Objectives: Serum biomarkers are valuable for clinical decision-making for patients with hepatocellular carcinoma (HCC), among which the most promising are AFP, AFP-L3, DCP, DKK-1, and GP73; however, the efficacy of using combined biomarkers remains controversial. This meta-analysis provides insights regarding this topic.

Methods: After systematically surveying the literature available in PubMed, Embase, and Cochrane Library, we identified 28 qualified articles published since January 2015. A random-effects model was used to assess pooled sensitivity, specificity, positive and negative likelihood ratios (PLRs and NLPs), and diagnostic odds ratio (DOR).

Results: Values under the summary receiver operating characteristic (SROC) curve varied in different panels of the five biomarkers. Importantly, the sum of sensitivity and specificity of AFP+GP73 was 1.76 (P= 0.0001), which was the best among all the panels. The sum of the triple biomarker panel of AFP, AFP-L3, and DCP was larger (1.64, P= 0.0001) than those of any double biomarker panels of AFP, AFP-L3, and DCP.

Conclusions: To the best of our knowledge, this is the first meta-analysis to focus solely on combination assays of multiple biomarkers in HCC. The combined assay of AFP and GP73 conferred the best outcome among all panels. The triple combined panel of AFP, AFP-L3, and DCP showed higher diagnostic potential than individual random double combinations of the three biomarkers. Multiple-biomarker combined assays will be clinically important for decision-making processes for HCC.