Reappraisal of the roles of alpha-fetoprotein in hepatocellular carcinoma surveillance using large-scale nationwide database and hospital-based information.

Background/purpose: Controversies over the use of alpha-fetoprotein (AFP) for detection of hepatocellular carcinoma (HCC) existed from guidelines. Using large-scale database and hospital-based information, we aimed to reappraise the role of AFP in HCC surveillance, including proportion of AFP elevation by stage of HCC, additional benefit of AFP in combination of ultrasonography (US) in the detection of early HCC, and survival in early HCC with high AFP levels.

Methods: This retrospective study enrolled 43,437 patients from database of the Taiwan Cancer Registry (TCR) and 4250 patients from Kaohsiung Chang Gung Memorial Hospital (KCGMH) between January 2011 and December 2017.

Results: The HCC cases in KCGMH accounted for 9.8% of the total cases in the TCR. Among both nationwide database and hospital-based information, the proportion of early HCC patients with an AFP level of ≥20 ng/mL was approximately 40%. In KCGMH, the proportion of patients with an AFP level of ≥20 ng/mL and a virus-related (hepatitis B and C) etiology was around 41.7%; furthermore, among patients with early HCC, those with an AFP level of ≥20 ng/mL had 4.7 years of median survival and 48.3% of the 5-year overall survival rate. By hospital electronic medical records review of early HCC cohort in KCGMH, approximately 10.9% of patients with AFP levels ≥20 ng/mL had US-undetectable early HCC.

Conclusion: This study suggested that AFP in combination with US would add an additional benefit as being a prompted role for detection of early HCC in patients with US-undetectable HCC.