Table 2. ISREA nodes and prediction metrics for detecting one disease among a population of five glomerular diseases (DKD, IMN, MN, RN, OT).
| Glomerular disease class | ISREA nodes (cm−1) | Random-chance accuracy* | Prediction accuracy | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|---|
| DKD | 400, 917, 940, 1,058, 1,558, 1,800 | 50% [35.3, 64.7] | 76.5% | 75.0% | 77.2% | 64.3% | 85.0% |
| IMN | 400, 510, 1,121, 1,488, 1,618, 1,623, 1,704, 1,800 | 50% [35.3, 64.7] | 82.4% | 83.3% | 81.8% | 71.4% | 90.0% |
| MN | 400, 463, 1,460, 1,483, 1,508, 1,681, 1,800 | 50% [17.7, 82.3] | 91.2% | 66.7% | 96.4% | 80.0% | 93.1% |
| RN | 400, 1,202, 1,232, 1,271, 1,323, 1,516, 1,717, 1,800 | 50% [8.8, 91.2] | 100% | 100% | 100% | 100% | 100% |
Note:
*
The values in brackets [ ] represent the calculated accuracies if either all positive or all negative predictions were returned.