QN-S3VM Method for Evaluation of Liver Functional Reserve

When doctors assess the liver reserve function of patients with liver cancer before hepatectomy, if the assessment is not appropriate, patients are prone to postoperative liver failure(POLF). Child-Pugh is the most commonly used method for clinical evaluation, but this method only relies on biochemical indicators, which may lead to poor prediction. This paper aims to evaluate the liver reserve function of patients with imaging omics information using an effective machine learning scheme. The performance of models including support vector machine (SVM), COP-Kmeans, MLP, SGD-S3VM, QN-S3VM are predicted by comparing accuracy and F1 Score. In particular, the QN-S3VM model with the best performance can achieve an accuracy of 0.91, which is higher than the Child-Pugh method. As such, the proposed intelligent diagnostic algorithm can provide routine diagnostic assistance and consultation for doctors to achieve the purpose of reducing the workload.