Abstract
Background
DNA methylation is under investigation as an early diagnostic biomarker for cancers such as hepatocellular carcinoma (HCC). p21-activated protein kinase 1 (PAK1) demonstrates a high methylation tendency in HCC. We assessed the diagnostic and prognostic performance of PAK1 methylation in liver transplantation (LT) recipients with and without HCC.
Methods
To assess PAK1 methylation, stored pretransplant blood samples from LT recipients were analyzed by droplet digital polymerase chain reaction.
Results
This study included 274 patients with HCC and 100 control patients without the disease. Ten-year survival rates in the HCC and control groups were 60.5% and 80.6%, respectively (P=0.001). The 10-year HCC recurrence rate was 39.0%. Ten-year survival rates in the HCC recurrence and nonrecurrence groups were 13.3% and 91.2%, respectively (P<0.001); median PAK1 methylation levels in the control and HCC groups were 50.0 and 108.0 copies (P=0.102). The area under the receiver operating characteristic curve was 0.599, and the Youden J index was 0.199, indicating 57.9% sensitivity and 62.0% specificity at a cutoff of 78 copies. With a cutoff of 5 copies, sensitivity was 94.1% and specificity was 14.0%. The positive likelihood ratio was 1.09 and the negative likelihood ratio was 0.42, indicating diagnostic accuracy below α-fetoprotein and protein induced by vitamin K absence or antagonist-II. PAK1 cutoffs of 5 and 10 copies did not affect HCC recurrence, but a cutoff of 2 copies appeared associated with lower recurrence.
Conclusions
These data suggest that PAK1 methylation is not a clinically useful biomarker for HCC in LT recipients. New DNA methylation biomarkers are required for early diagnosis.
Keywords: DNA methylation, Hepatocellular carcinoma, Early diagnosis, Liver transplantation, Prognosis
| HIGHLIGHTS |
|---|
|
INTRODUCTION
Hepatocellular carcinoma (HCC) is the cancer with the sixth highest incidence and the third highest mortality rate worldwide [1]. In Korea, HCC represents the sixth most common cancer [2]. Reflecting this, Korean patients at high risk of HCC are offered early screening based on ultrasonography and serum α-fetoprotein (AFP) tests. Nevertheless, these conventional methods have relatively low sensitivity—up to 63%—for early diagnosis of HCC, indicating that more sensitive methods are needed [2,3].
To address this, DNA methylation is under investigation as a biomarker for cancer diagnosis. DNA methylation is an epigenetic change in which a methyl group (-CH3) is added to cytosine (C) residues within a CpG sequence of DNA, thereby regulating transcription initiated via the gene promoter. Methylation can promote cancer development by inhibiting the expression of mRNA encoding tumor suppressor genes in cancer cells; as such, methylation is attracting attention as a new biomarker for cancer diagnosis. DNA methylation biomarkers can be detected early because cancer-specific methylation characteristics are frequently maintained throughout the course of cancer development, as well as due to the structural stability of DNA. Methylation can be identified on circulating tumor DNA (ctDNA) released from cancer tissues into the blood via cell death mechanisms, potentially enabling blood-based cancer diagnosis [4].
In this study, expression of p21-activated protein kinase 1 (PAK1), which exhibits a particularly high tendency for methylation in HCC, was selected for assessment as a new biomarker. The aim was to determine whether the detection of DNA methylation at a specific site on PAK1 could be used to diagnose HCC. Droplet digital polymerase chain reaction (ddPCR) was performed using ctDNA extracted from plasma samples, and these data were used to evaluate pretransplant diagnostic accuracy for HCC and the risk of HCC recurrence after liver transplantation (LT).
METHODS
Prior to LT, all participants provided consent for collection of blood samples and clinical data. The study protocol was approved by the Institutional Review Board of Asan Medical Center (IRB No. 2014-0898).
Study Design
This study used blood samples collected from patients before LT to analyze the diagnostic accuracy of evaluating the degree of PAK1 methylation for HCC. Additionally, it involved a retrospective clinical study designed to analyze the effect of methylation on posttransplant prognosis after HCC recurrence. The study was divided into two parts. First, an HCC diagnostic study was conducted using PAK1 methylation. The level of methylation before LT was measured in patients with pathologically confirmed HCC at the liver explant, a cutoff value was set, and the possibility of diagnosing HCC was evaluated. Second, a prediction study of HCC recurrence was conducted based on PAK1 methylation. The risk of posttransplant HCC recurrence was assessed according to the pretransplant PAK1 methylation expression level.
Selection of Hepatocellular Carcinoma Study Patients
Among adult recipients aged ≥18 years who underwent LT at Asan Medical Center in Seoul between January 2010 and December 2020, those who survived for more than 1 month after transplantation and had available frozen pretransplant blood samples were selected as study participants. The exclusion criteria were other liver carcinomas, such as intrahepatic bile duct cancer, or pathological complete necrosis of HCC caused by pretransplant treatment. Ultimately, the study HCC cohort comprised 273 patients.
Selection of Control Participants for Hepatocellular Carcinoma Diagnosis
Among adult recipients aged ≥18 years who underwent LT at Asan Medical Center from January 2010 to December 2020, those who were found to have no cancer (including HCC), who survived more than 1 month after transplantation, and for whom frozen pretransplant blood samples were available were selected as controls. Based on these criteria, 100 patients were randomly selected as the non-HCC control group.
Patient Follow-Up
The study group and control group were followed to monitor survival and HCC recurrence by reviewing medical records at the time of death or until December 2024.
Statistical Analysis
All statistical analyses were performed using SPSS ver. 22 (IBM Corp.) and MedCalc ver. 23.2.1 (MedCalc Software Ltd). Continuous variables were expressed as mean±standard deviation or as median (range), along with 95% confidence intervals (CIs). P-values less than 0.05 were considered to indicate statistical significance.
Sensitivity, defined as the proportion of HCC samples judged to be positive by the PAK1 methylation test, represented the primary evaluation variable used to assess the accuracy of the test. Specificity, denoting the proportion of non-HCC samples judged to be negative, was also employed to evaluate the diagnostic accuracy for non-HCC control samples. The positive likelihood ratio was defined as “sensitivity / (1 − specificity)” and the negative likelihood ratio as “(1 − sensitivity) / specificity”. The receiver operating characteristic (ROC) curve displays changes in the measured false positive and true positive rates as the model’s judgment criteria were continuously changed. The area under the ROC curve (AUC) was then determined. The Youden J index was calculated as “sensitivity + specificity – 1”, with the optimal cutoff found at the maximum value. Kaplan-Meier survival statistics were used to assess HCC recurrence and patient survival and were compared using the log-rank test.
Overview of ddPCR for PAK1 Methylation
DNA isolated from pretransplant patient plasma was sequenced following bisulfite treatment, and specific site methylation of the PAK1 gene was quantitatively examined by ddPCR. Methylation of specific sites within the PAK1 gene in plasma samples was analyzed qualitatively by ddPCR using the Droplet Digital PCR System (Bio-Rad). The assay comprised three major steps. (1) DNA extraction from patient plasma: after 5 mL of blood was collected, 2 mL of plasma was separated, and DNA was extracted using a circulating cell-free DNA (cfDNA) extraction kit. (2) Bisulfite conversion: the extracted DNA was subjected to bisulfite treatment for methylation analysis. Unmethylated cytosine is converted by bisulfite treatment, whereas methylated cytosine remains unchanged, enabling classification of the methylated and non-methylated status of the target sequence based on DNA base sequence differences. (3) ddPCR analysis: bisulfite-treated DNA was subjected to ddPCR to measure methylation at specific sites of the PAK1 gene.
ddPCR Procedures for PAK1 Methylation
Preparation and storage of samples
From each participant, 5 mL of blood was collected in an anticoagulant tube. After collection, the blood was allowed to stand for 4 hours to separate the plasma (hemolyzed specimens were not used). The minimum amount of plasma required for the test was 2 mL. The collected blood was transferred to a 15-mL tube and centrifuged at 2,000 × g for 15 minutes at 4 °C. After centrifugation, plasma was collected, with care taken not to touch the buffy coat to prevent white blood cell contamination. The buffy coat was removed completely by further centrifugation. Plasma from the centrifuged tube was transferred to a cryovial tube. The lid was checked to ensure it was properly closed, and sample information was recorded prior to storage at −70 °C or lower.
Extraction of cfDNA
DNA was extracted from 2 mL of plasma using a cell-free DNA extraction kit (Cobas cfDNA Sample Preparation Kit; Roche Korea). The extracted DNA was eluted to a final volume of 50 μL. For quality control, DNA was extracted using 10 μL of positive control (PC, provided) mixed with 2 mL of 1X phosphate-buffered saline to yield a PC substance. Nonmixed phosphate-buffered saline was used as a negative control substance (no template control [NTC]).
Bisulfite conversion
A bisulfite conversion kit (EZ DNA Methylation-Lightning Kit; Zymo Research) was used according to the manufacturer’s manual. Briefly, 50 μL of extracted DNA was used as the input and eluted to a final volume of 11 μL. For quality control, the DNA-extracted NTC and PC substances were also converted to bisulfite.
Digital polymerase chain reaction
The Droplet Digital PCR System was used in accordance with the manufacturer’s manual.
RESULTS
Clinicopathological Characteristics of the HCC Study and Control Groups
The clinicopathological characteristics of the 273 patients in the HCC study group and the 100 patients in the non-HCC control group are summarized in Table 1. When comparing the HCC recurrence (n=105) and nonrecurrence (n=168) groups, no significant differences were found in age, sex, protein induced by vitamin K absence or antagonist-II (PIVKA-II) level, model for end-stage liver disease (MELD) score, ABO blood group incompatibility, graft-to-recipient weight ratio, or graft type. In the HCC recurrence group, the proportions of patients with hepatitis B virus infection and those meeting the Milan criteria were significantly lower, whereas AFP level, maximum tumor size, tumor number, total tumor volume, and AFP–des-γ-carboxy-prothrombin (PIVKA-II)–tumor volume (ADV) score were significantly higher.
Table 1.
Clinicopathological features of the 273 study patients and 100 control patients
| Parameter | All HCC cases (A and B) | HCC recurrence group (A) | HCC nonrecurrence group (B) | Control non-HCC group | P-value (A versus B) |
|---|---|---|---|---|---|
| Case number | 273 | 105 | 168 | 100 | - |
| Age (yr) | 53.7±6.7 | 52.8±6.1 | 54.2±6.1 | 52.7±6.9 | 0.105 |
| Sex | 0.21 | ||||
| Male | 203 (74.4) | 90 (85.7) | 123 (73.2) | 76 (76.0) | |
| Female | 70 (25.6) | 25 (23.8) | 45 (26.8) | 24 (24.0) | |
| Background liver disease | 0.001 | ||||
| Hepatitis B virus infection | 204 (74.7) | 90 (85.7) | 114 (67.9) | 50 (50.0) | |
| C virus infection | 36 (13.2) | 6 (5.7) | 30 (17.) | 15 (15.0) | |
| Alcoholic liver disease | 17 (6.2) | 5 (4.8) | 12 (7.1) | 15 (15.0) | |
| Others | 16 (5.9) | 4 (3.8) | 12 (7.1) | 20 (20.0) | |
| MELD score | 12.0±6.4 | 11.6±6.1 | 12.3±6.6 | 14.1±6.2 | 0.41 |
| AFP (ng/mL) | 26.0 (1.0–42,200) | 34.3 (1.0–42,200) | 20.9 (1.2–5,097) | 6.6 (0.7–414) | 0.017 |
| PIVKA-II (mAU/mL) | 37 (9–20,000) | 56 (9–20,000) | 32 (9–20,000) | 21 (5–2,900) | 0.165 |
| Explant pathology | |||||
| Tumor number | 1 (1–25) | 2 (1–25) | 1 (1–12) | NA | <0.001 |
| Maximal tumor diameter (cm) | 2.7 (0.4–11.0) | 3.5 (0.7–11.0) | 2.3 (0.4–11.0) | NA | <0.001 |
| Total tumor volume (mL) | 28.7 (0.1–1,176) | 61.1 (0.7–1,176) | 16.7 (0.1–1,011) | NA | <0.001 |
| Macrovascular invasion | 12 (4.4) | 8 (7.6) | 4 (2.4) | NA | 0.08 |
| Microvascular invasion | 90 (33.0) | 62 (59.0) | 28 (16.7) | NA | <0.001 |
| Explant ADV score | 4.7±1.8 | 5.5±1.9 | 4.2±1.5 | NA | <0.001 |
| Within Milan criteria | 156 (57.1) | 38 (36.2) | 118 (70.2) | NA | <0.001 |
| ABO-incompatible transplantation | 57 (20.9) | 27 (25.7) | 30 (17.9) | 19 (19.0) | 0.161 |
| Graft-recipient weight ratio | 1.12±0.23 | 1.10±0.21 | 1.14±0.25 | 1.11±0.24 | 0.21 |
| Graft type | 0.23 | ||||
| Right liver graft | 252 (92.3) | 100 (95.2) | 152 (90.5) | 93 (93.0) | |
| Left liver graft | 6 (2.2) | 2 (1.9) | 4 (2.4) | 3 (3.0) | |
| Dual grafts | 15 (5.5) | 3 (2.9) | 12 (7.1) | 4 (4.0) |
Data are presented as mean±standard deviation, number (%), or median (range).
HCC, hepatocellular carcinoma; MELD, model for end-stage liver disease; AFP, α-fetoprotein; PIVKA-II, protein induced by vitamin K absence or antagonist-II; ADV, AFP–des-γ-carboxy-prothrombin (PIVKA-II)–tumor volume; NA, not available.
Survival Analysis of the HCC Study and Control Groups
The 1-year, 3-year, 5-year, and 10-year survival rates in the HCC study group were 91.9%, 75.5%, 68.8%, and 60.5%, respectively, whereas those in the control group were 95.0%, 91.9%, 87.6%, and 80.6%; these differences were statistically significant (P=0.001) (Fig. 1).
Fig. 1.
Comparison of overall survival curves for the hepatocellular carcinoma (HCC) study group and the non-HCC control group.
HCC Recurrence and Survival Analysis: HCC Study Group
The 1-, 3-, 5-, and 10-year HCC recurrence rates in the HCC study group were 22.0%, 35.1%, 37.8%, and 39.0%, respectively (Fig. 2). The 1-, 3-, 5-, and 10-year patient survival rates in the HCC recurrence group were 85.7%, 45.7%, 28.6%, and 13.3%, respectively, while those in the non-HCC recurrence group were 95.8%, 94.0%, 94.0%, and 91.2% (P<0.001) (Fig. 3).
Fig. 2.
Cumulative recurrence curves for hepatocellular carcinoma (HCC) in the HCC study group.
Fig. 3.
Comparison of overall survival curves for patients with and without recurrence in the hepatocellular carcinoma (HCC) study group.
PAK1 Methylation Expression Patterns and Diagnosis of HCC
The patterns of PAK1 methylation expression in the HCC study and control groups are shown in Fig. 4. The median copy number for PAK1 methylation in the control group was 50.0 (range, 1.2–2,940), with an average of 196.2±388.5. In comparison, the median copy number in the HCC group was 108.0 (range, 0–10,360), with an average of 342.5±860.7 (P=0.102). The ROC AUC was 0.599 (95% CI, 0.548–0.650; P=0.003), and the Youden J index was 0.199, indicating 57.9% sensitivity and 62.0% specificity at a cutoff value of 78 copies (Fig. 5).
Fig. 4.
Comparison of PAK1 methylation expression values between the hepatocellular carcinoma (HCC) study group and the non-HCC control group. CI, confidence interval.
Fig. 5.
Receiver operating characteristic (ROC) curve analysis of PAK1 methylation. AUC, area under the ROC curve.
When a cutoff of 5 copies was used—a value proposed by precedent pilot studies—the sensitivity was 94.1% (95% CI, 90.7%–96.6%) and the specificity was 14.0% (95% CI, 7.9%–22.4%). The positive likelihood ratio was 1.09 (95% CI, 1.01–1.19), while the negative likelihood ratio was 0.42 (95% CI, 0.21–0.83) (Table 2). Such low sensitivity and specificity precluded the calculation of a PAK1 methylation cutoff with clinical significance.
Table 2.
Comparison of receiver operating characteristic (ROC) curve analysis results for PAK1 methylation, AFP, and PIVKA-II
| Marker | ROC AUC | Youden index | At conventional cutoff | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AUC | 95% CI | P-value | Youden index J | Cutoff | Sensitivity (%) | Specificity (%) | Cutoff value | Sensitivity (%) | Specificity (%) | pLR | nLR | |
| PAK1 | 0.599 | 0.548–0.650 | 0.003 | 0.199 | 78 copies | 57.9 | 62.0 | 5 copies | 94.1 | 14 | 1.09 | 0.42 |
| AFP | 0.698 | 0.649–0.744 | <0.001 | 0.338 | 23.6 ng/mL | 49.8 | 84.0 | 7.5 ng/mL | 73.9 | 56 | 1.68 | 0.46 |
| PIVKA-II | 0.659 | 0.608–0.707 | <0.001 | 0.243 | 21 mAU/mL | 73.3 | 51.0 | 40 mAU/mL | 48.4 | 69 | 1.56 | 0.75 |
AFP, α-fetoprotein; PIVKA-II, protein induced by vitamin K absence or antagonist-II; AUC, area under the ROC curve; 95% CI, 95% confidence interval; pLR, positive likelihood ratio; nLR, negative likelihood ratio.
The distributions of AFP and PIVKA-II values are presented in Fig. 6. Table 2 compares the sensitivity and specificity according to the PAK1, AFP, and PIVKA-II test cutoff values. The results of the ROC curve analysis are shown in Fig. 7.
Fig. 6.
Comparison of (A) AFP and (B) PIVKA-II values between the hepatocellular carcinoma (HCC) study group and the non-HCC control group. AFP, α-fetoprotein; CI, confidence interval; PIVKA-II, protein induced by vitamin K absence or antagonist-II.
Fig. 7.
Comparison of receiver operating characteristic curves for PAK1 methylation, AFP, and PIVKA-II. AFP, α-fetoprotein; PIVKA-II, protein induced by vitamin K absence or antagonist-II.
Use of PAK1 Methylation Levels to Predict HCC Recurrence
When using a PAK1 methylation cutoff of 5 copies, no difference was found in the HCC recurrence rate (P=0.478) (Fig. 8). This remained the case even when a PAK1 methylation cutoff of 10 copies was applied (P=0.783) (Fig. 8).
Fig. 8.
Comparison of cumulative hepatocellular carcinoma (HCC) recurrence curves using a PAK1 methylation cutoff of (A) 5 copies or (B) 10 copies.
When applying a PAK1 methylation cutoff of 2 copies, a noticeable difference was observed in HCC recurrence rates, but it did not reach statistical significance due to the small sample size (n=7; P=0.142) (Fig. 9). Furthermore, no significant difference was noted in patient survival rates (P=0.595) (Fig. 9).
Fig. 9.
(A) Comparison of cumulative hepatocellular carcinoma (HCC) recurrence and (B) overall survival curves using a PAK1 methylation cutoff of 2 copies.
Posttransplant HCC Recurrence According to ADV Score
ADV score was associated with a proportional increase in HCC recurrence after LT (P<0.001) (Fig. 10), as well as a proportional decrease in the overall patient survival rate (P<0.001) (Fig. 10).
Fig. 10.
(A) Comparison of cumulative hepatocellular carcinoma (HCC) recurrence and (B) overall survival curves according to ADV score. ADV, α-fetoprotein–des-γ-carboxy-prothrombin–tumor volume.
DISCUSSION
DNA methylation is being evaluated as an optimized biomarker for early cancer diagnosis using liquid biopsy. CpG methylation, the process of attaching a methyl group to a cytosine residue within a CpG sequence, is an early regulatory event in gene expression and occurs continuously regardless of cancer stage. The structural stability of DNA enables straightforward detection by liquid biopsy; therefore, detection of specific methylated genes may facilitate early cancer diagnosis.
Methylation of PAK1, which is specific for HCC, was chosen for this study. First, a bioinformatics-based analysis was performed, and a candidate group of biomarkers was selected from a total of three microarray high-throughput datasets. A meta-analysis of DNA methylation biomarkers for HCC published in 2016 reported that methylation of p16, RASSF1A, GSTP1, p14, CDH1, APC, RUNX3, SOCS1, p15, MGMT, SFRP1, WIF1, PRDM2, DAPK1, RARβ, hMLH1, p73, DLC1, p53, SPINT2, OPCM, and WT1 correlated with HCC; however, PAK1 was not included [5].
PAK1 encodes p21-activated protein kinase 1, a member of the serine/threonine p21-activating kinase family. These proteins are important effectors that link Rho GTPase to cytoskeletal reconstruction and nuclear signaling and act as targets for the small GTP-binding proteins Cdc42 and Rac. This specific family member regulates cell motility and morphology. Overexpression of this gene has been reported in many cancer types, including HCC [6,7].
A study of HCC cell lines and a mouse xenograft model reported that PAK1 and Snail (a master regulator of epithelial-mesenchymal transition [EMT]) were upregulated in HCC cells; in contrast, PAK1 knockdown inhibited HCC cell proliferation, migration, and invasion, and increased cell death. PAK1 knockdown also inhibited tumor growth in vivo. Thus, PAK1 appears to promote EMT by increasing expression of Snail and by activating the β-catenin pathway, stimulating HCC progression [8].
The pilot studies conducted prior to this study included a PAK1 methylation validation test using plasma specimens, with diagnostic sensitivity and specificity for HCC confirmed by comparative analysis of HCC (up to stage II–IV) and normal specimens. In the 33 normal plasma specimens, PAK1 methylation was detected at 0–32 copies. By contrast, in the 67 plasma specimens from patients who underwent liver resection for HCC, PAK1 methylation was detected at a relatively very high level of 0–8,192 copies. The ROC AUC for HCC diagnosis was very high at 0.94, and the cutoff for HCC diagnosis was set at 5 copies. The present study was conducted based on these unpublished prior research data.
The results of the present study, conducted using samples from adult LT recipients, differed markedly from those of the previous studies. At a cutoff of 5 copies, the present study displayed a sensitivity of 94.1%, a specificity of 14.0%, a positive likelihood ratio of 1.09, and a negative likelihood ratio of 0.42. It was nearly impossible to calculate a clinically significant cutoff for PAK1 methylation because of such a low specificity. The ROC AUC for diagnosing HCC was low (0.599), and the Youden J index was 0.199, displaying a sensitivity of 57.9% and a specificity of 62.0% at a cutoff of 78 copies. With such low sensitivity and specificity, it was also impossible to derive a clinically significant cutoff value for PAK1 methylation. This is thought to largely stem from the different underlying liver diseases in patients with HCC. In the preceding studies, patients with HCC had chronic viral hepatitis or early cirrhosis at the Child-Pugh A level, whereas all participants in the present study had advanced cirrhosis at the Child-Pugh B or C level. Therefore, PAK1 methylation is considered to have limited clinical utility for diagnosing HCC. In this study, the ROC AUC for PAK1 methylation was lower than those for AFP and PIVKA-II.
In patients with advanced cirrhosis who are candidates for LT, the diagnostic utility of AFP and PIVKA-II—currently used in clinical practice—is lower than in the general patient population. In a previous single-center study, we divided 2,074 adult LT recipients into two groups, HCC (n=970 [46.8%]) and non-HCC (n=1,104 [53.2%]), over a 10-year study period. The MELD score distribution was <10 (n=464), ≥10 and <15 (n=632), ≥15 and <20 (n=355), ≥20 and <30 (n=340), and ≥30 (n=283). The median AFP and PIVKA-II levels before LT were 11.3 ng/mL and 26 mAU/mL, respectively, in the HCC group and 4.2 ng/mL and 22 mAU/mL in the non-HCC group. ROC curve analysis showed that the ROC AUC for AFP was 0.693, with a sensitivity of 64.5% and a specificity of 64.5% at a cutoff point of 6.8 ng/mL. The ROC AUC of AFP was inversely proportional to the MELD score. The ROC AUC of PIVKA-II was 0.546, with a sensitivity of 53.1% and a specificity of 51.8% at a cutoff of 25 mAU/mL. The ROC AUC of PIVKA-II, excluding patients with a MELD score ≥30, was <0.6. Therefore, the sensitivity of AFP and PIVKA-II was not high in the group of patients with a high MELD score [9].
In a previous multicenter study, we categorized 3,273 adult LT recipients according to the presence or absence of HCC and underlying liver disease. The median AFP and PIVKA-II levels were 6.3 ng/mL and 29 mAU/mL in the HCC group and 3.3 ng/mL and 35 mAU/mL in the non-HCC group (P<0.001 for AFP; P=0.037 for PIVKA-II). In patients with hepatitis B virus infection, the median levels were 6.0 ng/mL and 26 mAU/mL in the HCC group and 3.2 ng/mL and 21 mAU/mL in the non-HCC group (P<0.001 for both). Regarding hepatitis C virus infection, the values were 10.7 ng/mL and 37 mAU/mL in the HCC group and 2.6 ng/mL and 21 mAU/mL in the non-HCC group (P<0.001 and P=0.117, respectively). For alcoholic liver disease, the levels were 5.2 ng/mL and 61 mAU/mL in the HCC group and 6.4 ng/mL and 75 mAU/mL in the non-HCC group (P<0.001 and P=0.419, respectively). Finally, among patients with other liver diseases, the levels were 7.1 ng/mL and 32 mAU/mL in the HCC group and 3.3 ng/mL and 28 mAU/mL in the non-HCC group (P<0.001 and P=0.822, respectively). Therefore, because serum AFP and PIVKA-II are expressed highly variably in patients with end-stage liver disease before transplantation, the role of these tumor markers in the diagnosis of HCC is limited [10].
To date, the U.S. Food and Drug Administration has not approved a kit for the early diagnosis of HCC. The biotechnology company Epigenomics AG has developed an HCC diagnostic kit (HCCBloodTest), with results reported in a prospective clinical study from the United States. DNA methylation markers, including HCCBloodTest (Epigenomics) and a DNA methylation panel constructed through next-generation sequencing (NGS), were evaluated using a training/test design. The NGS panel algorithm included a training group (41 patients with HCC and 46 non-HCC controls with cirrhosis). Plasma samples were collected from patients with cirrhosis (Child-Pugh class A or B), with or without early-stage HCC (Barcelona Clinic Liver Cancer stage 0, A, or B). Subsequently, a blinded sample set was used for testing and analysis using a preset algorithm. The HCCBloodTest and NGS panel showed sensitivities of 76.7% and 57% at specificities of 64.1% and 97%, respectively. A post hoc analysis combining the NGS panel with AFP of 20 ng/mL achieved a sensitivity of 68% at a specificity of 97% (ROC AUC, 0.9). This suggests that methylation biomarkers of cell-free plasma DNA represent a novel alternative for surveillance of HCC [11].
Research on DNA methylation biomarkers has been underway since 2021. A Chinese study investigated 97 patients with HCC, 80 healthy controls, and 46 patients with chronic hepatitis B or C virus infection. The cfDNA level in the HCC group was significantly higher than that in the healthy controls. The sensitivity and specificity were 78.6% and 89.4%, respectively, based on a cutoff value of 15.7% for the cfDNA methylation ratio in the HCC group. The diagnostic accuracy was 85.3%, the positive predictive value was 81.9%, and the negative predictive value was 87.2%. The positive likelihood ratio for the 15.7% cutoff in the diagnosis of HCC was 7.40, and the negative likelihood ratio was 0.24. Therefore, the methylation-based analysis method suggests that it can be used as a liquid biopsy test for the diagnosis of HCC [12].
A Korean multi-institutional study conducted in 2023 identified HCC-specific methylation markers and developed a PCR-based method to detect them in cfDNA. Analysis of large-scale methylome data revealed that ring finger protein 135 (RNF135) and lactate dehydrogenase B (LDHB) represent universally applicable HCC methylation markers, with methylation not detected in other tissue types. These markers were used to develop methylation-sensitive high-resolution analysis (MS-HRM), and the diagnostic accuracy was tested using cfDNA from healthy individuals, at-risk patients, and patients with HCC. Combined analysis of MS-HRM RNF135 and LDHB detected 57% of HCC cases, outperforming the sensitivity of AFP (45%). When used with the AFP test, methylation analysis detected 70% of HCC cases. These results suggest that cfDNA methylation analysis can be used for HCC liquid biopsy [13].
In addition to the diagnostic role of PAK1 in HCC, we performed a prognostic analysis of HCC recurrence after LT. Typical prognostic factors for HCC recurrence include the size and number of HCC tumors, expression of blood biomarkers such as AFP and PIVKA-II, and 18F-fluorodeoxyglucose positron emission tomography findings. Previously, we reported the ADV score and the “tumor size and number, AFP, PIVKA-II, and positron emission tomography (SNAPP)” score, which combine these risk factors [14–25].
Herein, we found that when PAK1 methylation was very low (≤2 copies), the recurrence rate was also quite low. This suggests that low PAK1 expression may be associated with less aggressive tumor characteristics due to reduced EMT. In vitro, PAK1 promotes progression of HCC by upregulating Snail and promoting EMT, while PAK1 knockdown increases apoptosis of HCC and inhibits tumor growth in vivo [8].
In the era of personalized medicine and targeted cancer therapy, ddPCR plays a key role as an ultrasensitive method for detecting tumor genotype. As less invasive strategies for diagnosis, prognosis, and disease monitoring are sought, the number of studies assessing liquid biopsy approaches using ddPCR has increased significantly. ddPCR is a reliable and accurate tool for detecting nucleic acid-based markers derived from cfDNA, cell-free RNA, circulating tumor cells, extracellular vesicles, or exosomes when isolated from whole blood, plasma, and serum. As such, it may help to predict tumor recurrence or elucidate intratumoral heterogeneity and clonal evolution in response to therapy [26].
Early detection of cancer through blood tumor DNA analysis can significantly impact the morbidity and mortality of patients with cancer; however, the use of blood tumor DNA for early cancer diagnosis is challenging due to the small amounts released into the circulation and its dilution by DNA derived from nontumor cells. Recent development of new technologies such as ddPCR and NGS has greatly improved the sensitivity, specificity, and precision for rare sequence detection [27].
This study had several limitations. First, it had no group of healthy control participants with normal livers. This is a practical limitation because it is impossible to collect appropriate control blood samples unless a prospective study is conducted to obtain plasma from healthy individuals. Second, the plasma used in the study was stored at ultra-low temperatures for 2–12 years, so the possibility of DNA denaturation during storage cannot be entirely ruled out. Third, we cannot completely rule out the possibility of hemolysis in some samples during plasma separation. These limitations can be overcome only through a well-designed prospective study.
This study suggests that PAK1 methylation is not a clinically useful biomarker for HCC in patients with end-stage liver disease awaiting LT. However, very low PAK1 methylation may be associated with low posttransplant recurrence of HCC. Therefore, it remains necessary to identify new DNA methylation biomarkers that enable early diagnosis of HCC.
ARTICLE INFORMATION
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Funding/Support
This study was supported by research grant from the Korean Society for Transplantation (2026-00-03001-001).
Author Contributions
Conceptualization: SH. Data curation: SH, HJK. Formal analysis: SH. Investigation: ET, KJL, YKK. Methodology: ET, KJL, YKK. Project administration: SH. Supervision: SH; Visualization: SH, HJK. Writing–original draft: SH, HJK. Writing–review & editing: all authors. All authors read and approved the final manuscript.
References
- 1.Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2021;71:209-49. 10.3322/caac.21660 [DOI] [PubMed]
- 2.Park EH, Jung KW, Park NJ, Kang MJ, Yun EH, Kim HJ, et al. Cancer statistics in Korea: incidence, mortality, survival, and prevalence in 2021. Cancer Res Treat 2024;56:357-71. 10.4143/crt.2024.253 [DOI] [PMC free article] [PubMed]
- 3.Tzartzeva K, Singal AG. Testing for AFP in combination with ultrasound improves early liver cancer detection. Expert Rev Gastroenterol Hepatol 2018;12:947-9. 10.1080/17474124.2018.1512855 [DOI] [PubMed]
- 4.Laird PW. The power and the promise of DNA methylation markers. Nat Rev Cancer 2003;3:253-66. 10.1038/nrc1045 [DOI] [PubMed]
- 5.Zhang C, Li J, Huang T, Duan S, Dai D, Jiang D, et al. Meta-analysis of DNA methylation biomarkers in hepatocellular carcinoma. Oncotarget 2016;7:81255-67. 10.18632/oncotarget.13221 [DOI] [PMC free article] [PubMed]
- 6.Xu MH, Zheng YM, Liang BG, Xu WX, Cao J, Wang P, et al. Deubiquitination of CIB1 by USP14 promotes Lenvatinib resistance via the PAK1-ERK1/2 axis in hepatocellular carcinoma. Int J Biol Sci 2024;20:3269-84. 10.7150/ijbs.96031 [DOI] [PMC free article] [PubMed]
- 7.Tu J, Zhao Z, Xu M, Chen M, Weng Q, Ji J. LINC00460 promotes hepatocellular carcinoma development through sponging miR-485-5p to up-regulate PAK1. Biomed Pharmacother 2019;118:109213. 10.1016/j.biopha.2019.109213 [DOI] [PubMed]
- 8.Cao F, Yin LX. PAK1 promotes proliferation, migration and invasion of hepatocellular carcinoma by facilitating EMT via directly up-regulating Snail. Genomics 2020;112:694-702. 10.1016/j.ygeno.2019.05.002 [DOI] [PubMed]
- 9.Kim Y, Park YH, Hwang S, Kim KH, Ahn CS, Moon DB, et al. Diagnostic role of blood tumor markers in predicting hepatocellular carcinoma in liver cirrhosis patients undergoing liver transplantation. Ann Transplant 2016;21:660-7. 10.12659/AOT.900552 [DOI] [PubMed]
- 10.Kang WH, Hwang S, Kim JM, Lee KW, Joo DJ, You YK, et al. Diagnostic role of tumor markers for hepatocellular carcinoma in liver transplantation candidates: an analysis using the Korean Organ Transplantation Registry database. Ann Transplant 2022;27:e936937. 10.12659/AOT.936937 [DOI] [PMC free article] [PubMed]
- 11.Lewin J, Kottwitz D, Aoyama J, deVos T, Garces J, Hasinger O, et al. Plasma cell free DNA methylation markers for hepatocellular carcinoma surveillance in patients with cirrhosis: a case control study. BMC Gastroenterol 2021;21:136. 10.1186/s12876-021-01714-8 [DOI] [PMC free article] [PubMed]
- 12.Wang J, Yang L, Diao Y, Liu J, Li J, Li R, et al. Circulating tumour DNA methylation in hepatocellular carcinoma diagnosis using digital droplet PCR. J Int Med Res 2021;49:300060521992962. 10.1177/0300060521992962 [DOI] [PMC free article] [PubMed]
- 13.Kim SC, Kim DW, Cho EJ, Lee JY, Kim J, Kwon C, et al. A circulating cell-free DNA methylation signature for the detection of hepatocellular carcinoma. Mol Cancer 2023;22:164. 10.1186/s12943-023-01872-1 [DOI] [PMC free article] [PubMed]
- 14.Hwang S, Song GW, Ahn CS, Kim KH, Moon DB, Ha TY, et al. Quantitative prognostic prediction using ADV score for hepatocellular carcinoma following living donor liver transplantation. J Gastrointest Surg 2021;25:2503-15. 10.1007/s11605-021-04939-w [DOI] [PubMed]
- 15.Park GC, Hwang S, You YK, Choi Y, Kim JM, Joo DJ, et al. Quantitative prediction of posttransplant hepatocellular carcinoma prognosis using ADV score: validation with Korea-Nationwide Transplantation Registry Database. J Gastrointest Surg 2023;27:1353-66. 10.1007/s11605-023-05670-4 [DOI] [PubMed]
- 16.Hwang S, Moon DB, Kim KH, Ahn CS, Song GW, Jung DH, et al. Prognostic accuracy of the ADV score following resection of hepatocellular carcinoma with portal vein tumor thrombosis. J Gastrointest Surg 2021;25:1745-59. 10.1007/s11605-020-04800-6 [DOI] [PubMed]
- 17.Kang WH, Hwang S, Kaibori M, Kim JM, Kim KS, Kobayashi T, et al. Validation of quantitative prognostic prediction using ADV score for resection of hepatocellular carcinoma: a Korea-Japan collaborative study with 9200 patients. J Hepatobiliary Pancreat Sci 2023;30:993-1005. 10.1002/jhbp.1319 [DOI] [PubMed]
- 18.Hwang S, Song GW, Ahn CS, Kim KH, Moon DB, Ha TY, et al. Salvage living donor liver transplantation for hepatocellular carcinoma recurrence after hepatectomy: quantitative prediction using ADV score. J Hepatobiliary Pancreat Sci 2021;28:1000-13. 10.1002/jhbp.863 [DOI] [PubMed]
- 19.Cao S, Zhou Z, Chen C, Li W, Liu J, Xu J, et al. Early identification of hepatocellular carcinoma patients at high-risk of recurrence using the ADV score: a multicenter retrospective study. World J Surg Oncol 2024;22:240. 10.1186/s12957-024-03523-1 [DOI] [PMC free article] [PubMed]
- 20.Park GC, Hwang S, Park YH, Choi JU; Korean Liver Cancer Study Group. Validation of prognostic impact of ADV score for resection of hepatocellular carcinoma: analysis using Korea Liver Cancer Registry Database. Ann Surg Treat Res 2020;98:235-46. 10.4174/astr.2020.98.5.235 [DOI] [PMC free article] [PubMed]
- 21.Hwang S, Kim KH, Moon DB, Ahn CS, Ha TY, Song GW, et al. Prediction of post-resection prognosis using the ADV score for huge hepatocellular carcinomas ≥13 cm. J Liver Cancer 2021;21:45-57. 10.17998/jlc.21.1.45 [DOI] [PMC free article] [PubMed]
- 22.Jung DH, Hwang S, Lee YJ, Kim KH, Song GW, Ahn CS, et al. Small hepatocellular carcinoma with low tumor marker expression benefits more from anatomical resection than tumors with aggressive biology. Ann Surg 2019;269:511-9. 10.1097/SLA.0000000000002486 [DOI] [PubMed]
- 23.Hwang S, Lee KJ, Moon DB, Song GW, Jung DH, Kim YK, et al. Prognostic impact of serum soluble PD-1 and ADV score for living donor liver transplantation in patients with previously untreated hepatocellular carcinoma. Ann Surg Treat Res 2022;102:46-54. 10.4174/astr.2022.102.1.46 [DOI] [PMC free article] [PubMed]
- 24.Hwang S, Jung DH, Song GW. ADV score is a reliable surrogate biomarker of hepatocellular carcinoma in liver resection and transplantation. Ann Liver Transplant 2023;3:86-93. 10.52604/alt.23.0023 [DOI]
- 25.Kim SH, Moon DB, Park GC, Lee SG, Hwang S, Ahn CS, et al. Preoperative prediction score of hepatocellular carcinoma recurrence in living donor liver transplantation: validation of SNAPP score developed at Asan Medical Center. Am J Transplant 2021;21:604-13. 10.1111/ajt.16227 [DOI]
- 26.Olmedillas-López S, Olivera-Salazar R, García-Arranz M, García-Olmo D. Current and emerging applications of droplet digital PCR in oncology: an updated review. Mol Diagn Ther 2022;26:61-87. 10.1007/s40291-021-00562-2 [DOI] [PubMed]
- 27.Postel M, Roosen A, Laurent-Puig P, Taly V, Wang-Renault SF. Droplet-based digital PCR and next generation sequencing for monitoring circulating tumor DNA: a cancer diagnostic perspective. Expert Rev Mol Diagn 2018;18:7-17. 10.1080/14737159.2018.1400384 [DOI] [PubMed]