Table 4 Performance Evaluation for Developed Models Using Internal and External Dataset.
Model | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | F1-score (%) | Accuracy (%) | AUROC |
|---|---|---|---|---|---|---|---|
Internal Dataset | |||||||
AlexNet | 75.9 | 84.2 | 60.2 | 91.7 | 67.2 | 82.2 | 0.837 |
DenseNet-121 | 94.4 | 91.2 | 77.2 | 98.1 | 84.9 | 92.0 | 0.961 |
ResNet-50 | 85.1 | 90.0 | 73.0 | 95.0 | 78.6 | 88.8 | 0.960 |
ResNext-50 | 77.8 | 92.9 | 77.8 | 92.9 | 77.8 | 89.3 | 0.941 |
CheXNet | 88.9 | 90.6 | 75.0 | 96.2 | 81.3 | 90.2 | 0.954 |
EfficientNet -B3 | 90.7 | 87.1 | 69.0 | 96.7 | 78.3 | 88.0 | 0.954 |
VGG-19 | 75.9 | 87.1 | 65.0 | 91.9 | 70.0 | 84.4 | 0.870 |
External Dataset | |||||||
AlexNet | 57.5 | 70.6 | 18.9 | 93.3 | 28.5 | 69.2 | 0.736 |
DenseNet-121 | 75.7 | 88.0 | 43.1 | 96.8 | 54.9 | 86.7 | 0.927 |
ResNet-50 | 78.7 | 89.1 | 46.4 | 97.2 | 58.4 | 88.0 | 0.877 |
ResNext-50 | 75.7 | 88.7 | 44.6 | 96.8 | 56.1 | 87.3 | 0.912 |
CheXNet | 78.7 | 85.8 | 39.9 | 97.1 | 53.0 | 85.1 | 0.884 |
EfficientNet -B3 | 81.8 | 77.1 | 29.9 | 97.2 | 43.9 | 77.6 | 0.873 |
VGG-19 | 63.6 | 80.4 | 27.9 | 84.9 | 38.8 | 78.6 | 0.812 |
