Prediction of Ki-67 expression in hepatocellular carcinoma with machine learning models based on intratumoral and peritumoral radiomic features

Figure 5 Waterfall plot showing the model performance across the validation sets of the three models. A: Intratumoral model; B: Peritumoral model; C: Fusion model. Each bar in the plot represents the difference between the model’s predicted value and the established cut-off value. The bars extending above the Y = 0 line indicate model predictions of high Ki-67 expression, whereas the bars below the Y = 0 line signify predictions of low Ki-67 expression.