Table 3.
The prediction performances of models based on T1 + T2 and T1 + CEST.
| Input | ACC (%) | SEN (%) | SPE (%) | AUC |
|---|---|---|---|---|
| T1 + CEST | 80.03 ± 0.18** | 82.45 ± 2.55*** | 77.65 ± 4.97 | 0.8553 ± 0.0152* |
| T1 + T2 | 75.59 ± 1.82 | 74.77 ± 2.21 | 76.38 ± 3.48 | 0.8193 ± 0.0218 |
| T1 + CEST + annotation | 82.25 ± 1.57** | 83.36 ± 4.05* | 81.15 ± 4.08 | 0.8839 ± 0.0181* |
| T1 + T2 + annotation | 77.30 ± 2.40 | 75.89 ± 5.48 | 78.53 ± 4.48 | 0.8496 ± 0.0219 |
“*” indicates significant improvements of the models based on T1 + CEST over T1 + T2 at p < 0.05 level, based on one-sided paired t-test. **p<0.01; ***p<0.001.