To develop a multivariable model incorporating lipid profiles and endoscopic features of colorectal adenomatous polyps, and to validate its performance in predicting polyp recurrence risk and estimating polyp growth rate based on the surveillance interval.

A retrospective observational study was conducted, enrolling adult patients who had undergone adenomatous polypectomy and at least one follow-up colonoscopy at the Second Affiliated Hospital of Nanjing University of Chinese Medicine. Baseline and follow-up endoscopic data (polyp number, cumulative volume, maximum diameter, and surveillance interval) from the rectum and ascending colon were collected. Patients from the endoscopy center (n=795) and the colorectal surgery center (n=476) were assigned to the test group and validation group, respectively. Univariate and multivariate logistic regression were used to screen for independent predictors of recurrence and to construct a nomogram model. The discriminatory performance of the model was evaluated using the area under the receiver operating characteristic curve (AUC). Multiple linear regression, robust regression, and log-transformation models were employed to analyze the impact of various factors on the number, cumulative volume, and maximum diameter of polyps detected during follow-up.

A total of 1 271 patients were included (1 005 in the recurrence group and 266 in the non-recurrence group). Multivariate logistic regression showed that elevated total cholesterol (OR=2.51, 95% CI: 1.53~4.11), prolonged surveillance interval (OR=1.08, 95%CI: 1.04~1.12), and higher baseline polyp count (OR=1.15, 95%CI: 1.06~1.25) were independent risk factors for polyp recurrence, while decreased high-density lipoprotein cholesterol (OR=0.51, 95%CI: 0.27~0.95) was a protective factor. The nomogram model constructed based on these factors achieved AUCs of 0.736 and 0.710 in the training and validation sets, respectively. Multiple linear regression analysis further confirmed that the surveillance interval was a stable factor for predicting the number (β=0.13, P＜0.001), cumulative volume (β=2.94, P＜0.001), and maximum diameter (β=0.23, P＜0.001) of polyps at follow-up. The model predicted a median time of 25.08 months for the maximum polyp diameter to reach 10.0 mm. Subgroup analysis showed that the recurrence rates in the rectal group and the ascending colon group were 80.60% and 77.39%, respectively, with no statistically significant difference (P = 0.182). Polyp location was not identified as an independent predictor of recurrence (OR = 0.82, 95% CI: 0.61~1.08, P = 0.160).

The multivariable model prediction model, integrating lipid parameters with endoscopic features, has the potential to serve as an adjunctive tool for post-polypectomy recurrence risk assessment, providing a preliminary quantitative reference for individualizing surveillance intervals.