Table 2.
Performance measures of binary classifications with the chosen CPPLS-DA model.
| AUCa | Sensitivity | Specificity | Balanced accuracyb | ||
|---|---|---|---|---|---|
| SCN vs All | Cyst fluid | 1.000 | 1.000 | 1.000 | 1.000 |
| Plasma | 0.950 | 0.800 | 0.875 | 0.837 | |
| LGD vs All | Cyst fluid | 0.935 | 0.875 | 0.913 | 0.894 |
| Plasma | 0.825 | 1.000 | 0.812 | 0.906 | |
| HGD-Cancer vs All | Cyst fluid | 0.949 | 0.889 | 0.923 | 0.906 |
| Plasma | 0.854 | 0.636 | 1.000 | 0.818 | |
| IPMN vs SCN | Cyst fluid | 1.000 | 1.000 | 1.000 | 1.000 |
| Plasma | 1.000 | 1.000 | 1.000 | 1.000 | |
aArea Under the ROC Curve; b(Sensitivity + Specificity)/2.