Page 26 - EASL POSTGRADUATE COURSE
P. 26
Table 1. Validated fibrosis tests in NAFLD based upon routinely-available clinical and
biochemical variables. Adapted from Castera et al. [10].
Test n Parameters AUROC Cut-offs Sens Spec PPV NPV
NFS 733 IFG/diabetes, AST/ALT, 0.82-0.88 <-1.455 77% 71% 52% 88%
Age, BMI, platelets,
FIB-4 albumin 0.80 >0.676 43% 96% 82% 80%
<1.30 74% 71% 43% 90%
BARD 541 ALT, AST, platelets, Age 0.81
APRI 0.82 >2.67 33% 98% 80% 83%
827 BMI, AST/ALT, diabetes 2 - - 43% 96%
576 AST, platelets 1.0 67% 81% 31% 95%
NAFLD 235 Glucose, ALT, AST, 0.93-0.94 78% 96% 88% 92%
Fibrometer weight, age, platelets,
ferritin
Fibrotest 267 Age, sex, bilirubin, 0.81-0.92 >0.30 92% 71% 33% 98%
GGT, apolipoprotein A1, >0.70 25% 97% 60% 89%
haptoglobin,
α2-macroglobulin
Values based on the prediction of advanced fibrosis. Original studies with validation cohorts presented.
Sens = sensitivity; Spec = specificity.
Overall, simple serum-based tests are reasonably accurate at excluding cirrhosis with NPVs consistently
above 90%. More complex serum tests with multiple covariates have higher AUROC values and are
better at predicting or excluding lesser degrees of fibrosis such as F3+ (Table 1) [5].
Algorithms which incorporate serum measures of factors directly involved in fibrogenesis (e.g.
hyaluronate) have the theoretical advantage of being more specific for fibrosis. The accuracy of these
algorithms (outlined in Table 2) appears similar to algorithms based on multiple routine clinical and
biochemical variables, however direct comparative studies are lacking.
Table 2.Validated fibrosis tests in NAFLD based upon direct markers of fibrogenesis. Adapted
from Castera et al. [10].
Test n Parameters AUROC Cut-offs Sens Spec PPV NPV
Hyaluronate 112 Hyaluronate 0.80 50 ng/ml 69% 83% 75% 84%
Type IV collagen 7S 0.82 5 ng/ml 81% 71% 68% 78%
Type IV collagen 7S 0.97 89% 77% 100%
Hyaluronate 0.90 42 ng/ml 100% 90% 71% 94%
Hyaluronate 148
TIMP-1, 0.81 0.3576 80% 84% 57% 92%
ELF 192 hyalularonic acid,
terminal peptide of
Hepascore 242 procollagen III 0.37 76%
Hyaluronate,
α2-macroglobulin,
bilirubin, GGT, age,
sex
26 Postgraduate Course Syllabus • Metabolic Liver Disease
biochemical variables. Adapted from Castera et al. [10].
Test n Parameters AUROC Cut-offs Sens Spec PPV NPV
NFS 733 IFG/diabetes, AST/ALT, 0.82-0.88 <-1.455 77% 71% 52% 88%
Age, BMI, platelets,
FIB-4 albumin 0.80 >0.676 43% 96% 82% 80%
<1.30 74% 71% 43% 90%
BARD 541 ALT, AST, platelets, Age 0.81
APRI 0.82 >2.67 33% 98% 80% 83%
827 BMI, AST/ALT, diabetes 2 - - 43% 96%
576 AST, platelets 1.0 67% 81% 31% 95%
NAFLD 235 Glucose, ALT, AST, 0.93-0.94 78% 96% 88% 92%
Fibrometer weight, age, platelets,
ferritin
Fibrotest 267 Age, sex, bilirubin, 0.81-0.92 >0.30 92% 71% 33% 98%
GGT, apolipoprotein A1, >0.70 25% 97% 60% 89%
haptoglobin,
α2-macroglobulin
Values based on the prediction of advanced fibrosis. Original studies with validation cohorts presented.
Sens = sensitivity; Spec = specificity.
Overall, simple serum-based tests are reasonably accurate at excluding cirrhosis with NPVs consistently
above 90%. More complex serum tests with multiple covariates have higher AUROC values and are
better at predicting or excluding lesser degrees of fibrosis such as F3+ (Table 1) [5].
Algorithms which incorporate serum measures of factors directly involved in fibrogenesis (e.g.
hyaluronate) have the theoretical advantage of being more specific for fibrosis. The accuracy of these
algorithms (outlined in Table 2) appears similar to algorithms based on multiple routine clinical and
biochemical variables, however direct comparative studies are lacking.
Table 2.Validated fibrosis tests in NAFLD based upon direct markers of fibrogenesis. Adapted
from Castera et al. [10].
Test n Parameters AUROC Cut-offs Sens Spec PPV NPV
Hyaluronate 112 Hyaluronate 0.80 50 ng/ml 69% 83% 75% 84%
Type IV collagen 7S 0.82 5 ng/ml 81% 71% 68% 78%
Type IV collagen 7S 0.97 89% 77% 100%
Hyaluronate 0.90 42 ng/ml 100% 90% 71% 94%
Hyaluronate 148
TIMP-1, 0.81 0.3576 80% 84% 57% 92%
ELF 192 hyalularonic acid,
terminal peptide of
Hepascore 242 procollagen III 0.37 76%
Hyaluronate,
α2-macroglobulin,
bilirubin, GGT, age,
sex
26 Postgraduate Course Syllabus • Metabolic Liver Disease
