Table 2.
Summary of sensitivity and specificity across algorithms.
| Algorithm | Studies reporting sensitivity | Mean sensitivity (%) | Sensitivity ranka | Studies reporting specificity | Mean specificity (%) | Specificity ranka |
|---|---|---|---|---|---|---|
| ResNet | 1 | 94.0 | 1 | 1 | 93.0 | 1 |
| DenseNet | 2 | 91.5 | 2 | 2 | 85.5 | 5 |
| LDS | 1 | 89.0 | 3 | 1 | 52.0 | 8 |
| SVM | 5 | 86.8 | 4 | 5 | 86.6 | 3 |
| MP | 1 | 85.0 | 5 | 1 | 90.0 | 2 |
| Bayes | 2 | 83.0 | 6 | 2 | 86.0 | 4 |
| CNN | 2 | 80.0 | 7 | 2 | 82.0 | 7 |
| ELM | 1 | 69.0 | 8 | 1 | 84.0 | 6 |
| FFR | 0 | – | – | 0 | – | – |
| NN | 0 | – | – | 0 | – | – |
| LR | 0 | – | – | 0 | – | – |
a
Ranks are assigned in descending order of the mean sensitivity or specificity (1 = highest). Algorithms with no reported values are not ranked.