Table 4.
ROC curve analysis for significant comparisons. Only the scale with higher AUC is indicated
| Table 4 | lvPPA-HC | nfvPPA-HC | svPPA-HC | uPPA-HC | svPPA-lvPPA | svPPA- nfvPPA | svPPA-uPPA |
|---|---|---|---|---|---|---|---|
| Best Scale | ATL | PAL | ATL | ATL | ATL | ATL | ATL |
| Cutpoint | 1 | 2 | 2 | 1 | 3 | 2 | 3 |
| Sensitivity (%) | 88.64% | 57.89% | 100% | 100% | 83.87% | 100% | 83.87% |
| Specificity (%) | 86.67% | 88.89% | 97.78% | 86.67% | 95.45% | 73.68% | 81.82% |
| PPV (%) | 86.67% | 68.75% | 96.88% | 64.71% | 92.86% | 86.11% | 92.86% |
| NPV (%) | 88.64% | 83.33% | 100% | 100% | 89.36% | 100% | 64.29% |
| AUC | 0.921 | 0.817 | 0.999 | 0.953 | 0.940 | 0.925 | 0.868 |
ATL Anterior temporal left, PAL Posterior left, lvPPA Logopenic, nfvPPA Nonfluent, svPPA Semantic, uPPA Undetermined, HC Controls. PPV Positive predictive value, NPV Negative predictive value