Table 3.
Diagnostic accuracy of AFP, PIVKA-II, GPC-3 alone or in combination for the discrimination between patients with cirrhosis and those with HCC.
| Biomarkers | AUC (95% CI) | Cut-off ** | Se | Sp | +LR | -LR |
|---|---|---|---|---|---|---|
| AFP | 0.737 (0.688–0.783) | >9.7 ng/mL | 72% | 66% | 2.11 | 0.43 |
| PIVKA-II | 0.790 (0.744–0.832) | >73 mAU/mL | 68% | 84% | 4.11 | 0.39 |
| GPC-3 | 0.637 (0.584–0.688) | >73 pg/mL | 73% | 51% | 1.49 | 0.53 |
| AFP + PIVKA-II * | 0.822 (0.777–0.860) | >0.41 | 70% | 84% | 4.23 | 0.36 |
The biomarkers combination was carried out by logistic regression analysis. * The obtained formula of the model is: y = −5.09676 + 1.19356*Log AFP + 1.81397*Log PIVKA-II; the probability (p) of HCC is given by: p = 1/(1 + e−y). ** Identified by Youden J statistic. AUC values were calculated by receiver operating characteristic curve analysis. Abbreviations—alpha-fetoprotein (AFP), area under the curve (AUC), glypican-3 (GPC-3), hepatocellular carcinoma (HCC), protein induced by vitamin K absence or antagonist II (PIVKA-II), sensitivity (Se), specificity (Sp), positive likelihood ratio (+LR), negative likelihood ratio (−LR).