OBJECTIVE Hepatocellular carcinoma (HCC) develops with high incidence in patients with chronic liver disease (CLD), and particularly in those with cirrhosis. Currently, diagnosis and surveillance are mainly based on imaging methods. The aim of this study was to evaluate the diagnostic accuracy of highly sensitive measurement of α-fetoprotein (AFP), Lens culinaris agglutinin-reactive fraction of AFP (AFP-L3) and des-γ-carboxyprothrombin (DCP) alone and in combination, for HCC detection. In addition, a recently proposed statistical model, including these three biomarkers plus sex and age, the GALAD model, was applied. METHODS In a total of 98 patients (44 CLD patients without HCC [23 men, 21 women; mean age, 53.2 ± 13.4 years] and 54 patients with HCC [45 men, nine women; 69.5 ± 9.8 years]), AFP, AFP-L3 and DCP levels were determined using a highly sensitive assay on an μTASWako i30 immuno-analyzer. Areas under the curve (AUC), sensitivity and specificity were calculated and compared to assess diagnostic performance of the HCC biomarkers and of the GALAD model. RESULTS AFP, AFP-L3 and DCP serum levels were significantly elevated in HCC compared with CLD patients (P < 0.0001). AUC values were 0.891, 0.867 and 0.870, respectively. The combination of the three biomarkers resulted in an AUC of 0.947, whereas the GALAD model showed an AUC of 0.976 with a difference between AUC values of 0.029 (P = 0.028). CONCLUSIONS The combination of AFP, AFP-L3 and DCP is superior to a single biomarker in HCC detection. Furthermore, GALAD model performance is significantly higher than simple combination of these three biomarkers.