Development of a Prediction Model for Colorectal Adenoma Recurrence and Cancer Risk Using Combined Traditional Chinese and Western Medicine Features

What this Trial Is About

Researchers are developing a risk prediction model to estimate the chances of colorectal adenoma coming back or turning into cancer by combining features from Traditional Chinese Medicine TCM and Western medicine. This project, led by Dr. Wei Hongtao at Beijing Friendship Hospital, uses clinical data from both medical systems to build a model that can help identify patients at higher risk. The study involves recruiting patients with colorectal adenomas from multiple centers across the country and follows strict diagnostic standards from both TCM and Western medicine. Participants in this observational study are patients diagnosed with colorectal adenoma who have undergone endoscopic removal procedures like endoscopic mucosal resection, endoscopic submucosal dissection, or cold snare polypectomy, with pathology confirming the adenoma. Data collected include demographic information, TCM clinical features such as syndromes and tongue and pulse diagnosis, and Western clinical features including endoscopic data. Various algorithms analyze these multimodal data to build and evaluate the risk prediction model, aiming also to find key patient groups for better TCM treatment strategies. During the study, participants will be regularly observed and followed up for up to three years, with an average follow-up time of 1.5 years. Assessments include colonoscopies and laboratory tests to track recurrence of adenomas and advanced adenomas. The study collects comprehensive clinical data and monitors participants carefully to ensure data accuracy. Outcomes focus on the rate of advanced adenoma recurrence and overall adenoma recurrence during the study period. The project also includes training graduate students and publishing scientific papers based on the findings.

Prediction Model of Colorectal Adenoma Recurrence-Carcinogenesis Risk With TCM-WM Multimodal Feature Fusion