Table 5.
Diagnostic performance of different prediction models in the diagnosis of breast cancer and benign diseases.
| Model | ROC | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | AUC | 95% CI | Spe. (%) | Sen. (%) | PPV (%) | NPV (%) | Accuracy (%) | |
| Model 1 | BI-RADS | 0.856 | 0.769,0.943 | 76.67 | 84.44 | 84.44 | 76.67 | 81.33 |
| Model 2 | ΔT1%, T2 | 0.872 | 0.782,0.962 | 86.67 | 86.67 | 90.70 | 81.25 | 86.67 |
| Model 3 | ΔT1%, T2, ADC margin, internal enhancement pattern |
0.962 | 0.925,0.999 | 93.33 | 91.11 | 95.35 | 87.50 | 92.00 |
| Model 4 | ΔT1%, T2, ADC margin, internal enhancement pattern age, BMI, menopausal state |
0.989 | 0.974,0.999 | 96.67 | 95.56 | 97.73 | 93.55 | 96.00 |
Sen., sensitivity; Spe., specificity; PPV, positive predictive value; NPV, negative predictive value.