Table 3.
Results of ROC curve analysis and contingency table analysis for NT-proBNP and PFT
| Variable | Comparison groups |
Optimal cut point | Sensitivity (95% CI) |
Specificity (95% CI) |
+LR | -LR | AUC |
|---|---|---|---|---|---|---|---|
| NT-proBNP (ng/ml) |
PAH versus control | ≥ 209.8 | 92.9% (64.1%-99.6%) |
100% (85.9%-99.7%) |
- | 0.07 (0.01-0.47) |
0.93 (0.80-1.0) |
| PAH versus ILD | ≥ 360.5 | 85.7% (56.2%-97.0%) |
100.0% (71.8%-99.5%) |
- | 0.14 (0.04-0.52) |
0.92 (0.78-1.0) |
|
| DLCOcorr% | PAH versus control | < 70.3 | 100% (73.2%-99.3%)a |
100% (85.7%-99.7%)a |
- | - | 1.0 |
| PAH versus ILD | Nil | - | - | - | - | 0.42 (0.22-0.63) |
|
| FVC/DLCO | PAH versus control | ≥ 1.66 | 64.3% (35.6%-86.0%) |
96.7% (80.9%-99.8%) |
19.3 (2.7- .) |
0.37 (0.18-0.75) |
0.90 (0.78-1.0) |
| PAH versus ILD | ≥ 1.82 | 50.0% (24.0%-76.0%) |
94.4% (70.6%-99.7%) |
9.0 (1.2-64.9) |
0.53 (0.31-0.90) |
0.71 (0.51-0.90) |
|
| DLCO < 70.3% and FVC/DLCO ≥ 1.82 |
PAH versus control | - | 50.0% (24.0%-76.0%) |
100% (85.9%-99.7%) |
- | 0.63 (0.43-0.91) |
- |
| PAH versus ILD | - | 50.0% (24.0%-76.0%) |
94.4% (70.6%-99.7%) |
6.75 (0.9-50.2) |
0.66 (0.45-0.98) |
- | |
| DLCO < 70.3% and FVC/DLCO ≥ 1.82 (A) and/or NT-proBNP ≥ 209.8 (B) |
PAH versus control | - | 100% (73.2%-99.3%)a |
100% (85.9%-99.7%)a |
- | - | - |
| PAH versus ILD | - | 100% (73.2%-99.3%)a |
77.8% (51.9%-92.6%) |
4.50 (1.90-10.7) |
- | - | |
DLCO, diffusing capacity of lungs for carbon monoxide, % predicted; FVC, forced vital capacity, % predicted; ILD, interstitial lung disease; NT-proBNP, N-terminal pro brain natriuretic peptide; PAH, pulmonary arterial hypertension; (A) and (B) refer to the two components of the "composite" screening model. aBecause of the relatively small sample size, the upper limit of the 95% CI approximates but does not equal 100%.