Early Diagnosis Model of Colorectal Adenoma Using Laboratory Examinations

Construction and Validation of an Early Diagnosis Model for Colorectal Adenoma Based on Laboratory Examinations

Colorectal cancer has the third highest incidence and second highest mortality rate of all malignant tumors worldwide. Distinct from most other cancers, colorectal cancer can be prevented; colonoscopy-based identification and removal of adenomatous polyps is the most effective preventive measure. Early intestinal adenomas rarely cause specific symptoms, and many patients are diagnosed at advanced stages once symptoms emerge, leading to unsatisfactory treatment and prognosis. Colonoscopy, the main diagnostic tool for intestinal adenoma, is invasive, resulting in limited patient compliance, while grassroots hospitals face shortages of medical resources. There is an urgent demand for a convenient, affordable and well-tolerated early diagnostic method for intestinal adenoma. Artificial intelligence techniques can efficiently analyze routine clinical laboratory data. This study aims to establish an AI-based predictive model combining clinical information and laboratory test results to realize early identification of intestinal adenoma and optimize patient prognosis.

Colorectal cancer ranks the third in incidence and the second in mortality among all malignant tumors, constituting a major global public health concern. Unlike many other malignancies, colorectal cancer is preventable. Early detection and resection of adenomatous polyps via colonoscopy represent the most effective strategy for colorectal cancer prevention. Intestinal adenomas often present without specific clinical symptoms in the early stage. By the time patients seek medical care due to symptomatic manifestations, most have progressed to the middle or advanced stage, which exerts severe adverse impacts on subsequent therapeutic outcomes and long-term survival prognosis.

Although colonoscopy serves as the primary modality for the early diagnosis of intestinal adenoma, it is an invasive procedure associated with poor adherence among some patients. In addition, primary medical institutions are constrained by limited medical resources. Accurate and timely early diagnosis of intestinal adenoma is closely linked to targeted clinical intervention and improved patient survival outcomes. Therefore, it is critical to identify an early diagnostic approach for intestinal adenoma that boasts high patient acceptance, low technical barriers, convenience and cost-effectiveness. In recent years, with the advancement and wider accessibility of data analytics tools such as artificial intelligence (AI), growing research efforts have focused on addressing this clinical challenge using AI algorithms. Laboratory testing is routinely performed in clinical practice and delivers timely results, and the massive volume of laboratory data provides evidence supporting early disease diagnosis and prognostic prediction. Advances in artificial intelligence enable clinicians to convert abundant clinical data into practical predictive models to enhance diagnostic performance. Accordingly, integrated analysis of electronic medical records and laboratory results using artificial intelligence facilitates timely detection and early diagnosis of intestinal adenoma, and ultimately improves patient survival prognosis.

Inclusion Criteria:

• Patients aged ≥ 18 years undergoing colonoscopy, with pathologically confirmed diagnosis of colorectal adenoma or non-adenomatous lesions (e.g., inflammatory polyps, hyperplastic polyps, chronic inflammation, etc.);
• Capable of reading, understanding and signing the informed consent form;
• The investigator judges that the subject can understand the procedures of this clinical study, and is willing and able to cooperate with and complete all study procedures.

Exclusion Criteria:

• Unavailable clinical data including laboratory test results;
• Neoplastic lesions other than colorectal adenoma;
• Non-first diagnosis of colorectal adenoma;
• Autoimmune diseases;
• Repeated enrolled subjects (duplicate patients);
• Pregnancy or breastfeeding status;
• Failure to obtain informed consent;
• The investigator considers that the subject has high-risk diseases or other special conditions unsuitable for participating in this clinical trial.