Development and External Validation of a Nomogram Including Body Composition Parameters for Predicting Early Recurrence of Hepatocellular Carcinoma After Hepatectomy

Abstract

Body composition, including adipose and muscle tissues, evaluated by computer tomography is correlated with the prognosis of hepatocellular carcinoma (HCC). However, its relationship with early recurrence (ER) remains unclear. This study aimed at establishing and validating a nomogram based on body composition and clinicopathological indices to predict ER of HCC. One hundred ninety-five patients from institution A formed the training cohort and internal validation cohort, and 50 patients from institution B formed the external validation cohort. Independent predictors of ER were identified using LASSO and Cox regression analyses. The performance of nomogram was evaluated using the calibration curve, concordance index (C-index), area under the curve (AUC), and decision curve analysis (DCA). After data screening, the nomogram was constructed using eight independent predictors of ER, including the tumor size, alpha fetoprotein, body mass index, Edmondson Steiner grade, visceral adipose tissue radiodensity, intermuscular adipose tissue index, intramuscular adipose tissue content, and skeletal muscle area. The calibration curve exhibited excellent concordances, with C-indices of 0.808 (95%CI: 0.771-0.860), 0.802 (95%CI: 0.747-0.942), and 0.804 (95%CI: 0.701-0.861) in training, internal validation, and external validation cohorts, respectively. In addition, compared to conventional staging systems and pure clinical model, the nomogram exhibited a higher AUC and wider range of threshold probabilities in DCA, which indicated better discriminative ability and greater clinical benefit. Finally, patients with nomogram scores of <183.07, 183.07-243.09, and >243.09 were considered to have low, moderate, and high risks of ER, respectively. The nomogram exhibits excellent ER predictive ability for patients with HCC who underwent hepatectomy.