关键词: 肺癌, 手术, 营养不良, 预测模型, 列线图

Abstract: Objective: To investigate risk factors for malnutrition in postoperative lung cancer patients, and to develop and validate a predictive model for malnutrition risk in this population. Methods: Using a convenience sampling method, a total of 368 postoperative lung cancer patients admitted to a tertiary grade A hospital in Tianjin from May 10, 2024, to April 30, 2025, were enrolled as study subjects. They were randomly divided into a modeling group (n=259) and a validation group (n=109) in a 7：3 ratio. Univariate and multivariate analyses were performed to identify risk factors for malnutrition in postoperative lung cancer patients. A nomogram was developed to construct a risk prediction model for malnutrition in this population. The performance of the model was evaluated using the receiver operating characteristic (ROC) curve, Hosmer-Lemeshow test, calibration curve, and decision curve analysis (DCA). Results: The incidence of malnutrition among the 368 postoperative lung cancer patients was 26.36%. Age, BMI, albumin, hemoglobin, pathological type, and tumor stage were identified as risk factors for postoperative malnutrition (all P<0.05). In the modeling group, area under curve (AUC) of the modeling model was 0.902(95%CI：0.861~0.942), with a sensitivity of 0.905 and a specificity of 0.754. The Hosmer-Lemeshow goodness of fit test showed a good fit (χ2=7.894, P=0.444). The calibration curve showed high consistency between predicted and observed probabilities. The decision curve analysis (DCA) indicated that the model had good clinicalutility. In the validation group, the AUC was 0.862 (95%CI：0.820~0.903), indicating satisfactory discriminative ability of the model. Conclusion: The malnutrition risk prediction model developed for postoperative lung cancer patients demonstrates good predictive performance, thereby providing a decision-making tool for early, precise, and individualized nutritional interventions in clinical practice.

Key words: Lung cancer, Surgery, Malnutrition, Predictive model, Nomogram