Clinical factors associated with advanced colorectal neoplasia in average-risk, asymptomatic individuals.

Abstract

89 Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality in the U.S. Advanced precancerous lesions and CRC are classified as advanced colorectal neoplasia (ACN). Risk factors for ACN in average-risk, screen-eligible individuals remain unclear due to reliance on small, retrospective, heterogenous cohorts. Using data from PREEMPT CRC—a large prospective, multicenter clinical study of a CRC screening blood test—we applied a Conditional Inference Tree (CIT) model to identify clinical risk factors and their interactions. Methods: We analyzed over 28,000 average-risk adults aged 45-85 years with disease status determined by colonoscopy and histopathology from PREEMPT CRC. Factors considered for analysis included age, sex, race, ethnicity, BMI, smoking status, pack-year history, alcohol use, aspirin, NSAID, diabetes, hyperlipidemia, chronic kidney disease, lower vitamin D, fatty liver, and hypertension. The CIT model identified risk factors through a two-step permutation testing process: selecting the variable most significantly associated with ACN (i.e., the smallest p-value <0.05), then determining the optimal split to define clinically distinct subgroups (Hothorn 2006). Results: The primary risk factors for ACN were sex and BMI, followed by smoking status, cumulative smoking exposure, and age. BMI and smoking status were significant factors for both sexes. In females, current smokers had a higher ACN prevalence than former or never smokers (10.6% vs 6.7% for BMI ≤ 30.54 and 14.3% vs 8.9% for BMI > 30.54). In males with BMI > 32.36, current and former smokers had a higher ACN prevalence than never smokers (18.7% vs 14.0%). Among males with BMI ≤ 32.36, risk was further stratified by smoking exposure and age. Males with > 36 pack-years had an ACN prevalence of 22.6%. For male never-smokers with BMI 26.61-32.36, ACN prevalence nearly doubled from 6.2% for those aged ≤ 50 years to 12.0% for those aged > 50 years. Conclusions: Sex, BMI, and smoking were the strongest clinical predictors of ACN, with additional risk stratification by cumulative smoking exposure and age in men. These findings highlight the importance of behavioral intervention and other risk mitigation strategies to reduce the incidence of disease.