TABLE 3.
Classification performance of our proposed method on single-center dataset.
| Sites | Accuracy (%) | Sensitivity (%) | Specificity (%) | Precision (%) |
| NYU | 71.4 ± 5.7 | 75.0 ± 1.5 | 68.4 ± 9.4 | 66.7 ± 9.8 |
| OHSU | 80.8 ± 3.8 | 76.9 ± 7.7 | 84.6 ± 7.7 | 83.3 ± 8.3 |
| KKI | 62.5 ± 8.3 | 63.6 ± 9.1 | 61.5 ± 7.7 | 58.3 ± 8.4 |
| YALE | 71.4 ± 3.6 | 74.0 ± 2.9 | 68.9 ± 4.4 | 68.9 ± 2.5 |
| USM | 81.6 ± 2.9 | 84.8 ± 4.1 | 76.0 ± 0.9 | 86.6 ± 0.3 |
| Olin | 76.4 ± 3.0 | 75.0 ± 3.9 | 78.6 ± 1.4 | 83.3 ± 0.0 |
| Pitt | 73.2 ± 1.8 | 75.0 ± 2.6 | 71.4 ± 6.4 | 72.4 ± 5.4 |
| Leuven | 74.6 ± 3.2 | 68.8 ± 5.4 | 80.6 ± 0.7 | 78.6 ± 0.7 |
| UCLA | 79.5 ± 3.1 | 77.8 ± 3.7 | 75.0 ± 2.3 | 79.2 ± 2.3 |
| Caltech | 67.6 ± 2.7 | 71.4 ± 0.0 | 62.5 ± 6.3 | 71.4 ± 3.6 |
| CMU | 66.7 ± 3.6 | 64.3 ± 7.1 | 69.2 ± 0.0 | 69.2 ± 2.2 |
| MaxMun | 65.4 ± 3.8 | 69.2 ± 0.3 | 61.5 ± 6.7 | 64.3 ± 4.9 |
| SBL | 66.7 ± 3.4 | 71.4 ± 1.2 | 62.7 ± 6.2 | 62.5 ± 4.9 |
| SDSU | 61.6 ± 2.7 | 60.1 ± 5.7 | 67.9 ± 3.8 | 65.6 ± 3.2 |
| Stanford | 67.7 ± 4.2 | 51.6 ± 5.7 | 76.9 ± 5.3 | 64.2 ± 3.7 |
| UM | 63.8 ± 5.6 | 70.5 ± 4.3 | 53.2 ± 5.7 | 75.6 ± 4.9 |
| Trinity | 68.8 ± 5.5 | 72.9 ± 4.6 | 60.5 ± 5.7 | 73.3 ± 3.7 |
| AVERAGEE | 70.6 ± 4.0 | 71.7 ± 4.1 | 69.4 ± 4.8 | 71.7 ± 4.0 |