Table 1 Summary of the different AI-ECG algorithms of included studies
Study | Year | Machine learning technique | AUC | Specificity (%) | Sensitivity (%) |
|---|---|---|---|---|---|
Adedinsewo et al. [21] | 2020 | Convolutional neural network | 0.890 | 87 | 74 |
Attia et al. [47] | 2021 | Convolutional and residual neural network | 0.767 | 10.2 | 98 |
Cohen-Shelly et al. [37] | 2021 | Convolutional neural network | 0.850 | 74 | 78 |
Cordeiro et al. [42] | 2021 | Deep neural network | 0.945 | 85 | 87.6 |
Kwon et al. [41] | 2021 | Residual neural network | 0.901 | – | – |
Kwon et al. [45] | 2020 | Convolutional neural network | 0.873 | – | – |
Lin et al. [44] | 2021 | Convolutional neural network | 0.986 | 69.2 | 88.9 |
Potter et al. [20] | 2021 | Random forest classifier | 0.830 | 72 | 85 |
Rabinstein et al. [22] | 2021 | – | – | 75 | 63 |
Shrivastava et al. [16] | 2021 | Convolutional neural network | 0.955 | 44.8 | 98.8 |
Siontis et al. [15] | 2021 | Convolutional neural network | 0.980 | 95 | 92 |