Risk factor and nomogram for predicting the probability of a permanent stoma after laparoscopic intersphincteric resection for ultralow rectal cancer

Abstract

Objective: A permanent stoma can seriously affect patients' quality of life. Clinicians need to consider the risk of a permanent stoma when making clinical decisions. This study analyzed preoperative predictors of a permanent stoma after laparoscopic intersphincteric resection for low rectal cancer (LISR), and a prediction model was constructed validated. Methods: This was a retrospective study that analyzed clinical data of 331 ultralow rectal cancer patients who were diagnosed with primary rectal adenocarcinoma by endoscopy and pathology, including 218 males and 113 female, (58.8±11.2) years and (23.7±3.1) kg/m2. The patients underwent LISR with a preventive stoma from January 2012 to December 2020. Patients with multiple primary colorectal cancers, who underwent emergency surgery for intestinal obstruction or bleeding or perforation, and did not complete 18 months follow up were exclucled. R software was used to randomly select 234 patients as the modeling group with a ratio of approximately 7:3, and the remaining 97 patients comprised the validation group. The stoma site was determined by the surgeon before the operation, and the ileum 30 cm from the ileocecal valve was selected. The rates of a permanent stoma for the entire group and the preoperative clinical factors that may affect the permanency of a stoma in the modeling group were determined. A permanent stoma was defined as failure to close the stoma at 18 months after surgery. Multivariate logistic regression analysis was used to analyze the preoperative independent risk factors for a permanent stoma after LISR. R software was used to create the nomogram model, and the predictive ability of the nomogram model was evaluated by receiver operating characteristic (ROC) curve analysis. Results: Among the 331 patients who underwent LISR, 37 (26 cases in the modeling group and 11 cases in the validation group, 11.2%) developed a permanent stoma for the following reasons: anastomotic stenosis due to leakage (16 cases, 43.2%), distant metastasis (16 cases, 43.2%), intolerant to stoma closure surgery (3 cases, 8.1%), stenosis due to postoperative radiation (1 case, 2.7%), and poor recovery of anorectal function (1 case, 2.7%). Univariate analysis showed that preoperative neoadjuvant chemoradiotherapy, poorly differentiated tumor, cT3 stage, and distant metastasis were associated with a permanent stoma. Multivariate logistic regression analysis showed that neoadjuvant chemoradiotherapy [OR=3.078, 95% confidence interval (CI): 1.326-7.147; P=0.009], cT3 stage (OR=2.257, 95%CI: 1.001-5.091; P=0.049), and stage IV cancer (OR=16.180, 95%CI: 2.753-95.102; P=0.002) were independent risk factors for permanent stoma after LISR. Based on the selected risk factors, a nomogram model for predicting permanent stoma was constructed. The area under the ROC curve of the modeling group was 0.793, the optimal cut-off value was 0.890, the sensitivity was 0.577, and the specificity was 0.885. The area under the ROC curve of the validation group was 0.953. The corrected curves of the modeling group and the validation group showed a good degree of fit. Conclusion: Neoadjuvant chemoradiotherapy, cT3 stage, and distant metastasis are independent predictors of a permanent stoma after LISR, and the nomogram model is helpful to predict the probability of a permanent stoma. Patients with high-risk factors should be adequately informed of the risk of a permanent stoma before colorectal surgery.