TABLE 3.
Optimum cutoffs based on the ROC curves and their performancesa
| Assay (n) | Cutoff | Sensitivity | Specificity | Overall agreement (%) |
|---|---|---|---|---|
| Genscript cPass (125) | Manufacturer’s cutoff, 30% | 71% | 94.23% | 81.6% |
| Based on ROC cutoff, 20% | 92.75% | 87.5% | 90.4% | |
| P value | <0.001 | 0.065 | <0.05 | |
| Diasorin-TrimericS IgG (103) | Manufacturer’s cutoff, 33.8 BAU/ml | 84.91% | 90% | 87.37% |
| Based on ROC cutoff, 40 BAU/ml | 84.90% | 96% | 90.29% | |
| P value | 0.998 | 0.092 | 0.506 | |
| Roche Elecsys (96) | Manufacturer’s cutoff, 0.8 U/ml | 100% | 41.67% | 70.83% |
| Based on ROC cutoff, 5 U/ml | 93.75% | 70.83% | 82.29% | |
| P value | <0.05 | <0.001 | 0.0616 |
a
The other three tests showed ROC-based cutoffs similar to the manufacturer’s values.