Table 5.
ROC analysis for predicting the absence of biochemical remission when using preoperative PRL level as a diagnostic marker.
| Parameter | All patients (n = 109) | Micro (n = 75) | Macro (n = 34) |
|---|---|---|---|
| Analysis after 1 month | |||
| Cut-off value | >299.5 | >200 | >345 |
| Sensitivity, % | 67.5 | 44.2 | 70.5 |
| Specificity, % | 66.6 | 82.6 | 88.2 |
| NPV, % | 77.9 | 85.1 | 85.7 |
| PPV, % | 54.0 | 39.5 | 75.0 |
| Diagnostic accuracy, % | 66.9 | 56.0 | 79.4 |
| AUROC | 0.680 | 0.616 | 0.744 |
| 95% CI | 0.577–0.783 | 0.487–0.745 | 0.561–0.927 |
| P (AUROC) | 0.002 | 0.112 | 0.015 |
| Analysis after 12 months | |||
| Cut-off value | >351 | >210 | >396 |
| Sensitivity, % | 60.0 | 53.3 | 90.0 |
| Specificity, % | 71.4 | 36.6 | 62.5 |
| NPV, % | 85.7 | 75.8 | 93.7 |
| PPV, % | 38.4 | 17.3 | 50.0 |
| Diagnostic accuracy, % | 68.8 | 40.0 | 70.5 |
| AUROC | 0.599 | 0.477 | 0.725 |
| 95% CI | 0.459–0.738 | 0.293–0.660 | 0.558–0.892 |
| P (AUROC) | 0.136 | 0.781 | 0.041 |
NPV, negative predictive value; PPV, positive predictive value; AUROC, area under ROC-curve, 95% CI—95% confidence interval for AUROC.