ObjectiveTo explore the causes of colon-anal anastomotic stenosis in patients with low rectal cancer after prophylactic ileostomy under complete laparoscopy. MethodsA total of 194 patients with low rectal cancer who received complete laparoscopic radical resection of rectal cancer combined with preventive ileostomy in our hospital from January 2020 to December 2020 were selected as the study objects, and were divided into non-stenosis group (n=136) and stenosis group (n=58) according to postoperative colon-anal anastomosis stenosis. The clinical data of the two groups were compared. Univariate and multivariate logistic regression were used to analyze the factors affecting postoperative colon-anal anastomotic stenosis, and stepwise regression was used to evaluate the importance of each factor. The risk prediction model of postoperative colon-anal anastomotic stenosis was constructed and evaluated. ResultsIn the stenosis group, the proportion of males, tumor diameter ＞3 cm, NRS2002 score >3 points, manual anastomosis, left colic artery not preserved, anastomotic leakage, pelvic infection and patients undergoing neoadjuvant radiotherapy and neoadjuvant chemotherapy were higher than those in the non-stenosis group (P<0.05). The results of univariate logistic analysis showed that female and preserving the left colonic artery were the protective factors for postoperative colon-anal anastomotic stenosis (P<0.05), and the tumor diameter >3 cm, NRS2002 score >3 points, manual anastomosis, anastomotic leakage, pelvic infection, neoadjuvant radiotherapy and neoadjuvant chemotherapy were the risk factors for postoperative colon-anal anastomotic stenosis (P<0.05). Multivariate logistic regression analysis showed that gender, tumor diameter, NRS 2002 score, anastomotic mode, anastomotic leakage, and pelvic infection were independent influencing factors for postoperative colon-anal anastomotic stenosis (P<0.05). Stepwise regression analysis showed that the top three factors affecting postoperative colon-anal anastomotic stenosis were NRS 2002 score, gender and anastomotic leakage. Multivariate Cox risk proportional model analysis showed that the multivariate model composed of NRS 2002 score, gender and anastomotic leakage had a good consistency in the risk assessment of postoperative colon-anal anastomotic stenosis. Based on this, a risk prediction model for postoperative colon-anal anastomotic stenosis was constructed. The results of strong influence point analysis show that there are no data points in the modeling data that have a strong influence on the model parameter estimation (Cook distance <1). Receiver operating characteristic curve results showed that the model had good differentiation ability, the area under curve was 0.917, 95%CI was (0.891, 0.942). The calibration curve was approximately a diagonal line, showing that the model has good predictive power (Brier value was 0.097). The results of the clinical decision curve showed that better clinical benefits can be obtained by using the predictive model to identify the corresponding risk population and implement clinical intervention. ConclusionThe prediction model based on NRS 2002 score, gender and anastomotic fistula can effectively evaluate the risk of colon-anal anastomotic stenosis after preventive ileostomy in patients with low rectal cancer under complete laparoscopy.