Colorectal cancer (CRC) is one of the most commonly diagnosed cancers, and remains the second leading cause of cancer death, despite the fact that rates have declined and that being among the most preventable of cancers. Current screening recommendations are based only on age, family history of CRC, and previous screening results, whereas risk varies substantially in the population and most cases occur in those without a positive family history. The recent substantial progress in identifying CRC associated common genetic variants has offered new opportunities to refine risk-prediction models and provide more accurate risk stratification for those at risk. We developed risk-prediction models for men and women based on age, family history, and 28 CRC-associated common genetic variants identified to date, accounting for history of endoscopy, in 5,811 cases and 6,302 controls, and validated the models in an independent study of 866 cases and 869 controls. The genetic risk score derived from CRC-associated common variants provides an independent risk predictor and significantly improves the discriminatory accuracy over the model that includes only family history for both men (AUC = 0.60 vs. 0.51, p-value = 4.6e-6) and women (AUC=0.57 vs. 0.52, p-value=0.02). Including the genetic risk score may also improve risk stratification. Considering, for example, the average 10-year risks of 60-year-old men (2.68%) and women (1.49%) in the absence of endoscopy as the thresholds for screening, then the age to start screening for men with positive family history would be 54, and 61 years old otherwise. For women the starting age would be 57 and 61 years old, respectively. Including additional genetic variants information leads to a wider range, from 51 years old for those with positive family history and the genetic risk scores at the 90% to 68 years old for those without positive family history and the genetic risk scores at the 10%. For women, the range would be from 53 years old to 66 years old, respectively. We also derived estimates of 10-year absolute risk for men and women given their age, family history status, risk alleles carried, and endoscopy. To illustrate, for a 60-year-old man who does not have positive family history, but who carries 28 risk alleles (top 20%), and has not had an endoscopy, risk for developing CRC is 5.49% (95% CI, 3.63 to 7.35); however, if he has undergone endoscopy, his risk is reduced to 1.67% (95% CI, 1.10 to 2.25). For a 60-year-old woman who has a similar profile, risk of developing CRC is 2.41% (95% CI, 1.98 to 2.83) in the absence of endoscopy and is reduced to 1.23% (95% CI, 1.00 to 1.46) if she has undergone endoscopy. Our development and validation of a risk-prediction model are first steps towards utilizing genetic information for more tailored screening and prevention strategies, which should improve the cost-effectiveness of screening and, perhaps, adherence.

Citation Format: Li Hsu, Jihyoun Jeon, Hermann Brenner, Sonja Berndt, Andrew T. Chan, Jenny Chang-Claude, Stephen B. Gruber, Tabitha Harrison, Michael Hoffmeister, Polly Newcomb, John Potter, Fredrick Schumacher, Martha Slattery, Duncan Thomas, Emily White, Yingye Zheng, Ulrike Peters, Colorectal Transdisciplinary Study; Genetics and Epidemiology of Colorectal Cancer Consortium. Assessing colorectal cancer risk using known genetic susceptibility variants and family history accounting for endoscopy. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2921. doi:10.1158/1538-7445.AM2014-2921