Skip to main content

Table 4 Accuracy, sensitivity, specificity and AUC scores of MVI predictive models for training, validation and external sets

From: Predicting microvascular invasion in hepatocellular carcinoma: a deep learning model validated across hospitals

Model Training set Validation set External set
Accuracy Sensitivity Specificity AUC Accuracy Sensitivity Specificity AUC Accuracy Sensitivity Specificity AUC
ResNet-18(AP) 0.95 0.91 0.97 0.98 0.68 0.96 0.56 0.82 0.66 0.8 0.62 0.75
ResNet-18(AP + CF) 0.97 0.94 0.98 0.97 0.72 0.96 0.62 0.85 0.71 0.82 0.67 0.78
SVM(CF) 0.71 0.82 0.66 0.78 0.77 0.71 0.8 0.78 0.7 0.77 0.67 0.76
SVM(AP + CF) 0.93 0.92 0.93 0.98 0.6 0.93 0.46 0.7 0.57 0.9 0.47 0.68
  1. ResNet-18 model built with AP images and CF generalized well and produced the best metric scores on the external validation set. SVM support vector machine; AP arterial phase; CF clinical factors; AUC area under the curve.