Optimizing colorectal cancer screening through polygenic risk score-based risk stratification: evidence from a population-based cohort and screening trial

Abstract Background Accurate risk stratification of colorectal cancer (CRC) is essential for precise screening. Polygenic risk scores (PRS) hold promise for improving predictive efficacy in CRC. However, the real-world applicability of a risk-adapted CRC screening strategy based on the PRS remains underexplored. Therefore, we aimed to evaluate the optimized PRS in a large prospective cohort in China and assess its utility for risk-adapted CRC screening. Methods We evaluated multiple PRS construction strategies using East Asian genome-wide association study data and well-established PRSs to select an optimal score, which was then assessed in 100,639 eligible participants from the China Kadoorie Biobank. The risk-adapted screening strategy assigns high-risk individuals to colonoscopy and low-risk individuals to fecal immunochemical testing (FIT), with FIT-positive cases referred for colonoscopy. We assessed the screening performance of the PRS, Asia–Pacific Colorectal Screening score, and their combination-based risk-adapted screening strategies against a standard FIT-based strategy among 2,821 participants in the TARGET-C CRC screening trial. Results The combined PRS (i.e., PRS121) demonstrated the best predictive performance (C-index = 0.602) for CRC. Individuals in the highest PRS quintile (top 20%) exhibited a 2.69-fold increased CRC risk compared with those in the lowest quintile. The high PRS and unfavorable lifestyle group conferred the highest risk (hazard ratio = 3.32, 95% confidence interval: 1.80–6.11). In the TARGET-C trial, the PRS121 effectively distinguished patients with advanced neoplasia from controls (top 20%, odds ratio = 3.66). The PRS-based risk-adapted screening strategy improved AN detection compared with FIT-only screening, with higher sensitivity (69.4% vs. 54.1%, P = 7.6 × 10–4) and detection rate (16.7% vs. 13.1%, P = 0.024). Notably, PRS-based screening identified 21.1% of AN cases that were missed by FIT, and integrated risk stratification using PRS and APCS increased this proportion to 37.9%, demonstrating substantial improvement in detecting FIT-negative lesions. Conclusions PRS enables effective risk stratification for colorectal cancer and improves detection of advanced neoplasia by identifying high-risk individuals missed by FIT, supporting its utility in precision risk-adapted screening. Trial registration Chinese Clinical Trial Registry: ChiCTR1800015506. Registered on 4 April 2018.