Fig. 7 | Scientific Reports

Fig. 7

From: Noncontrast CT-based deep learning for predicting intracerebral hemorrhage expansion incorporating growth of intraventricular hemorrhage

Fig. 7

Comparison of prediction performance between the 2D-ResNet-101 model and baseline machine learning models.Comparison of AUC, sensitivity, and specificity between the 2D-ResNet-101 model and four baseline prediction models in the internal- (A) and external-testing sets (B). McNemar’s chi-squared test was used to compare differences in sensitivity and specificity, while the DeLong test was used to assess the differences between AUC. *Indicate a significant difference (p < 0.05).

Back to article page