Influencing factors and construction of risk prediction model of malnutrition in patients with gastric cancer undergoing chemotherapy

Objective  To analyze the risk factors of malnutrition in patients with gastric cancer undergoing chemotherapy and construct a nomogram prediction model.

Methods  A total of 152 patients who received complete chemotherapy were selected as study subjects. After complete chemotherapy, the patients were divided into well-nourished group (n=46) and undernourished group (n=106) according to the nutritional evaluation results. Logistic regression analysis was used to screen the influencing factors of malnutrition in patients with gastric cancer undergoing chemotherapy. A nomogram model was constructed to predict the risk of malnutrition in patients with gastric cancer undergoing chemotherapy. Receiver operating characteristic (ROC) curve and calibration curve were used to verify the differentiation and consistency of the nomogram model.

Results  Among 152 patients with gastric cancer treated with chemotherapy, 106 were malnourished, the incidence rate was 69.74%. Male (95%CI, 1.232 to 6.292, P=0.014), no parenteral nutrition support during chemotherapy (95%CI, 1.528 to 7.781, P=0.003), hypoproteinemia (95%CI, 1.193 to 76.786, P=0.033), early satiety (95%CI, 1.259 to 82.652, P=0.030) were independent risk factors for malnutrition in patients with gastric cancer undergoing chemotherapy. The area under ROC curve was 0.757 (95%CI, 0.680 to 0.833). The slope of calibration curve was close to 1, and the Hosmer-Lemeshow goodness of fit test was good (χ²=5.834, P=0.442).

Conclusion  Male, lack of parenteral nutrition support during chemotherapy, hypoproteinemia and early satiety are independent risk factors for malnutrition inpatients with gastric cancer undergoing chemotherapy. The prediction nomogram model built on this basis has good differentiation and consistency.