Risk factors and clinical prediction models for short-term recurrence after endoscopic surgery in patients with colorectal polyps.

Abstract

Background: Colorectal polyps (CPs) are important precursor lesions of colorectal cancer, and endoscopic surgery remains the primary treatment option. However, the short-term recurrence rate post-surgery is high, and the risk factors for recurrence remain unknown.

Aim: To comprehensively explore risk factors for short-term recurrence of CPs after endoscopic surgery and develop a nomogram prediction model.

Methods: Overall, 362 patients who underwent endoscopic polypectomy between January 2022 and January 2024 at Nanjing Jiangbei Hospital were included. We screened basic demographic data, clinical and polyp characteristics, surgery-related information, and independent risk factors for CPs recurrence using univariate and multivariate logistic regression analyses. The multivariate analysis results were used to construct a nomogram prediction model, internally validated using Bootstrapping, with performance evaluated using area under the curve (AUC), calibration curve, and decision curve analysis.

Results: CP re-occurred in 166 (45.86%) of the 362 patients within 1 year post-surgery. Multivariate logistic regression analysis showed that age (OR = 1.04, P = 0.002), alcohol consumption (OR = 2.07, P = 0.012), Helicobacter pylori infection (OR = 2.34, P < 0.001), polyp number > 2 (OR = 1.98, P = 0.005), sessile polyps (OR = 2.10, P = 0.006), and adenomatous pathological type (OR = 3.02, P < 0.001) were independent risk factors for post-surgery recurrence. The nomogram prediction model showed good discriminatory (AUC = 0.73) and calibrating power, and decision curve analysis showed that the model had good clinical benefit at risk probabilities > 20%.

Conclusion: We identified multiple independent risk factors for short-term recurrence after endoscopic surgery. The nomogram prediction model showed a certain degree of differentiation, calibration, and potential clinical applicability.