Performance of serum α-fetoprotein levels in the diagnosis of hepatocellular carcinoma in patients with a hepatic mass

ObjectivesThe role of serum α-fetoprotein (AFP) measurements in the diagnosis of hepatocellular carcinoma (HCC) remains controversial. Some guidelines have advised against the use of AFP in the diagnosis of HCC. This study was conducted to evaluate the performance of AFP in the diagnosis of HCC, and to identify the optimal cut-off value of serum AFP in the diagnosis of HCC in patients with a hepatic mass.MethodsPatients who presented during the period from May 1997 to March 2003 with hepatic lesions, for whom paired data on serum AFP values at baseline and lesion histology were available, were reviewed. The performance of AFP in the diagnosis of HCC was determined using receiver operating characteristic curve analysis.ResultsData for a total of 805 patients were evaluated. The mean AFP value was 26 900 ng/ml (range: 0–1 965 461 ng/ml). The histological diagnosis was HCC in 557 patients. The optimal AFP cut-off value was 10 ng/ml (for sensitivity of 82.6% and specificity of 70.4%). At a cut-off level of 200 ng/ml, sensitivity, specificity, and positive and negative predictive values were 47.7%, 97.1%, 97.5% and 44.4%, respectively. The diagnostic performance of AFP remains similar in patients with chronic hepatitis B virus infection, despite a lower negative predictive value. Common aetiologies of liver lesions associated with elevated AFP include cholangiocarcinoma and neuroendocrine tumours.ConclusionsIn Asian patients with suspicious liver lesions, the cut-off AFP level of 200 ng/ml is useful to achieve a diagnosis of HCC with high specificity and reasonable sensitivity. The measurement of serum AFP should not be excluded from guidelines for the diagnosis of HCC.