Development and validation of a clinical factors and body fat distribution-based nomogram to predict refractoriness of transarterial chemoembolization in hepatocellular carcinoma.

ABSTRACT

Background: Transarterial chemoembolization (TACE) is an important treatment modality for hepatocellular carcinoma (HCC). However, some patients may develop TACE refractoriness during treatment. We aimed to construct a prediction model incorporating computed tomography (CT) body composition and clinical factors to preoperatively predict the risk of developing TACE refractoriness in patients with HCC, enabling the rapid identification of patients at high risk of TACE refractoriness. Methods: This study included 128 HCC patients treated with TACE who were randomly assigned to the training (n=89) and validation groups (n=39) in a 73 ratio. Multiple body-composition parameters were outlined from CT images of the third lumbar vertebra level of each patient. Standardized values of body-composition parameters were calculated, such as visceral-to-subcutaneous adipose tissue area ratio (VSR). Multifactor logistic regression analysis was performed to identify independent predictors of TACE-refractoriness in patients and to develop predictive models. High- and low-risk subgroup analyses were performed for the predictive model. Results: Alpha-fetoprotein (AFP) level (P=0.041), tumor size (P=0.001), and VSR (P=0.043) were independent risk factors for TACE refractoriness. The combined clinical-body composition model had an area under the curve (AUC) value of 0.875 in the training cohort and an AUC value of 0.837 in the validation cohort. Calibration curves and decision curves revealed the specific optimal performance and clinical utility of the combined model. Subgroup analysis showed differences in predicted TACE refractoriness between the high- and low-risk groups (P<0.001). Conclusions: The combined clinical-body fat distribution model has the good performance in predicting a patient's risk of TACE refractoriness preoperatively and can help clinicians make the best clinical decisions in advance for the treatment of high-risk patients.