Construction of a nutritional risk prediction model for ovarian cancer patients based on Logistic regression and decision tree

WANG Jing; HU Shiyin; GONG Qian; LI Hongxia; LIU Yi; HE Linsheng

doi:10.12102/j.issn.1009-6493.2025.22.004

Abstract：Objective:To construct a nutritional risk prediction model for ovarian cancer patients based on logistic regression and classification and regression tree(CART).Methods:A total of 326 ovarian cancer patients admitted to the oncology department of a tertiary hospital in Jiangxi province from November 2023 to September 2024 were selected as the study subjects by convenience sampling.General data of the patients were collected.The nutritional risk of the patients was assessed.A predictive model for nutritional risk in ovarian cancer patients was established based on Logistic regression and decision tree models.Influencing factors were analyzed.Results:The incidence of nutritional risk in ovarian cancer patients was 50.92%.The Logistic regression analysis showed that average monthly income,tumor pathological type,chemotherapy,and hypoproteinemia were influencing factors of nutritional risk in ovarian cancer patients(P<0.05).The decision tree model results indicated that chemotherapy,tumor pathological type,average monthly income,hypoalbuminemia,age,and employment status were influencing factors of nutritional risk in ovarian cancer patients.The area under the receiver operating characteristic curve(AUC)for the Logistic model was 0.800.The accuracy was 0.733.The sensitivity was 0.819.The specificity was 0.644.The Youden's index was 0.463.The AUC for the decision tree model was 0.802.The accuracy was 0.742.The sensitivity was 0.831.The specificity was 0.650.The Youden's index was 0.481.Conclusions:Both the Logistic regression and decision tree models showed good discriminatory ability.The combined use of two models was beneficial for the early identification and management of nutritional risk in ovarian cancer patients.It could provide a reference for the comprehensive prevention,treatment,and rehabilitation management of gynecological malignant tumors.