From: Preoperative prediction of lymph node metastasis using deep learning-based features
 | Training (n = 109, 37 positive SLNs) | ||||||
AUC | Sensitivity | Specificity | PPV | NPV | Accuracy | YI | |
CR | 0.91 | 0.89 | 0.82 | 0.82 | 0.89 | 0.85 | 0.71 |
DLB | 0.93 | 0.89 | 0.86 | 0.86 | 0.89 | 0.88 | 0.75 |
 | Validation (n = 54, 18 positive SLNs) | ||||||
AUC | Sensitivity | Specificity | PPV | NPV | Accuracy | YI | |
CR | 0.87 | 0.72 | 0.83 | 0.68 | 0.86 | 0.80 | 0.56 |
DLB | 0.89 | 0.83 | 0.83 | 0.71 | 0.91 | 0.83 | 0.67 |
 | Testing (n = 35, 12 positive SLNs) | ||||||
AUC | Sensitivity | Specificity | PPV | NPV | Accuracy | YI | |
CR | 0.77 | 0.58 | 0.78 | 0.58 | 0.78 | 0.71 | 0.37 |
DLB | 0.83 | 0.58 | 0.87 | 0.70 | 0.80 | 0.77 | 0.45 |