Background Early diagnosis of hepatocellular carcinoma enhances its effective management. Aim Evaluation of Midkine as a biomarker for diagnosis of hepatocellular carcinoma. Methods: 90 subjects devided into three groups, Group I included 40 HCV patients with liver cirrhosis, Group II included 40 HCV cirrhotic patients with hepatocellular carcinoma and Group III included 10 healthy subjects as a control group. Demographic, laboratory and imaging data were collected. All cirrhotic cases were evaluated by Child-Pugh and MELD scores while BCLC score and Okuda staging were applied for hepatocellular carcinoma cases. Serum Midkine was measured by ELISA technique. Results HCC group had significant elevation in Midkine level compared when to Cirrhotic and Control groups (3.5±2.5 ng/ml versus 1±0.7 ng/ml and 0.1±0.1 ng/ml) (p = 0.000). No significant correlations were found between Midkine and age, sex, site or size of focal lesion, Child classification, MELD score, Okuda staging or BCLC score. ROC analysis showed that the best cut-off value for Midkine was 1.33ng/ml and for AFP was 41.3ng/ml. Area Under the Curve was higher in Midkine than AFP (0.921 and 0.79 respectively) with higher specificity of AFP than Midkine (97.5% and 82.5% respectively) and higher sensitivity of Midkine than AFP (87.5 and 62.5% respectively). By combination of serum AFP and Midkine, AUC was 0.94 with specificity 97.5 % and sensitivity 87.5%
Conclusion Midkine may be a sensitive biomarker for diagnosis of hepatocellular carcinoma and combination between alpha-fetoprotein and Midkine increases the accuracy in diagnosis.