Analysis of risk factors for anastomotic leakage and construction of a nomogram model in patients undergoing rectal cancer surgery

Gu Tiantian; Wang Huan; Wang Feng; Huang Yuan; Cao Xijuan; Liang Xiaoli; Xu Xunxun; Liu Yao; Li Dongbao

doi:10.11986/j.issn.1673-873X.2025.05.10

Abstract：Objective To identify risk factors for anastomotic leakage(AL)following rectal cancer surgery and to construct a nomogram model to aid clinical decision-making.Methods We retrospectively analyzed 170 rectal cancer patients treated from July 2021 to June 2024 at a hospital in Jiangsu Province.Patients were categorized into an AL group(n=20)or a nonAL group(n=150)according to postoperative presence or absence of leakage.Logistic regression was used to identify independent risk factors for AL,which were then incorporated into a nomogram.Receiver operating characteristic(ROC)curves were used to assess the predictive value of nomogram models for postoperative AL in rectal cancer patients and validation of the model was also performed.Results Compared with non-AL group,those in the AL group were more often male,aged ≥60 years,had combined diabetes,presented with preoperative intestinal obstruction,had tumors with a distal margin＜7 cm from the anal verge,and had received neoadjuvant therapy;they were less likely to have received a protective diverting stoma,had longer time of operation,and lower preoperative serum albumin levels(P＜0.05).Multivariate analysis identified male sex(OR=1.998;95％CI:1.396-2.860),combined diabetes(OR=2.190;95％CI:1.503-3.191),preoperative combined intestinal obstruction(OR=2.507;95％CI:1.671-3.761),tumor distance from anal verge ≥7 cm(OR=0.331;95％CI:0.214-0.514),neoadjuvant therapy(OR=2.273;95％CI:1.569-3.292),and albumin levels did not decrease(OR=0.373;95％CI:0.246-0.567)as independent influence factors for AL(P＜0.05).The resulting nomogram demonstrated strong discrimination with a concordance index(Cindex)of 0.885(95％CI:0.833-0.937),good calibration(Hosmer-Lemeshow goodness of fit test:x2=7.543,df=13,P=0.586;calibration slope is close to 1).Decision curve analysis showed a net clin-ical benefit rate＞0 when the threshold probability was＞0.05.ROC curve analysis showed that the area under the curve(AUC)of the nomogram model for predicting postoperative AL in rectal cancer patients was 0.846(0.794-0.898),with a specificity of 67.35％and a sensitivity of 91.82％.Conclusion Risk factors for postoperative AL in rectal cancer patients in-clude being male,combined diabetes,preoperative combined intestinal obstruction,tumor proximity to the anal verge,neoad-juvant therapy,and low serum albumin.The nomogram based on these factors offers high predictive accuracy for AL risk in this patient population.