Abstract
Background:
Hepatocellular carcinoma (HCC) is associated with a poor 5-year survival mainly due to detection at late stages. Better non-invasive surveillance methods are needed to improve early detection and maximize survival. We performed a strict assessment of DNA methylation markers (DMMs) for HCC detection.
Methods:
A total of 385 samples from liver tissues and blood were analyzed. Genome-wide Methylated DNA sequencing (MeD-seq) was initially performed on 46 liver tissues, followed by the validation using quantitative methylation-specific PCR (qMSP) on 175 liver tissues. The selected DMMs with and without ASAP/GAAD score were further evaluated in 180 blood samples. Additionally, MeD-seq was performed to validate the results on blood.
Results:
MeD-seq revealed a substantial number of differentially methylated regions (DMRs) in HCC tissues compared to non-HCC controls. By qMSP, the top 5 DMMs demonstrated strong performance in distinguishing cirrhotic HCC from cirrhosis controls in tissue (AUC 0.842 to 0.957). However, evaluation of these DMMs in blood showed lower performance in early HCC detection compared to cirrhosis in both the training (sensitivity 26.7–43.3%, 81.3% specificity) and validation cohorts (sensitivity 16.2–43.2%, 85.7% specificity). The addition of DMMs to the ASAP/GAAD score only provided an additional 5.4% sensitivity in the validation cohort compared to the ASAP/GAAD score alone. These findings were confirmed using MeD-seq analysis in blood samples, which revealed no detectable DMRs between cirrhotic HCC and cirrhosis controls. Interestingly, DNA methylation patterns in blood of healthy individuals differed strongly from both groups (cirrhosis and cirrhotic HCC).
Conclusion:
DNA methylation patterns in liver tissue were distinctly different between HCC and controls. In blood, DMMs contributed minimally to early-stage HCC detection compared to cirrhosis, whether used alone or in combination with the ASAP/GAAD score. It is likely that high baseline DNA methylation related to cirrhosis and possibly the low input of tumor-related DNA impacts the use of DMMs in early HCC detection in blood.
Original language | English |
---|---|
Article number | 144 |
Journal | Journal of Experimental and Clinical Cancer Research |
Volume | 44 |
Issue number | 1 |
DOIs | |
Publication status | Published - 15 May 2025 |
Bibliographical note
Publisher Copyright:© The Author(s) 2025.