Plasma protein biomarkers for early risk prediction of colorectal cancer

Abstract

Background: Prospective studies are necessary to determine the association between circulating proteins and colorectal cancer (CRC) risk, offering non-invasive biomarkers for early CRC risk prediction and interventions. Methods: The Olink Explore-3072 platform was employed to measure 2,923 plasma proteins in 53,073 UK Biobank participants. Cox regression analyses assessed associations between normalized protein expression and CRC risk. We developed a protein score for CRC risk prediction and compared it with a traditional risk model (incorporating age, sex, body mass index, smoking status, alcohol consumption, physical activity, and family history of CRC) using Receiver Operating Characteristic analysis. Results: Over an average 10.8-year follow-up, 652 CRC cases occurred. After multivariable adjustment, seven proteins showed significant associations with CRC risk (false discovery rate-corrected P values < 0.05). Hazard ratios (HRs) for CRC per one standard deviation increase in levels of these proteins ranged from 1.18 (95% confidence interval [CI], 1.07–1.29) to 1.38 (95% CI, 1.23–1.55). Compared with the lowest quintile, the highest quintile of the seven-protein score was associated with an HR for CRC of 4.27 (95% CI, 3.17–5.75). The protein score model and traditional risk factor model showed similar predictive performance for 5-year CRC risk (both areas under the curve [AUC] =0.68), while the combined model showed an AUC of 0.70. Conclusions: This study established a panel of seven plasma proteins that may help identify high-risk individuals for CRC. Impact: This study highlights the potential of plasma protein biomarkers as non-invasive tools for early detection and personalized prevention strategies of CRC.