Abstract: Objective To explore the relationship between blood indicators and the types of colorectal polyps, and to construct a diagnostic model for early diagnosis of colorectal cancer in the elderly patients. Methods The data of the elderly patients who preformed electronic colonoscopy and polypectomy in Jiangsu Province Hospital from January to June 2020 were collected and compared between the groups with different pathological types, including baseline information (gender, age), blood indicators (blood cell counter, liver enzyme, bilirubin, lipid indices, nutrition state, electrolytes, tumor makers, coagulation indices, inflammatory indicators) and the area, maximum diameter of the polyps. Binary Logistic regression was used to analyze the influencing factors of adenomatous polyps. A polyp diagnostic model was established. Results The levels of serum potassium, monocytes, lymphocyte /monocyte (LM ratio), platelet, neuron-specific enolase (NSE), polyp size, and maximum polyp diameter in the elderly patients with adenomatous polyps were significantly different from those in the elderly with other types of of polyps. A binary Logistic regression model was established to plot the ROC curve and Calibration curve. The AUC value of the serological model to diagnose adenomatous polyps was 0.824(P=0.000) . Conclusions The independent risk factors of adenomatous polyps in the elderly patients include serum potassium, monocytes, LM ratio, platelet, NSE, polyp area and maximum diameter. And the pre-endoscopic diagnostic model of colorectal polyps shows significant value in diagnosing adenomatous polyps.

Key words: diagnostic models, blood potassium, aged, adenomatous polyps, colorectal polyps