Table 1.
Summary of detection performance using different networks.
| Pathologic Grade I (total 55) | Pathologic Grade II (total 41) | |||
|---|---|---|---|---|
| Model type | ACR/SCR/TCAB/TCA/TCB | ACR/SCR/TCAB/TCA/TCB | ||
| Predicted grade I | 51/44/38/40/16 | 6/10/16/12/26 | ||
| Predicted grade II | 4/11/17/15/39 | 35/31/25/29/15 | ||
| Model type | Accuracy | Sensitivity | Specificity | AUC |
|---|---|---|---|---|
| ACR | 0.90 (0.80–1.00) | 0.85 (0.67–1.00) | 0.93 (0.75–1.00) | 0.91 |
| SCR | 0.78 (0.70–1.00) | 0.76 (0.57–1.00) | 0.80 (0.75–1.00) | 0.83 |
| TCAB | 0.66 (0.50–0.90) | 0.61 (0.50–1.00) | 0.69 (0.50–1.00) | 0.73 |
| TCA | 0.72 (0.60–0.90) | 0.71 (0.57–1.00) | 0.73 (0.50–1.00) | 0.79 |
| TCB | 0.32 (0–0.50) | 0.37 (0–0.57) | 0.29 (0–0.50) | 0.45 |
ACR, SCR, TCAB, TCA, and TCB stand for asymmetric CNN with ratio (18:2), symmetric CNN with ratio (10:10), traditional CNN with T1-CE and T2-FLAIR, traditional CNN with T1-CE, and traditional CNN with T2-FLAIR, respectively.