Evaluating the combined diagnostic power of alpha-fetoprotein and protein induced by vitamin K absence or antagonist-II for hepatocellular carcinoma

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide, particularly prevalent in Asia and sub-Saharan Africa, where its incidence and mortality rates remain high (1,2). Globally, HCC ranks fourth in cancer-related deaths (1,3). The etiology is diverse, with chronic hepatitis B (CHB) being one of the primary risk factors (4,5). Additionally, alcoholic liver disease, non-alcoholic fatty liver disease (NAFLD), and cirrhosis are also significant contributors to HCC (6-8). HCC typically develops without noticeable early symptoms and is often detected only at advanced stages, significantly reducing treatment efficacy and patient survival rates. Early diagnosis is crucial for improving the prognosis of HCC patients (9,10); thus, identifying effective early biomarkers holds substantial clinical significance. There is an urgent global need to enhance early detection rates for HCC (11), posing a significant challenge to public health systems.

In terms of biomarkers, alpha-fetoprotein (AFP) is a traditional and widely used HCC marker; however, its sensitivity and specificity remain inadequate (12), particularly for diagnosing early-stage or small tumors (13,14). These limitations have prompted researchers to explore more effective biomarkers to enhance the accuracy of HCC diagnosis. Liquid biopsy techniques, such as circulating tumor DNA (ctDNA), circulating tumor cells (CTCs), and miRNA detection, are emerging as a non-invasive approach for early HCC diagnosis. These methods provide crucial genetic insights into liver tumors, aiding early cancer detection and predicting recurrence (15,16). Recent studies indicate that HCC immune evasion is closely linked to tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), and regulatory T cells (Tregs) (17,18). Targeting these immune cells may enhance anti-tumor immunity and improve treatment outcomes.

Protein induced by vitamin K absence or antagonist-II (PIVKA-II), also known as des-gamma-carboxy prothrombin, has emerged as a promising biomarker for HCC diagnosis in recent years. Unlike AFP, PIVKA-II expression significantly increases during liver carcinogenesis (19), which is associated with abnormal hepatocyte vitamin K metabolism (20). Studies have shown that PIVKA-II levels remain low in non-HCC patients but are markedly elevated in those with HCC, especially in patients where AFP levels are not significantly elevated, making PIVKA-II detection particularly valuable (21,22). Moreover, PIVKA-II demonstrates high sensitivity and specificity, making it a strong candidate for early HCC diagnosis (23,24). Thus, applying PIVKA-II enhances diagnostic accuracy and serves as a crucial complementary marker when AFP is insufficient.

Combining AFP with other biomarkers, such as PIVKA-II, is a key strategy to enhance the sensitivity and specificity of HCC diagnosis. Studies have shown that this combined testing approach effectively compensates for the limitations of using a single biomarker (24), particularly in improving accuracy in early-stage HCC diagnosis (9). By utilizing a multi-biomarker strategy, the strengths of each marker are leveraged, significantly increasing diagnostic sensitivity and specificity, especially in heterogeneous HCC cases (24,25). A new meta-analysis indicates that combined detection of AFP and PIVKA-II offers higher sensitivity and specificity than either biomarker alone. Single biomarkers often fall short for small tumors or early-stage HCC, while their combination significantly improves diagnostic accuracy and reduces missed diagnoses. Advances in technology have also led researchers to explore multi-biomarker panels, including AFP, PIVKA-II, and hepatocyte growth factor (HGF), which further enhance diagnostic precision and provide more reliable detection in heterogeneous patient populations (26,27).

This study systematically evaluated the diagnostic value of HGF, AFP, and PIVKA-II in HCC, focusing on the added value of combining PIVKA-II with AFP. Unlike previous studies, it incorporated large-scale patient data and advanced statistical methods, including meta-analysis, for a comprehensive assessment. The study introduced PIVKA-II as a key diagnostic biomarker and explored novel biomarker combinations, such as AFP + PIVKA-II + HGF, to enhance diagnostic accuracy. Establishing a more precise strategy for early HCC detection held strong translational potential, providing clinically actionable insights to improve patient outcomes and reduce disease burden. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-863/rc).

Methods

Study design and participants

This study was conducted at the Second Clinical Hospital of Lanzhou University, investigating 210 cases of HCC, 270 cases of CHB, and 92 post-treatment HCC cases from June 2021 to February 2023. Additionally, from March to June 2023, we enrolled 88 patients with chronic viral hepatitis as a control group and 91 HCC patients as a validation cohort (Figure S1). A power analysis was conducted with α=0.05 and power (1 − β) =0.80 to ensure an 80% probability of correctly rejecting the null hypothesis. Using G*Power, the required sample size was determined to be 40 per group, which aligns with the collected data. Eligible participants were Chinese citizens aged 18 years and older of any gender. The diagnoses of HCC and CHB were made according to the “Diagnostic and treatment guidelines for primary HCC” and the “Guidelines for the prevention and treatment of chronic hepatitis B”, respectively (28). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Clinical Ethics Committee of the Second Hospital of Lanzhou University (No. D2023A-459), and informed consent was waived due to the retrospective nature of the study.

Exclusion criteria included pregnant women, patients with autoimmune diseases, liver cysts, bleeding or thrombotic disorders, cholangiocarcinoma, metastatic liver cancer, or other concurrent malignancies. Patients with severe heart, lung, or brain diseases or those who had taken oral vitamin K or warfarin within 2 weeks before or after the PIVKA-II assessment were also excluded.

Specimen collection and analysis

Venous blood samples were collected from study participants and centrifuged at 3,000 rpm for 10 minutes to separate the serum for subsequent HGF assessment. Serum samples were obtained preoperatively or before radiofrequency ablation treatment and were immediately stored at −80 ℃ to prevent degradation, with strict precautions taken to avoid repeated freeze-thaw cycles.

Comprehensive clinicopathological data were collected, including demographic information (age, gender), tumor characteristics (size, number), presence of portal vein tumor thrombus, evidence of distant metastasis (particularly lymph node involvement), and complications such as gastrointestinal bleeding, hepatorenal syndrome, hepatic encephalopathy, HCC nodule rupture and hemorrhage, and secondary infections.

Measurement of serum AFP and PIVKA-II

Serum levels of AFP and PIVKA-II were quantitatively measured using electrochemiluminescence on a Roche Cobas p501 analyzer (Roche Diagnostics, Mannheim, Germany), following the manufacturer’s protocol and specific Roche reagents.

Measurement of serum HGF concentration

Serum concentrations of HGF were evaluated using an enzyme-linked immunosorbent assay (ELISA) kit (Abclonal, Wuhan, China), following the manufacturer’s instructions. The plate was washed three times with phosphate-buffered saline with Tween (PBST), then 100 µL of diluent (for blank control), standards and serum samples were added and incubated at 37 ℃ for 2 hours. After the washing step, 100 µL of biotin-labeled antibody was added and incubated at 37 ℃ for 1 hour. The plate was washed three more times, followed by adding 100 µL of biotinylated antibody, and incubated at 37 ℃ for 30 minutes. After three washes, 100 µL of substrate and 50 µL of stop solution were added. The absorbance at 450 nm was recorded using a microplate reader (Thermo, Waltham, MA, USA).

Meta-analysis literature search strategy

We developed a comprehensive literature search strategy to systematically evaluate the diagnostic efficacy of serum HGF, AFP, and PIVKA-II in HCC and compare the performance of these biomarkers across various studies. The literature search was conducted across major medical databases, including PubMed, Web of Science, Embase, and China National Knowledge Infrastructure (CNKI), covering all relevant studies from the inception of these databases up to January 2024. To ensure the search’s thoroughness and accuracy, we utilized carefully selected keywords and Boolean expressions such as “hepatocellular carcinoma”, “HGF”, “AFP”, “PIVKA-II”, “diagnostic biomarkers”, “ROC curve”, “sensitivity”, “specificity”, and “AUC”. The search strategy was tailored to each database’s characteristics, incorporating keyword searches in titles and abstracts and searches using subject terms to capture all relevant studies. Additionally, manual searches were performed, including a review of reference lists from relevant studies and abstracts from vital academic conferences to identify important research that the database searches may have missed. This comprehensive and rigorous search strategy allowed us to gather sufficient high-quality data, enabling a meta-analysis of the diagnostic performance of HGF, AFP, and PIVKA-II across different populations and study conditions, thus verifying their universality and consistency in HCC diagnosis.

Inclusion and exclusion criteria for literature

Strict inclusion and exclusion criteria were applied to ensure the rigor and specificity of this systematic review and meta-analysis. Studies included had to involve adult patients diagnosed with HCC and assess the diagnostic performance of serum biomarkers such as HGF, AFP, and PIVKA-II. Priority was given to studies reporting receiver operating characteristic (ROC) curves, sensitivity, specificity, and area under the curve (AUC). HCC diagnosis must be based on clear clinical criteria with well-defined detection methods and results.

Inclusion criteria were strictly defined to minimize potential confounding factors (e.g., comorbidities, concurrent treatments), excluding patients with significant comorbidities or recent chemotherapy that could influence biomarker levels. The study included randomized controlled trials (RCTs), cohort studies, case-control studies, and other observational studies with appropriate control groups. Studies in CHB patients assessing these biomarkers’ diagnostic performance were also included as controls.

Multivariate regression models or propensity score matching were used to adjust for confounders and control their impact on results. Comprehensive clinical data on comorbidities, medication use, and treatment history were systematically collected to account for relevant factors. Sensitivity analyses were conducted to assess result consistency under different assumptions, ensuring the robustness of findings.

For exclusion criteria, we eliminated studies that did not meet the inclusion requirements. It included studies involving non-adult patients or other types of liver diseases. Additionally, studies needing complete data, failing to conduct ROC analysis, or missing crucial diagnostic metrics such as AUC were excluded. We also excluded review articles, case reports, opinion pieces, and studies with only abstracts or those without full-text availability. Duplicate publications were removed to prevent data redundancy and bias, ensuring the reliability and consistency of the meta-analysis results.

Literature coding and quality assessment

In this study, each included article was meticulously coded to ensure systematic organization and accuracy for subsequent analysis. The coding process covered basic information such as author names, publication year, country or region of the study, sample size, and study design type. Specific biomarker details (e.g., HGF, AFP, PIVKA-II), measurement methods, detection tools, control group settings, patient clinical characteristics, and key statistical outcomes were recorded. Data coding was systematically organized using Excel spreadsheets, ensuring clarity and easy analysis. The quality of each study was assessed using the Cochrane risk of bias tool, evaluating aspects such as selection bias, detection bias, and reporting bias to ensure high reliability and internal validity of the results.

The quality of randomization and allocation concealment in the included studies was assessed to minimize selection bias. Randomization ensures baseline balance between groups and reduces confounding using stratified randomization and minimization. The former stratify participants by key variables (e.g., age, sex, disease severity) before random assignment within each stratum, while the latter dynamically adjusts allocations based on real-time participant characteristics using a computational algorithm to enhance balance and reduce bias. Studies with unclear randomization methods were classified as having a high risk of bias or uncertain risk.

Allocation concealment prevents researchers from predicting or influencing group assignments. It was assessed using the sealed envelope method and central randomization. The sealed envelope method employs opaque, sealed, and sequentially numbered envelopes to ensure allocation remains unknown before enrollment. Central randomization is conducted by an independent entity (e.g., a computerized system or third-party center) to further reduce selection bias. Studies lacking a clear description of allocation concealment were classified as having uncertain risk, while those with predictable allocation (e.g., transparent envelopes or prior knowledge of assignments) were rated as having high selection bias risk.

Data extraction

Data extraction focused on retrieving each study’s diagnostic performance metrics for HGF, AFP, and PIVKA-II. These metrics included, but were not limited to, the area under the ROC curve, sensitivity, specificity, positive predictive value, and negative predictive value. Additionally, we extracted information on detection methods, sample size, patient grouping, and other key statistical data reported in each study. Continuous and categorical variables were recorded separately to ensure data completeness and accuracy. Two researchers conducted the data extraction process independently, with any discrepancies resolved through discussion or adjudication by a third party. This detailed extraction process ensured the consistency and reliability of the data used in the meta-analysis.

Statistical analysis

Statistical analyses were performed using SPSS 26.0 (SPSS, Chicago, IL, USA). Non-normally distributed data were presented as a median and interquartile range (IQR), while categorical data were expressed as n (%) and analyzed using the χ² test. Right-skewed variables underwent natural log transformation for improved statistical accuracy. The Mann-Whitney U test was used for two-group comparisons, the Kruskal-Wallis H test for multiple groups, and Spearman correlation analysis was used to assess relationships between quantitative variables.

Diagnostic performance was evaluated using ROC curves, with AUC, sensitivity, and specificity calculated. Multivariable logistic regression assessed the independent contributions of HGF, AFP, PIVKA-II, and clinical variables [age, sex, aspartate aminotransferase (AST), albumin (ALB)] to HCC diagnosis, including only variables with P<0.05 in univariable analysis. Stepwise regression was used for variable selection, and β coefficients and P values were calculated. Model fit was assessed using the Hosmer-Lemeshow test, and VIF was used to detect collinearity. R software was employed to generate nomograms and further evaluate predictive performance.

Meta-analysis was conducted using RevMan 5.4 and R software, with r values and 95% confidence intervals (CIs) as effect size indicators. Cochran’s Q test and I² statistics were used to assess heterogeneity, further analyzed by Tau² and Chi² tests. Sensitivity analysis evaluated result robustness, while funnel plots and Egger’s test assessed publication bias. Data were pooled using a random-effects model to estimate mean differences (MDs) between experimental and control groups, ensuring result reliability. A significance level of P<0.05 was applied to all tests.

Results

Serum AFP and PIVKA-II levels in HCC patients

Table 1 compares the baseline characteristics of 210 HCC patients and 270 CHB patients. Our study found that serum levels of PIVKA-II and AFP were significantly elevated in HCC patients compared to those with CHB (P<0.001), as shown in Figure 1. The median serum AFP level in HCC patients was 354.09 ng/mL [95% confidence interval (CI): 25.80–2,414.03], compared to 3.03 ng/mL (95% CI: 2.12–4.54) in CHB patients. The median serum PIVKA-II level in HCC patients was 3,103.92 mAU/mL (95% CI: 358.86–18,353.68), whereas in CHB patients, it was 25.30 mAU/mL (95% CI: 20.55–30.85). Univariate analysis, based on the “2020 CSCO guidelines for the diagnosis and treatment of primary liver cancer”, highlighted an increased risk of liver cancer in men over 40 years and women over 45 years. There were statistically significant differences in gender and the proportion of high-risk individuals (P<0.001).

Figure 1 Comparison of serum AFP and PIVKA-II levels between HCC and CHB patients. (A) Serum AFP and PIVKA-II levels in HCC patients and CHB patients. (B) Serum AFP and PIVKA-II levels in the preoperative and postoperative groups. AFP, alpha-fetoprotein; CHB, chronic hepatitis B; HCC, hepatocellular carcinoma; PIVKA-II, protein induced by vitamin K absence or antagonist-II.

Diagnostic accuracy of AFP and PIVKA-II for HCC

The diagnostic potential of AFP and PIVKA-II as independent biomarkers for HCC was evaluated using ROC curves, with optimal cut-off values determined by the Youden index (Figure 2A-2C). To distinguish HCC from CHB, the cut-off value for PIVKA-II was set at 10.695 ng/mL, achieving a sensitivity of 96.2% and a specificity of 96.3%. The cut-off value for AFP was established at 59.475 mAU/mL, with a sensitivity of 82.9% and a specificity of 93.7%. PIVKA-II demonstrated superior diagnostic performance, with an AUC of 0.979 (95% CI: 0.963–0.994), compared to an AUC of 0.923 (95% CI: 0.895–0.950) for AFP. Combining AFP with PIVKA-II further enhanced diagnostic accuracy, yielding an AUC of 0.989 (95% CI: 0.979–0.998), with sensitivity and specificity increasing to 98.6% and 91.1%, respectively. This combined approach significantly improved detection efficacy, outperforming AFP alone.

Figure 2 ROC curve analysis of AFP and PIVKA-II in HCC diagnosis and their correlation with tumor characteristics. (A) The AUC and diagnostic accuracy of AFP and PIVKA-II for HCC patients versus CHB patients. (B) Serum AFP and PIVKA-II levels in different tumor diameters. (C) Correlation of AFP, PIVKA-II and tumor diameter. ns, not significant (P>0.05). AFP, alpha-fetoprotein; AUC, area under the curve; CI, confidence interval; CHB, chronic hepatitis B; HCC, hepatocellular carcinoma; PIVKA-II, protein induced by vitamin K absence or antagonist-II; ROC, receiver operating characteristic.

Relationship between AFP, PIVKA-II, and HCC clinical characteristics

Analysis of HCC patients revealed a significant association between serum levels of AFP and PIVKA-II and various clinicopathological features (Table 2). Serum PIVKA-II concentrations showed a strong correlation with tumor size (P<0.001) (Figure 1). Patients with lymph node and distant metastasis had significantly higher serum levels of AFP and PIVKA-II than those without metastasis (P<0.001). Elevated PIVKA-II levels were also associated with complications (P=0.003), and portal vein thrombosis correlated with higher levels of both AFP and PIVKA-II (P<0.001). In contrast, PIVKA-II levels were not significantly related to age (P=0.51), sex (P=0.29), or tumor number (P=0.14). Similarly, AFP levels showed no significant association with age (P=0.54), sex (P=0.76), tumor number (P=0.68), or comorbidities (P=0.04).

Spearman rank correlation analysis further delineated the relationship between serum AFP and PIVKA-II levels and tumor diameter in HCC patients. A weak positive correlation was observed between serum levels of PIVKA-II and AFP (r=0.338, P<0.001). The correlation between tumor diameter and serum levels of PIVKA-II (r=0.568, P<0.001) and AFP (r=0.299, P<0.001) was stronger, with PIVKA-II showing a more robust association than AFP.

Value of PIVKA-II in evaluating treatment efficacy in HCC patients

A retrospective study involving 92 HCC patients treated between June 2021 and February 2023 assessed the impact of various treatment modalities on serum levels of AFP and PIVKA-II. Significant differences in AFP and PIVKA-II levels were observed before and after treatment (Figure 1A,1B). Patients were categorized based on the treatment received: liver resection (n=29), transarterial chemoembolization (TACE) (n=50), and other surgical procedures (n=13) (Figure S2). Demographic data for these patients are presented in Table 3. There was a significant difference in tumor size distribution across the treatment groups (P=0.02). While preoperative AFP and PIVKA-II levels were consistent across treatment types, postoperative PIVKA-II levels varied. The postoperative PIVKA-II level was lowest in the hepatectomy group, with a median of 30.82 mAU/mL (95% CI: 18.06–56.94), compared to 180.64 mAU/mL (95% CI: 47.57–1,074.80) in the TACE group and 248.54 mAU/mL (95% CI: 144.04–1,058.42) in the other treatment group (P<0.001).

Development of a nomogram for predicting HCC incidence

Multivariate analysis identified age, sex, AST, ALB, and the natural logarithm of AFP (LnAFP) and natural logarithm of PIVKA-II (LnPIVKA-II) as significant independent predictors of HCC risk in patients with hepatitis B virus (HBV) infection (P<0.05). These variables were incorporated into a risk prediction model using R software. Given the wide range of AFP and PIVKA-II levels compared to other indicators, natural logarithm transformations (Ln) were applied to normalize the data. A nomogram combining LnAFP and LnPIVKA-II was developed to represent HCC risk, as shown in Figure 3A-3D visually. Based on multivariate logistic regression analysis from the training cohort, a predictive equation was established: risk score = −7.151 + 1.291 × LnPIVKA-II + 1.274 × LnAFP − 0.013 × AST − 0.198 × ALB + 0.133 × age − 2.120 × sex (0= male, 1= female). The model’s diagnostic performance, illustrated by the ROC curve, showed an AUC of 0.995 (95% CI: 0.990–1.000), a Youden index of 1.959, with a sensitivity of 0.981 and a specificity of 0.978, indicating strong discriminative ability. Figure 3A-3D displays the calibration curve, demonstrating close alignment with actual outcomes, which indicates high predictive accuracy. The Hosmer-Lemeshow test confirmed the model’s goodness-of-fit, with a P value >0.05, indicating an excellent fit for the nomogram.

Figure 3 Nomogram for predicting HCC incidence based on multivariate logistic regression analysis. (A) Predictive modelling of HCC based on logistic regression analysis (nomogram). (B) The AUC and diagnostic accuracy of predictors (gender, age, LnAFP, LnPIVKA-II, AST, and ALB) and predictive models. (C) The AUC and diagnostic accuracy of the validation set. (D) Goodness of fit of the nomogram model. ALB, albumin; AST, aspartate aminotransferase; AUC, area under the curve; CI, confidence interval; HCC, hepatocellular carcinoma; LnAFP, natural logarithm of alpha-fetoprotein; LnPIVKA-II, natural logarithm of protein induced by vitamin K absence or antagonist-II.

Validation of the HCC prediction model

To confirm the statistical model’s predictive value, we selected a cohort of 91 HCC patients and 88 viral hepatitis controls between March and June 2023. ROC curve analysis was performed using R software, and the results demonstrated an AUC of 0.986 (95% CI: 0.970–1.000), confirming the model’s strong predictive capability, as shown in Figure 3A-3D.

Evaluation of combined diagnostic efficacy of HGF, AFP, and PIVKA-II in HCC patients

This study also evaluated the diagnostic utility of combining HGF, AFP, and PIVKA-II for a comprehensive assessment of HCC. Data from patients admitted between March and June 2023 indicated that PIVKA-II outperformed AFP and HGF in detecting HCC, as shown in Figure 4A,4B. PIVKA-II demonstrated a sensitivity of 0.901, a specificity of 0.989, and an AUC of 0.970 (95% CI: 0.943–0.997). AFP showed a specificity of 0.886, a sensitivity of 0.769, and an AUC of 0.890 (95% CI: 0.842–0.939). HGF had a sensitivity of 0.440, a specificity of 0.818, and an AUC of 0.617 (95% CI: 0.534–0.699), with statistically significant differences (P=0.007<0.05). The AUC of HGF greater than 0.5 indicates its potential predictive value in HCC diagnosis (Figure 4A,4B). The combined use of AFP and PIVKA-II resulted in a sensitivity of 0.956, a specificity of 0.875, and an AUC of 0.977 (95% CI: 0.954–1.000). When HGF was combined with AFP and PIVKA-II, the sensitivity and specificity were 0.967 and 0.784, respectively, with an AUC of 0.976 (95% CI: 0.951–1.000). The combination of all three biomarkers demonstrated the highest sensitivity, with an AUC comparable to that of AFP and PIVKA-II alone, highlighting the synergistic diagnostic value of using HGF alongside AFP and PIVKA-II in HCC diagnosis.

Figure 4 Diagnostic efficacy of combined HGF, AFP, and PIVKA-II in differentiating HCC from CHB patients. (A) The AUC and diagnostic accuracy of AFP, PIVKA-II, and HGF in patients with HCC and patients with CHB of the validation set. (B) Correlation of HGF with AFP, PIVKA-II, and other serum markers. γ-GGT, γ-glutamyl transpeptidase; AFP, alpha-fetoprotein; ALB, albumin; AST, aspartate aminotransferase; AUC, area under the curve; CHB, chronic hepatitis B; HCC, hepatocellular carcinoma; HGF, hepatocyte growth factor; PIVKA-II, protein induced by vitamin K absence or antagonist-II.

Significance of HGF in HCC patients

This section explores the relationship between serum HGF levels and traditional HCC biomarkers, including AFP and PIVKA-II, as well as various liver function tests. As shown in Figure 4A,4B, Spearman correlation analysis revealed a positive correlation between serum HGF and PIVKA-II levels (r=0.308, P=0.003), indicating a significant association between the two. In contrast, no significant correlation was found between serum HGF and AFP levels (r=0.169, P=0.11), suggesting their clinical independence. Additionally, HGF levels were positively correlated with several liver function parameters, including total bilirubin (r=0.414, P<0.001), AST (r=0.483, P<0.001), and γ-glutamyl transferase (γ-GGT) (r=0.425, P<0.001), highlighting HGF’s potential as an indicator of liver function deterioration. A negative correlation was also observed between HGF and ALB levels (r=−0.349, P<0.001), emphasizing its prognostic value in assessing liver synthesis function.

Summary of literature screening and quality assessment

Seventy articles were identified through database searches in PubMed, Web of Science, CNKI, and Wan Fang Data (Figure 5A). After removing duplicates, 56 articles remained for further screening. Based on the predetermined criteria, 32 non-randomized studies and studies unrelated to HCC diagnosis were excluded after reviewing the titles and abstracts. The remaining 24 articles were subjected to full-text review, resulting in the exclusion of 17 articles due to insufficient data, lack of full-text availability, or inconsistencies with the study’s evaluation criteria. Ultimately, seven articles were deemed eligible and included in the quantitative meta-analysis. The basic characteristics of these articles, including first author, publication year, sample size, and treatment method, were summarized in Table 4.

Figure 5 Literature screening process and quality assessment results. (A) Flowchart of literature inclusion. (B) Summary of bias risk assessment for included studies. CNKI, China National Knowledge Infrastructure.

The quality of the included studies was assessed using the Cochrane risk of bias tool (Figure 5B). Potential biases in each study, including random sequence generation, allocation concealment, blinding, and outcome assessment, were classified as low, high, or unclear risk. Studies with inadequate randomization or allocation concealment were classified as having a high risk of selection bias. Cases with insufficient information to determine bias risk were marked as needing clarification. Conversely, studies with proper randomization, adequate allocation concealment, and appropriate blinding were classified as having a low risk of bias. This systematic evaluation ensured that each study underwent a comprehensive and impartial quality review.

Diagnostic efficacy of AFP in HCC

This meta-analysis systematically evaluated the diagnostic efficacy of AFP in HCC and compared AFP performance across various studies. The results showed that AFP levels in the experimental group were significantly higher than those in the control group, with a pooled MD of 200.00 (95% CI: 65.83–334.16), supporting the potential of AFP as a diagnostic marker for HCC (Figure 6A). However, due to significant heterogeneity between studies (I²=100%), these findings should be interpreted cautiously and applied in clinical practice considering specific contexts.

Figure 6 Meta-analysis of AFP in HCC diagnosis. (A) Forest plot of the meta-analysis for AFP in HCC diagnosis. (B) Sensitivity analysis forest plot for AFP in HCC diagnosis. (C) Publication bias assessment results for AFP in HCC diagnosis. AFP, alpha-fetoprotein; CI, confidence interval; HCC, hepatocellular carcinoma; MD, mean difference; SD, standard deviation.

In sensitivity analysis, recalculating the pooled effect size by sequentially excluding each study (Figure 6B, Table 5) showed that the MD ranged from 172.90 to 220.09, with all P values remaining significant (P<0.01). This indicates that despite high heterogeneity, the overall results are robust, with no single study significantly influencing the overall conclusions of the meta-analysis.

Funnel plot analysis indicated some asymmetry in data point distribution, suggesting the possibility of publication bias. However, most studies were clustered near the center of the funnel plot (Figure 6C), demonstrating a degree of robustness in the meta-analysis results. Thus, the results support the efficacy of AFP in diagnosing HCC. Although signs of bias exist, the conclusions still hold clinical relevance.

Discussion

HCC is the fifth most common cancer worldwide and has a very high mortality rate (1,3). Traditionally, AFP has been widely used as a diagnostic marker for HCC; however, its sensitivity and specificity are not ideal, particularly for early-stage HCC detection (20,21). In recent years, PIVKA-II has emerged as a novel biomarker, showing promising diagnostic potential (22,23,36). This study aimed to comprehensively evaluate the diagnostic value of HGF, AFP, and PIVKA-II in HCC, focusing on their performance individually and in combination.

Our findings indicated that levels of AFP, PIVKA-II, and HGF were significantly elevated in HCC patients compared to those with CHB, consistent with previous studies that reported increased levels of AFP and PIVKA-II in HCC patients (24,37). However, this study further quantified these biomarkers using electrochemiluminescence and ELISA techniques and assessed their diagnostic performance through ROC curve analysis. The results confirmed that PIVKA-II had an AUC of 0.970, demonstrating superior diagnostic performance compared to AFP (AUC =0.890) and HGF (AUC =0.617). These findings suggest that PIVKA-II may be a more valuable biomarker than traditional AFP for diagnosing HCC.

The results of this study indicate that HGF performs relatively poorly in HCC diagnosis, with an AUC of 0.617. This finding differs slightly from previous studies that occasionally reported HGF as a potential marker for tumor growth and invasion. These discrepancies may be due to differences in sample selection, geographic variations, or measurement techniques. Although the AUC value for HGF is not high, its combination with AFP and PIVKA-II significantly improved diagnostic sensitivity and specificity, highlighting the importance of using a combined biomarker approach in research.

Further analysis of sensitivity and specificity revealed that the combined use of PIVKA-II and AFP significantly increased the diagnostic sensitivity and specificity for HCC to 98.6% and 91.1%, respectively. This study has significant scientific and clinical value.

Scientifically, it is the first to comprehensively evaluate the combined application of HGF, AFP, and PIVKA-II in HCC diagnosis, revealing the correlation between HGF and PIVKA-II and validating the relationship between HGF and liver function indicators. These findings provide fundamental data for future research on the biological mechanisms of HGF in HCC. Clinically, this study demonstrates that PIVKA-II outperforms AFP in diagnostic performance, particularly in AFP-insensitive patients. Additionally, the combination of AFP and PIVKA-II significantly improves the sensitivity and specificity of early HCC diagnosis, offering a new strategy for screening and follow-up. While HGF alone has limited diagnostic value, its combination with other biomarkers may have greater clinical significance. This study validates and expands existing knowledge and provides a more precise biomarker combination strategy for early HCC diagnosis, offering new insights into biomarker selection and combination in clinical practice. With the advancement of new technologies and biomarkers, HCC diagnosis will become more accurate, improving early detection, treatment outcomes, and patient prognosis.

The meta-analysis results further support the importance of AFP in HCC diagnosis. Despite high heterogeneity among the included studies (I²=100%), the pooled MD showed that AFP levels in HCC patients were significantly higher than those in the control group, supporting the reliability of AFP as a diagnostic biomarker for HCC. However, the substantial heterogeneity suggests that variations in study design, sample characteristics, and detection techniques across studies may impact the diagnostic accuracy of AFP. Therefore, while AFP remains a valuable tool for HCC detection, combining it with other diagnostic methods or biomarkers (such as PIVKA-II) may enhance diagnostic precision.

The main limitation of this study lies in its retrospective design. Selection bias may affect sample representativeness, and the limited sample size restricts the generalizability of the findings. Information bias could arise due to variations in biomarker measurement methods, timing, and laboratory techniques across different records, potentially impacting data accuracy and comparability. Moreover, the absence of prospective follow-up data prevents the assessment of biomarker dynamics during disease progression and their role in clinical decision-making. While PIVKA-II demonstrated superior diagnostic performance, its consistency across different populations, disease stages, and surgical types requires further validation. Additionally, the diagnostic value of HGF as a standalone biomarker is limited, and its interactions with other potential biomarkers remain insufficiently explored, which may lead to an underestimation of its role in HCC diagnosis.

This study has significant implications for future research and clinical practice. Future studies should expand sample sizes and incorporate prospective validation in independent cohorts to enhance the robustness and generalizability of the findings, particularly across different populations and disease stages. Further investigations are needed to elucidate the role of HGF in HCC progression and its relationship with PIVKA-II, as well as to integrate clinical data and multi-omics analysis for developing more precise HCC early detection and risk prediction models, advancing precision medicine. Clinically, this study confirms that PIVKA-II combined with AFP significantly improves the sensitivity and specificity of early HCC diagnosis, providing a more effective screening tool for high-risk individuals, such as patients with CHB. Integrating AFP, PIVKA-II, and HGF could further optimize personalized screening and monitoring strategies, enhancing detection rates and enabling more flexible management approaches for high-risk populations. In the future, biomarker integration with multi-omics data could refine HCC clinical management and precision therapy, ultimately improving patient survival and quality of life. This study introduces a new biomarker combination strategy for HCC early diagnosis and personalized screening, offering valuable insights for clinical applications and scientific advancements.

Conclusions

This study evaluated the diagnostic value of HGF, AFP, and PIVKA-II in HCC. Results showed that PIVKA-II demonstrated the highest diagnostic performance (AUC =0.970), surpassing AFP (AUC =0.890) and HGF (AUC =0.617). Logistic regression analysis identified LnAFP, LnPIVKA-II, AST, ALB, sex, and age as independent predictors of HCC, and the constructed risk prediction model exhibited high diagnostic accuracy (AUC =0.995, sensitivity =98.1%, specificity =97.8%). Additionally, HGF levels correlated significantly with liver function indicators, suggesting its potential role in reflecting liver function status (Figure 7).

Figure 7 Molecular mechanism of PIVKA-II, AFP, and HGF in HCC diagnosis. AFP, alpha-fetoprotein; HCC, hepatocellular carcinoma; HGF, hepatocyte growth factor; PIVKA-II, protein induced by vitamin K absence or antagonist-II.

PIVKA-II showed superior diagnostic value for HCC, particularly in early-stage detection. However, its performance may vary across populations, disease stages, and geographic factors, necessitating further validation. Future studies should conduct large-scale, independent validations across diverse populations and explore the applicability of these biomarkers in other chronic liver diseases, such as cirrhosis and fatty liver disease, to refine HCC early screening and personalized diagnostic strategies.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-863/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-863/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-863/prf

Funding: This study was supported by the National Natural Science Foundation of China (No. 32060231), the Gansu Province Health Industry Scientific Research Project (No. GSWSKY2023-03), the Gansu Province Natural Science Foundation (Nos. 21JR11RA115 and 23JR5RA1632), the Cuiying Scientific and Technological Innovation Program of Lanzhou University Second Hospital (No. CY2024-MS-B15), and the Young Scholars Innovation Foundation of Lanzhou (No. 2024-QN-19).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-863/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Clinical Ethics Committee of the Second Hospital of Lanzhou University (No. D2023A-459), and informed consent was waived due to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Brown ZJ, Tsilimigras DI, Ruff SM, et al. Management of Hepatocellular Carcinoma: A Review. JAMA Surg 2023;158:410-20. [Crossref] [PubMed]
2. Kulik L, El-Serag HB. Epidemiology and Management of Hepatocellular Carcinoma. Gastroenterology 2019;156:477-491.e1. [Crossref] [PubMed]
3. Sagnelli E, Macera M, Russo A, et al. Epidemiological and etiological variations in hepatocellular carcinoma. Infection 2020;48:7-17. [Crossref] [PubMed]
4. Hu X, Wen Y. The hepatic steatosis, steatohepatitis and metabolic dysfunction: distinct roles in hepatocellular carcinoma occurrence in chronic hepatitis B patients. Hepatol Int 2024;18:1338-9. [Crossref] [PubMed]
5. Sayiner M, Golabi P, Younossi ZM. Disease Burden of Hepatocellular Carcinoma: A Global Perspective. Dig Dis Sci 2019;64:910-7. [Crossref] [PubMed]
6. Mauro E, Forner A. Hepatocellular adenoma in men: Is it time for a precision approach? Liver Int 2021;41:2246-8. [Crossref] [PubMed]
7. Golabi P, Paik JM, Eberly K, et al. Causes of death in patients with Non-alcoholic Fatty Liver Disease (NAFLD), alcoholic liver disease and chronic viral Hepatitis B and C. Ann Hepatol 2022;27:100556. [Crossref] [PubMed]
8. Toh MR, Wong EYT, Wong SH, et al. Global Epidemiology and Genetics of Hepatocellular Carcinoma. Gastroenterology 2023;164:766-82. [Crossref] [PubMed]
9. Piñero F, Dirchwolf M, Pessôa MG. Biomarkers in Hepatocellular Carcinoma: Diagnosis, Prognosis and Treatment Response Assessment. Cells 2020;9:1370. [Crossref] [PubMed]
10. Menyhárt O, Nagy Á, Győrffy B. Determining consistent prognostic biomarkers of overall survival and vascular invasion in hepatocellular carcinoma. R Soc Open Sci 2018;5:181006. [Crossref] [PubMed]
11. Wang W, Wei C. Advances in the early diagnosis of hepatocellular carcinoma. Genes Dis 2020;7:308-19. [Crossref] [PubMed]
12. Johnson P, Zhou Q, Dao DY, et al. Circulating biomarkers in the diagnosis and management of hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol 2022;19:670-81. [Crossref] [PubMed]
13. Hanif H, Ali MJ, Susheela AT, et al. Update on the applications and limitations of alpha-fetoprotein for hepatocellular carcinoma. World J Gastroenterol 2022;28:216-29. [Crossref] [PubMed]
14. Zhang J, Chen G, Zhang P, et al. The threshold of alpha-fetoprotein (AFP) for the diagnosis of hepatocellular carcinoma: A systematic review and meta-analysis. PLoS One 2020;15:e0228857. [Crossref] [PubMed]
15. Carpi S, Daniele S, de Almeida JFM, et al. Recent Advances in miRNA-Based Therapy for MASLD/MASH and MASH-Associated HCC. Int J Mol Sci 2024;25:12229. [Crossref] [PubMed]
16. Sartorius K, Sartorius B, Winkler C, et al. The biological and diagnostic role of miRNA's in hepatocellular carcinoma. Front Biosci (Landmark Ed) 2018;23:1701-20. [Crossref] [PubMed]
17. Wang L, Zhu L, Liang C, et al. Targeting N6-methyladenosine reader YTHDF1 with siRNA boosts antitumor immunity in NASH-HCC by inhibiting EZH2-IL-6 axis. J Hepatol 2023;79:1185-200. [Crossref] [PubMed]
18. Tang B, Zhu J, Wang Y, et al. Targeted xCT-mediated Ferroptosis and Protumoral Polarization of Macrophages Is Effective against HCC and Enhances the Efficacy of the Anti-PD-1/L1 Response. Adv Sci (Weinh) 2023;10:e2203973. [Crossref] [PubMed]
19. Svobodova S, Karlikova M, Topolcan O, et al. PIVKA-II as a Potential New Biomarker for Hepatocellular Carcinoma - A Pilot Study. In Vivo 2018;32:1551-4. [Crossref] [PubMed]
20. Hemken PM, Sokoll LJ, Yang X, et al. Validation of a novel model for the early detection of hepatocellular carcinoma. Clin Proteomics 2019;16:2. [Crossref] [PubMed]
21. Feng H, Li B, Li Z, et al. PIVKA-II serves as a potential biomarker that complements AFP for the diagnosis of hepatocellular carcinoma. BMC Cancer 2021;21:401. [Crossref] [PubMed]
22. Xu F, Zhang L, He W, et al. The Diagnostic Value of Serum PIVKA-II Alone or in Combination with AFP in Chinese Hepatocellular Carcinoma Patients. Dis Markers 2021;2021:8868370. [Crossref] [PubMed]
23. Liu S, Sun L, Yao L, et al. Diagnostic Performance of AFP, AFP-L3, or PIVKA-II for Hepatitis C Virus-Associated Hepatocellular Carcinoma: A Multicenter Analysis. J Clin Med 2022;11:5075.
24. Piratvisuth T, Tanwandee T, Thongsawat S, et al. Multimarker Panels for Detection of Early Stage Hepatocellular Carcinoma: A Prospective, Multicenter, Case-Control Study. Hepatol Commun 2022;6:679-91. [Crossref] [PubMed]
25. Zong J, Fan Z, Zhang Y. Serum Tumor Markers for Early Diagnosis of Primary Hepatocellular Carcinoma. J Hepatocell Carcinoma 2020;7:413-22. [Crossref] [PubMed]
26. Kim DY, Toan BN, Tan CK, et al. Utility of combining PIVKA-II and AFP in the surveillance and monitoring of hepatocellular carcinoma in the Asia-Pacific region. Clin Mol Hepatol 2023;29:277-92. [Crossref] [PubMed]
27. Zheng K, Dai L, Zhao Y, et al. Methylated SEPT9 combined with AFP and PIVKA-II is effective for the detection of HCC in high-risk population. BMC Gastroenterol 2023;23:260. [Crossref] [PubMed]
28. Zhou J, Sun H, Wang Z, et al. Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2022 Edition). Liver Cancer 2023;12:405-44. [Crossref] [PubMed]
29. Zuo N, Hu X. miR-122 Levels and Their Relationship with Serum AFP and HGF in Hepatocellular Carcinoma Patients. Sichuan Medical Journal 2013;34:1640-2.
30. Wan B, Mu QM, He W. The Effect of Serum AFP, CA19-9 and HGF Levels on Prognosis of Patients with Hepatocellular Carcinoma After Operation. Labeled Immunoassays and Clinical Medicine 2020;27:114-7.
31. Chai S. Application of Alpha-Fetoprotein Combined with Alpha-Fetoprotein-L3, Alkaline Phosphatase, and Acid Hydrolase in Hepatocellular Carcinoma. Clinical Research 2024;32:122-4.
32. Ha P, Su XC, Fu BP. Qualitative diagnosis by multi-slice spiral CT and serum AFP and VEGF detection combination in patients with intrahepatic space-occupying lesions. Journal of Practical Hepatology 2024;27:251-4.
33. Ren H, Dong J, Liu LW, et al. Risk factors for hepatocellular carcinoma in hepatitis B cirrhosis patients given treatment with human umbilical cord mesenchymal stem cells. Chinese Journal of Tissue Engineering Research 2018;22:1350-6.
34. Xu HY, Meng JL, Fu SQ, et al. Expression and clinical significance of miR-652-5p in serum of hepatocellular carcinoma patients. International Journal of Digestive Diseases 2023;43:269-74.
35. Zhang M, Li GF, Meng QH. Prediction of TACE Efficacy for Hepatocellular Carcinoma Using DCE-MRI Combined with Serum AFP, AFP-L3, and CCR3. Journal of Imaging Research and Medical Applications 2024;8:151-3.
36. Tian S, Chen Y, Zhang Y, et al. Clinical value of serum AFP and PIVKA-II for diagnosis, treatment and prognosis of hepatocellular carcinoma. J Clin Lab Anal 2023;37:e24823. [Crossref] [PubMed]
37. Perne MG, Sitar-Tăut AV, Alexescu TG, et al. Diagnostic Performance of Extrahepatic Protein Induced by Vitamin K Absence in the Hepatocellular Carcinoma: A Systematic Review and Meta-Analysis. Diagnostics (Basel) 2023;13:816. [Crossref] [PubMed]