Fig. 4.

Performance of binary and multi-class classification models in HCC treatment response. A illustrates the performance of binary classification models, with each point representing a different machine learning model applied to T1wT1w, T2wT2w, or combined T1w + T2w MRI datasets. The x-axis indicates the AUROC and the y-axis shows the accuracy of each model. The symbols correspond to different models: squares for SVM, circles for LR, triangles for RF, and diamonds for XGBoost. B displays the performance of multi-class classification models, using the same symbol