Artificial Neural Network Model for Predicting 5-year Mortality after Surgery for Hepatocellular Carcinoma and Performance Comparison with Logistic Regression Model: A Nationwide Taiwan Database Study

Abstract

Despite the improving prediction methods reported in outcome prediction studies of hepatocellular carcinoma (HCC) surgery, few studies have used longitudinal data for periods exceeding two years. Additionally, most studies have analyzed populations in the US or in the OECD countries, which may substantially differ from those in Taiwan. The purpose of this study was to validate the use of artificial neural network (ANN) models for predicting 5-year mortality in HCC and to compare their predictive capability with that of logistic regression (LR) models. This study compared LR and ANN models based on initial clinical data for 22,926 HCC surgery patients, including age, gender, Charlson co-morbidity index (CCI), chemotherapy, radiotherapy, hospital volume, surgeon volume, length of stay and outcome. A global sensitivity analysis was also performed to assess the relative significance of input parameters in the system model and to rank the importance of variables. The ANN model outperformed the LR model in all performance indices. The most influential (sensitive) parameter affecting in-hospital survival was surgeon volume followed by hospital volume and CCI. In comparison with the conventional LR model, the ANN model in this study was more accurate in predicting 5-year mortality and had higher overall performance indices.