摘要结直肠癌（CRC）是我国常见的消化道恶性肿瘤之一，基于流行病学危险因素及多组学生物标志物的CRC风险预测模型，通过评估个体的发病风险，可以进一步识别高危人群，并针对不同风险人群制定个性化的筛查方案，从而有效降低CRC的发病率和死亡率。然而，传统风险预测模型通常采用全人群无差别风险评估的方式，易造成资源浪费且影响筛查成效。相比之下，采用多阶段、多层次结构的序贯式分层模型，通过逐级富集，将有限资源集中于高风险人群的精准筛查，可提高结肠镜利用效率和筛查准确性，有望成为CRC筛查的有效策略。本文就CRC传统风险预测模型和序贯式风险预测模型进行综述，为从传统模型到序贯式分层模型的转变提供依据。

abstractsColorectal cancer (CRC) is one of the most common malignancies affecting the digestive system in China. Risk prediction models for CRC based on epidemiological risk factors and multi-omics biomarkers enable the assessment of an individual′s likelihood of developing the disease. Such models can help identify high-risk populations and facilitate personalized screening strategies tailored to different risk levels, thereby effectively reducing the incidence and mortality of CRC. However, traditional risk prediction models often adopt a uniform risk assessment approach for the entire population, which may lead to inefficient resource allocation and reduced screening effectiveness. In contrast, a sequential, multi-level stratified model employing a progressive risk enrichment strategy can concentrate limited resources on precise screening of high-risk groups. This approach enhances the efficiency and accuracy of colonoscopy screening and may represent a promising strategy for CRC prevention. This review summarizes both traditional and sequential risk prediction models for CRC, providing evidence to support the transition from conventional frameworks to sequential stratified models.