Table III.
Receiver operating characteristic and accuracy findings of four electrodiagnostic parameters in detecting carpal tunnel syndrome.
| Variables | Diagnostic test | ROC curve | p-value* | ||||
|---|---|---|---|---|---|---|---|
| Cut-off | Sensitivity | Specificity | PPV | NPV | AUC | ||
| Terminal Latency Index | ≤ 0.368 | 75.4 (62.2 to 85.9) | 87.7 (76.3 to 94.9) | 86.0 (75.1 to 92.6) | 78.1 (69.2 to 85.0) | 0.88 (0.81 to 0.94) | 0.001 |
| Residual latency | > 2.245 | 85.9 (74.2 to 93.7) | 82.5 (70.1 to 91.3) | 83.3 (73.8 to 89.8) | 85.4 (75.3 to 91.8) | 0.93 (0.87 to 0.97) | 0.001 |
| MSUMLD | > 0.1 | 94.8 (85.6 to 98.9) | 60.0 (45.9 to 73.0) | 70.5 (63.4, 76.7) | 91.9 (78.7 to 84.7) | 0.88 (0.75 to 0.89) | 0.001 |
| FdifMU | 0.5 | 87.9 (76.7 to 95.0) | 92.9 (83.0 to 98.1) | 92.7 (83.1 to 97.0) | 88.3 (79.0 to 93.8) | 0.93 (087 to 0.97) | 0.001 |
*
Receiver operating characteristic (ROC) curve was used to estimate optimal cutoff points and p-values for studied parameters. The optimal cutoff point was estimated with regard to the highest possible logical specificity and sensitivity.
AUC, area under curve; FdifMU, median ulnar F-wave latency difference; MSUMLD, median sensory latency-ulnar motor latency difference; NPV, negative predictive value; PPV, positive predictive value.