Skip to main content

Table 2 The predictive power of multiple biomarkers in the diagnosis of patients with RA-ILD vs. RA-non-ILD

From: Identification of biomarkers by machine learning classifiers to assist diagnose rheumatoid arthritis-associated interstitial lung disease

 

Cut-off

AUC

(95%CI)

P

Sen

(95%CI)

Spe

(95%CI)

 + LR

(95%CI)

 − LR

(95%CI)

Youden

index

Z

P′

Age

(years)

 > 49

0.814

(0.731–0.880)

 < 0.001

93.33

(85.10–97.80)

67.95

(56.40–78.10)

2.91

(2.10–4.00)

0.10

(0.04–0.20)

0.613

3.248

0.001

KL-6

(U/mL)

 > 373.65

0.749

(0.660–0.824)

 < 0.001

61.33

(49.40–72.40)

78.21

(67.40–86.80)

2.81

(1.80–4.40)

0.49

(0.40–0.70)

0.396

4.256

 < 0.001

D-dimer (ng/mL)

 > 716

0.749

(0.660–0.824)

 < 0.001

59.46

(47.40–70.70)

82.19

(71.50–90.20)

3.34

(2.00–5.70)

0.49

(0.40–0.70)

0.417

4.196

 < 0.001

CA199

(KU/L)

 > 8.01

0.727

(0.637–0.805)

 < 0.001

56.52

(44.00–68.40)

77.36

(63.80–87.70)

2.50

(1.50–4.30)

0.56

(0.40–0.80)

0.339

4.523

 < 0.001

All

 

0.928

(0.865–0.968)

 < 0.001

83.82

(72.90–91.60)

81.63

(68.00–91.20)

4.56

(2.50–8.30)

0.20

(0.10–0.30)

0.655

  
  1. P value indicates that the AUC of each indicator has statistical significance. Z and P′ indicated that the AUC of each indicator had a statistical difference with the AUC of combined detection of the four indicators
  2. Sen, sensitivity; Spe, specificity; LR, likelihood ratio