Risk model based on lifestyle factors shows discriminative ability for colon cancer and precancerous lesions

This study aimed to develop and externally validate a non-invasive predictive nomogram for stratifying individuals at elevated risk for colon cancer and precancerous lesions, with the goal of optimizing risk-stratified screening protocols. The model was derived from a clinical screening cohort and validated using the National Health and Nutrition Examination Survey (NHANES) database. The modeling cohort consisted of 400 participants who underwent colonoscopy, comprising 272 healthy controls and 128 patients histopathologically diagnosed with colon cancer or precancerous lesions. External validation was performed on 284 individuals from the NHANES database (191 healthy controls and 93 self-reported cases). All predictors were selected based on their non-invasive nature and clinical accessibility. Significant intergroup differences were observed in age, body mass index (BMI), smoking history, alcohol consumption, and high-fat diet (all P 55 years, BMI >25 kg/m², smoking, alcohol use, and a high-fat diet as independent risk factors for colonic neoplasia. The derived nomogram exhibited robust discriminative ability in both the modeling (AUC, 0.765) and validation (AUC, 0.761) cohorts. Decision curve analysis demonstrated that intervention guided by the nomogram yielded superior net clinical benefit at risk thresholds of >0.15 and >0.11 in the modeling and validation cohorts, respectively. This novel, non-invasive nomogram provides a reliable and pragmatic tool for individualized risk assessment of colon cancer and precancerous lesions. Its strong performance in both internal and external validations supports its potential utility in enhancing risk-stratified screening and early intervention strategies in diverse populations.