Establishment of clinical predictive model based on the study of influence factors in patients with colorectal polyps.

Abstract

Colorectal cancer (CRC) is the most common gastrointestinal malignancy and is generally thought to be caused by the transformation of colorectal polyps. It has been shown that early detection and removal of colorectal polyps may reduce the mortality and morbidity of colorectal cancer. Based on the risk factors associated with colorectal polyps, an individualized clinical prediction model was built to predict and evaluate the possibility of developing colorectal polyp. A case-control study was conducted. Clinical data were collected from 475 patients who underwent colonoscopy at the Third Hospital of Hebei Medical University from 2020 to 2021. All clinical data were then divided into training sets and validation sets by using R software (7:3). A multivariate logistic analysis was performed to identify the factors associated with colorectal polyps according to the training set, and a predictive nomogram was created by R software based on the multivariate analysis. The results were internally validated by receiver operating characteristic (ROC) curves, calibration curves, and externally validated by validation sets. Multivariate logistic regression analysis showed that age (OR = 1.047, 95% CI = 1.029-1.065), history of cystic polyp (OR = 7.596, 95% CI = 0.976-59.129), and history of colorectal diverticulums (OR = 2.548, 95% CI = 1.209-5.366) were independent risk factors for colorectal polyps. History of constipation (OR = 0.457, 95% CI = 0.268-0.799) and fruit consumption (OR = 0.613, 95% CI 0.350-1.037) were protective factors for colorectal polyps. The nomogram demonstrated good accuracy for predicting colorectal polyps, with both C index and AUC being 0.747 (95% CI = 0.692-0.801). The calibration curves showed good agreement between the predicted risk by the nomogram and real outcomes. Both internal and external validation of the model showed good results. In our study, the nomogram prediction model is reliable and accurate, which can help early clinical screening of patients with high-risk colorectal polyps, improve polyp detection rate, and reduce the incidence of colorectal cancer (CRC).