Biomarkers in HCC Surveillance: Gastroenterology | May 2026
Introduction and Summary
This commentary discusses the role of serum biomarkers in hepatocellular carcinoma surveillance, particularly after a randomised trial by Hirode et al. evaluated whether adding AFP, AFP-L3, and DCP to routine ultrasound improves early HCC detection.
The original trial found that adding these biomarkers to biannual ultrasound did not significantly improve early-stage HCC detection compared with ultrasound alone. This challenges the routine use of biomarker panels in all high-risk patients.
However, the authors of this letter argue that biomarkers should not be dismissed too quickly. They highlight important methodological issues that may influence interpretation.
Problem Statement
HCC surveillance is difficult because ultrasound has variable sensitivity, especially in obesity, cirrhosis, and nodular liver disease. Biomarkers may help, but their value depends on patient risk, biomarker thresholds, tumour biology, and timing of measurement.
The key question is not simply whether biomarkers should be added to ultrasound for everyone, but which patients may benefit most and how biomarkers should be interpreted longitudinally.
Key Points Raised by the Authors
- Baseline HCC Risk Was Not Clearly Stratified
The authors note that although both trial groups appeared balanced clinically, baseline HCC risk scores were not clearly reported. This matters because lower-risk patients, especially those without cirrhosis, may reduce the apparent benefit of biomarkers.
- Risk-Based Surveillance May Be Better
Instead of a uniform surveillance strategy, the authors suggest using validated HCC risk scores to adjust surveillance intensity, biomarker cutoffs, and cost-effectiveness.
- AFP-L3 May Have Limited Added Value
Although AFP-L3 showed relatively better individual diagnostic accuracy, the GALAD score, which includes AFP-L3, numerically performed slightly worse than the ASAP model, which excludes it. This raises the possibility that AFP-L3 may add limited independent value in some populations.
- AFP-L3 Should Be Tested in Specific Subgroups
The authors suggest evaluating AFP-L3, particularly in patients with negative ultrasound and AFP <20 ng/mL, where its incremental value would be clinically more meaningful.
- Ultrasound False Positives Remain a Major Problem
A high proportion of positive ultrasound findings did not lead to HCC diagnosis. Biomarkers may be useful not only for detecting cancer but also for helping distinguish benign ultrasound abnormalities from true malignancy.
- More Direct Comparison Is Needed
The authors request performance data for ultrasound alone within the biomarker arm, so the true incremental value of biomarkers can be better quantified.
- Longitudinal Biomarker Trends May Be More Informative
Rather than analysing biomarkers only from baseline, aligning biomarker changes to the actual date of HCC diagnosis may reveal rising trends before cancer detection.
Clinical Relevance
This letter provides an important caution: a negative overall trial result does not mean biomarkers have no role in HCC surveillance. Their value may be greatest in selected high-risk patients, ultrasound-negative patients, patients with poor ultrasound visualisation, or those showing rising biomarker trends over time.
For clinicians, the key message is that biomarkers should not replace ultrasound, and routine addition for all patients may not be justified. However, risk-adapted and longitudinal biomarker strategies may still improve future surveillance models.
Conclusion
The commentary supports the trial’s important finding that routine addition of AFP, AFP-L3, and DCP to ultrasound may not significantly improve early HCC detection in a broad surveillance population. However, it argues that further subgroup analysis, risk stratification, ultrasound false-positive assessment, and longitudinal biomarker kinetics are needed.
The future of HCC surveillance may not be “ultrasound versus biomarkers,” but rather personalised surveillance using risk scores, imaging quality, biomarker combinations, and dynamic trends over time.
