Skip to main content

Table 2 ROC curve area comparison

From: Performance of a multiplexed dual analyte immunoassay for the early detection of non-small cell lung cancer

A. From a logistic regression analysis

Biomarker

Transformation

AUC a

Standard error

95% Wald Confidence Limits

Probability b

Lower

Upper

NY-ESO-1

Raw

0.60

0.04

0.52

0.67

0.0102

CEA

Raw

0.79

0.03

0.74

0.85

<0.0001

CA125

Raw

0.70

0.03

0.63

0.77

<0.0001

CYFRA 21-1

Raw

0.69

0.04

0.62

0.76

<0.0001

4 Biomarker Modelc

Raw

0.83

0.03

0.78

0.88

<0.0001

B. From a MoM analysis

Biomarker

Transformation

AUC a

Standard error

95% Confidence Interval

Probability b

Lower

Upper

NY-ESO-1

Raw

0.60

0.04

0.52

0.67

0.01007

CEA

Raw

0.79

0.03

0.74

0.85

<0.0001

CA125

Raw

0.70

0.03

0.63

0.77

<0.0001

CYFRA 21-1

Raw

0.69

0.04

0.62

0.76

<0.0001

4 Biomarker Modelc

Raw

0.81

0.03

0.75

0.86

<0.0001

  1. aAUC.
  2. bProbability of difference to AUC = 0.5.
  3. c4 Biomarker model = NY-ESO-1, CEA, CA125, and CYFRA 21-1.