Table 2.
Receiver operating characteristic curve analysis of the accuracy of biomarkers for identifying parapneumonic effusions.
| Biomarker | CRP optimal cut-off value (mg/dL) | Sensitivity (%) |
Specificity (%) | PPV∗ (%) | NPV∗∗ (%) | AUC∗∗∗ |
|---|---|---|---|---|---|---|
| PE† versus Mal‡, HF§, and LTx¶ | >1.38 | 84.2 | 71.5 | 37.6 | 96.7 | 0.85 |
| PE versus LTx | >1.93 | 75.7 | 56.5 | 71.4 | 61.9 | 0.67 |
| PE versus Mal | >0.88 | 87.8 | 64.7 | 40.8 | 95 | 0.84 |
| PE versus HF | >0.49 | 93.9 | 72.7 | 72 | 94.1 | 0.92 |
| HF versus Mal, PE, and LTx | <0.64 | 79.5 | 59.4 | 32.4 | 92.2 | 0.76 |
| Emp£ versus Mal, LTx, and uncomplicated PE | >2.31 | 83.3% | 74.7% | 8.3% | 99.3% | 0.7 |
†PE: parapneumonic effusions; ‡Mal: malignancy; £Emp: empyema;§ HF: heart failure; ¶LTx: lung transplant; ∗PPV: positive predictive value; ∗∗NPV: negative predictive value; ∗∗∗AUC: area under the curve.