Skip to main content

Table 4 Diagnostic performance of biomarkers alone and in combination with clinical factors in the development cohort (JHMI/UTSMC) and in the model 5 in the validation cohort (NCI EDRN)

From: Validation of a novel model for the early detection of hepatocellular carcinoma

Model

Predictor variables

AUC

AUC 95% CI

SE

SP

SE (SP = 0.90)

SP (SE = 0.90)

SP (SE = 0.75)

Development cohort

 1

AFP

0.88

0.84–0.93

0.86

0.77

0.64

0.64

0.81

 2

PIVKA-II

0.87

0.82–0.90

0.86

0.72

0.51

0.65

0.77

 3

Age, gender, AFP

0.93

0.90–0.96

0.94

0.76

0.78

0.78

0.91

 4

Age, gender, PIVKA-II

0.91

0.87–0.94

0.93

0.72

0.67

0.72

0.86

 5

Age, gender, AFP, PIVKA-II

0.95

0.93–0.98

0.93

0.84

0.84

0.83

0.97

Validation cohort

 5a

Age, gender, AFP, PIVKA-II

All

0.87

0.85–0.90

0.74

0.85

0.67

0.54

0.84

Viral etiology

0.86

0.83–0.89

0.79

0.80

0.62

0.61

0.82

Non-viral etiology

0.87

0.83–0.91

0.75

0.91

0.75

0.58

0.91

 5b

Age, gender, AFP, PIVKA-II

All

0.85

0.81–0.88

0.70

0.86

0.63

0.51

0.79

Viral etiology

0.86

0.82–0.89

0.81

0.80

0.60

0.61

0.83

Non-viral etiology

0.83

0.77–0.90

0.68

0.91

0.68

0.57

0.74

  1. aAll cancers
  2. bEarly-stage cancers (BCLC stage 0 and A)