Table 3 SLNB reduction rates using the ANN model to predict disease-free axilla. Possible SLNB reduction rate corresponding to cut-offs at maximum negative predictive value, false negative rate 5 and 10%, respectively
From: Artificial neural network models to predict nodal status in clinically node-negative breast cancer
N0 vs. N+ n = 800 | ||||
|---|---|---|---|---|
Cut-off Max NPV 0.95 | TP | TN | FP | FN |
No. | 283 | 57 | 457 | 3 |
SLNB Reduction Rate | (TN + FN) / (TP + TN + FP + FN) = 7.50% | |||
False Negative Rate | FN / (TP + FN) = 1.05% | |||
Cut-off NPV 0.90 | TP | TN | FP | FN |
No. | 272 | 128 | 386 | 14 |
SLNB Reduction Rate | (TN + FN) / (TP + TN + FP + FN) = 17.75% | |||
False Negative Rate | FN / (TP + FN) = 5% | |||
Cut-off NPV 0.87 | TP | TN | FP | FN |
No. | 258 | 190 | 324 | 28 |
SLNB Reduction Rate | (TN + FN) / (TP + TN + FP + FN) = 27.25% | |||
False Negative Rate | FN / (TP + FN) = 10% | |||