Table 2.
Diagnostic ability of the developed models.
| Model | Training cohort (n = 204) | Validation cohort (n = 88) | ||||||
|---|---|---|---|---|---|---|---|---|
| AUC (95% CI) | ACC | SEN | SPE | AUC (95% CI) | ACC | SEN | SPE | |
| Clinical model |
0.841 (0.787, 0.896) |
0.760 | 0.734 | 0.790 | 0.766 (0.667, 0.863) |
0.659 | 0.674 | 0.643 |
| T1_tumor signature |
0.859 (0.810, 0.908) |
0.819 | 0.972 | 0.642 | 0.818 (0.732, 0.904) |
0.716 | 0.891 | 0.524 |
| T1_edema signature |
0.768 (0.706, 0.829) |
0.716 | 0.927 | 0.474 | 0.673 (0.569, 0.777) |
0.648 | 0.957 | 0.310 |
| T2_tumor signature |
0.858 (0.809, 0.907) |
0.794 | 0.927 | 0.642 | 0.762 (0.666, 0.858) |
0.693 | 0.849 | 0.524 |
| T2_edema signature |
0.787 (0.727, 0.846) |
0.750 | 0.936 | 0.537 | 0.711 (0.613, 0.809) |
0.682 | 0.957 | 0.381 |
| CE-T1_tumor signature |
0.811 (0.752, 0.870) |
0.755 | 0.725 | 0.789 | 0.731 (0.628, 0.835) |
0.648 | 0.630 | 0.667 |
| CE-T1_edema signature |
0.760 (0.699, 0.821) |
0.770 | 0.982 | 0.526 | 0.734 (0.652, 0.817) |
0.659 | 1.000 | 0.286 |
| Tumor signature | 0.917 (0.877,0.958) |
0.878 | 0.973 | 0.768 | 0.872 (0.799, 0.944) |
0.750 | 0.848 | 0.643 |
| Edema signature | 0.808 (0.747, 0.869) |
0.775 | 0.973 | 0.547 | 0.704 (0.597, 0.811) |
0.659 | 0.978 | 0.310 |
| Fusion signature | 0.979 (0.959, 0.999) |
0.956 | 0.991 | 0.916 | 0.902 (0.841, 0.964) |
0.818 | 0.891 | 0.738 |
| HMDT | 0.985 (0.968, 1) |
0.961 | 0.973 | 0.947 | 0.917 (0.861, 0.972) |
0.852 | 0.848 | 0.857 |
AUC, area under the curve; CI, confidence interval; ACC, accuracy; SEN, sensitivity; SPE, specificity.