanced colorectal neoplasm, they reported higher prediction values (AUC = 0.63) in the combined environmental-genetic score model compared with single environmental score (AUC = 0.584, p = 0.0002).…”

Use of Polygenic Risk Scores to Select Screening Intervals After Negative Findings From Colonoscopy

“…Further research, such as modeling by microsimulation analyses, should address the impact of risk-adapted differentiation and potential prolongation of screening intervals on effectiveness and costeffectiveness of colonoscopy-based screening. Future studies should also evaluate whether the combination of PRSs and environmental risk scores 29,30 could better define personalized screening intervals and take additional molecular characteristics of CRC cases into account. Adjusted for age, sex, education, body mass index, participation in a health checkup, family history of colorectal cancer, smoking, ever regular use of nonsteroidal anti-inflammatory drugs, and ever regular use of hormone replacement therapy.…”

Whole-blood DNA Methylation Markers for Risk Stratification in Colorectal Cancer Screening: A Systematic Review

“…Compared to methylation of SEPT9 [22,23] in tumor-derived cell-free DNA, or to other types of molecular markers such as micro RNA [60,61], genetic [62], or proteomic markers [63,64], performance of whole-blood methylation markers for diagnosis/risk stratification of colorectal neoplasms seems much poorer. In the past 10–15 years large progress has been made in risk stratification of CRC by genetic risk variants and their combination in polygenetic risk scores [62,65,66,67,68]. Reproducible genetic risk variants have mostly been identified by hypothesis-free screening of genetic variants through genome-wide arrays in large scale international consortia including tens of thousands of CRC cases and controls, with thorough correction for multiple testing and independent validation of promising risk variants and risk scores, rather than by candidate gene approaches.…”