Abstract
Malnutrition in patients is associated with reduced tolerance to treatment-related side effects and higher risks of complications, directly impacting patient prognosis. Consequently, a pressing requirement exists for the development of uncomplicated yet efficient screening methods to detect patients at heightened nutritional risk. The aim of this study was to formulate a concise nutritional risk prediction model for prompt assessment by oncology medical personnel, facilitating the effective identification of hepatocellular carcinoma patients at an elevated nutritional risk. Retrospective cohort data were collected from hepatocellular carcinoma patients who met the study's inclusion and exclusion criteria between March 2021 and April 2022. The patients were categorized into two groups: a normal nutrition group and a malnutrition group based on body composition assessments. Subsequently, the collected data were analyzed, and predictive models were constructed, followed by simplification. A total of 220 hepatocellular carcinoma patients were included in this study, and the final model incorporated four predictive factors: age, tumor diameter, TNM stage, and anemia. The area under the ROC curve for the short-term nutritional risk prediction model was 0.990 [95% CI (0.966-0.998)]. Further simplification of the scoring rule resulted in an area under the ROC curve of 0.986 [95% CI (0.961, 0.997)]. The developed model provides a rapid and efficient approach to assess the short-term nutritional risk of hepatocellular carcinoma patients. With easily accessible and swift indicators, the model can identify patients with potential nutritional risk more effectively and timely.