Skip to main content
. 2017 Oct 13;12(10):e0185855. doi: 10.1371/journal.pone.0185855

Table 2. Shows the AUC, sensitivity, and specificity, of each ROC-analysis and percentage of cases correctly identified, and combination of biomarkers.

AUC (CI) P values Sens. % Spec. % PPV. % NPV. % DA %
Biomarkers: UC (n = 60) vs. CD (n = 72)
 BGM 0.95 (0.90–0.98) <0.001 93 87 86 94 90
 EL-NE 0.85 (0.78–0.91) <0.001 78 78 74 81 75
 C5M 0.61 (0.52–0.69) <0.05 48 76 63 64 55
 Pro-C5 0.65 (0.56–0.73) <0.01 85 43 55 78 60
 VICM, see Mortensen et al. [14]
 CRP 0.56 (0.47–0.65) NS 100 39 56 100 54
 BGM + VICM* 0.98 (0.93–1.00) <0.001 94 96 94 96 94
Biomarkers: UC (n = 60) vs. Colonic CD (n = 48)
 BGM 0.94 (0.88–0.98) <0.001 83 95 93 88 88
 EL-NE 0.87 (0.79–0.93) <0.001 60 93 88 74 87
 C5M 0.62 (0.53–0.72) NS 79 52 57 76 52
 Pro-C5 0.69 (0.59–0.78) <0.01 79 48 55 74 69
 VICM, see Mortensen et al. [14]
 CRP 0.52 (0.39–0.69) NS 45 54 43 60 53
 BGM+VICM* 0.97 (0.92–1.0) <0.001 96 91 90 96 94
Biomarkers: IBD (n = 132) vs. IBS (n = 22)
 BGM 0.70 (0.62–0.78) <0.001 60 91 98 28 85
 EL-NE 0.77 (0.69–0.83) <0.001 78 65 93 34 85
 C5M 0.56 (0.48–0.64) NS 62 57 89 21 85
 Pro-C5 0.71 (0.59–0.83) <0.001 78 61 92 33 85
 VICM 0.59 (0.51–0.66) NS 83 41 89 29 85
 CRP 0.51 (0.43–0.59) NS 51 58 88 17 85
 EL-NE+BGM+Pro-C5* 0.83 (0.76–0.88) <0.001 68 91 98 32 85
Biomarkers: IBD (n = 132) vs. healthy donors (n = 24)
 BGM 0.61 (0.53–0.69) <0.01 46 79 92 21 85
 EL-NE 0.65 (0.57–0.72) <0.01 52 96 99 27 85
 C5M 0.91 (0.85–0.95) <0.001 84 92 98 51 85
 Pro-C5 0.93 (0.88–0.96) <0.001 86 88 97 54 85
 C5M+Pro-C5+EL-NE* 0.98 (0.95–0.99) <0.001 86 88 97 54 85
Biomarkers: IBS (n = 22) vs. healthy donors (n = 24)
 BGM 0.64 (0.49–0.78) <0.05 91 58 67 88 64
 EL-NE 0.75 (0.60–0.86) NS 68 96 83 75 76
 C5M 0.92 (0.80–0.98) <0.001 96 83 84 96 92
 Pro-C5 0.83 (0.69–0.92) <0.001 96 63 70 94 83
 C5M+Pro-C5+EL-NE* 0.96 (0.86–1.0) <0.001 96 88 88 96 98

Abbreviations: AUC: Area under the curve, CI: Confidence interval, Sens: Sensitivity, Spec: Specificity, PPV: Positive predictive value, NPV: Negative predictive value, DA: Diagnostic accuracy

* Best logistic regression model with combination of biomarkers.

sample size is too low for the control groups (IBS and healthy donors) resulting in overestimation of the diagnostic value.