METTL4 and METTL5 as biomarkers for recurrence-free survival in hepatocellular carcinoma patients

ABSTRACT

Background

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related deaths, with high rates of postoperative recurrence. Identifying reliable biomarkers for predicting recurrence is critical for improving patient outcomes. This study investigates the predictive value of m6A methylation-related genes, METTL4 and METTL5, on HCC recurrence after surgery.

Research Design and Methods

We analyzed METTL4 and METTL5 expression in HCC and adjacent non-cancerous tissues using the TCGA database and evaluated their levels in surgical samples from 67 hCC patients. A recurrence risk model was developed and validated in an external cohort of 65 patients.

Results

METTL4 and METTL5 were significantly overexpressed in HCC tissues. High expression correlated with shorter recurrence-free survival (RFS). The model stratified patients into high, medium, and low-risk groups with 3-year RFS rates of 18.75%, 69.70%, and 93.75%, respectively.

Conclusions

METTL4 and METTL5 expression levels are strong predictors of HCC recurrence. The risk model offers a novel approach for postoperative management of HCC.

1. Introduction

Liver cancer is one of the most prevalent malignancies worldwide, with higher incidence rates in developing countries compared to developed countries [1]. Primary liver cancer (PLC) accounts for over 50% of total liver cancer incidence [2], with hepatocellular carcinoma (HCC) being the most common type of primary liver cancer, characterized by high recurrence rates and poor prognosis [3,4]. Significant progress has been made in controlling liver cancer with advancements in early diagnosis, surgical resection of small HCCs, medical diagnostics, imaging, and non-surgical treatments [5]. However, surgical resection remains the most effective treatment for liver cancer, with a 5-year survival rate of 50–70% following early HCC resection [6]. Extensive research on HCC-specific molecular biomarkers has further revealed the underlying molecular mechanisms of HCC, which is crucial for early diagnosis, prognosis monitoring, and the development of new molecular-targeted drugs. Nonetheless, despite the roles of α-fetoprotein (AFP) and PIVKA-II in warning of HCC recurrence, some patients are not accurately covered [7]. Thus, it is imperative to investigate biomarkers capable of accurately and consistently predicting postoperative HCC recurrence.

RNA methyltransferase complexes (also known as “writers”), RNA demethylases (sometimes known as “erasers”), and N6-methyladenosine (m6A) “readers” control the process of m6A RNA methylation. A dynamic regulatory network of m6A is formed by the “writers” inscribing methylation sequences, the “readers” decoding these sequences and displaying their consequences, and the “erasers” displaying demethylation activity [8]. Previous research has documented the influence of genes linked to m6A methylation on the biological functions of HCC, including as angiogenesis, invasion, migration, and proliferation [9]. METTL3, a known m6A methyltransferase, promotes HCC tumorigenesis and metastasis both in vitro and in vivo [10]. Additionally, Wilms’ tumor 1-associating protein (WTAP) has been suggested to play a crucial role in HCC progression through m6A-HuR-dependent epigenetic silencing of the ETS proto-oncogene 1 [11]. However, the roles of METTL4 and METTL5 in HCC recurrence have not been elucidated. METTL4 is a methyltransferase that mediates nuclear m6A methylation deposition, activating the epithelial-mesenchymal transition process and accelerating tumor metastasis [12]. As one of the most well-studied 18S rRNA methyltransferases, METTL5 is crucial for the development of the embryo and is involved in the translation of proteins and the production of polysomes [13,14]. Besides, METTL5 increases protein translation in cancer, decreases apoptosis, and stimulates cell cycle progression [15]. Recent studies have found that METTL5 enhances glycolysis and cell proliferation in HCC [13]. Most current research on METTL4 and METTL5 focuses on their roles in tumor malignancy, with few studies investigating their potential as biomarkers for early diagnosis, prognosis, and recurrence prediction in clinical settings.

Therefore, this study focuses on the genes METTL4 and METTL5, exploring their relevance to liver cancer in the TCGA database, and collecting samples from 67 hCC postoperative patients from the Department of Hepatobiliary and Pancreatic Surgery, Hangzhou First People’s Hospital. We intend to explore the expression levels of METTL4 and METTL5 in cancerous and adjacent non-cancerous tissues, as well as their relationship to postoperative tumor recurrence in HCC. The purpose of this work is to indicate the alarming role of METTL4 and METTL5 in recurrence and to develop a prognostic model that would provide new insights for early prediction of HCC recurrence.

2. Materials and methods

2.1. Patients and tissue samples

The training cohort included 67 patients who underwent surgery at the Department of Hepatobiliary and Pancreatic Surgery, Hangzhou First People’s Hospital, between 12 January 201512 January 2015, and 25 August 201625 August 2016. The validation cohort comprised 65 patients who underwent surgery at the Department of Hepatobiliary and Pancreatic Surgery, Second Affiliated Hospital of Dalian Medical University, between 26 August 201726 August 2017, and 28 February 202228 February 2022. Inclusion criteria were confirmed HCC diagnosis via pathological examination of surgical specimens; complete medical records, including age, gender, AFP levels, and histological differentiation; and tissue samples suitable for immunohistochemical staining. In total, 132 patients were followed up and evaluated for survival information in this study. This study was reviewed and approved by both the Second Affiliated Hospital of Dalian Medical University Institutional Review Board and Clinical Application of Medical Technology and Scientific Research of Hangzhou First People’s Hospital, with approval number 2022–166 and KY-20211110-0103-01. All participants provided informed consent, and the study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki.

2.2. Immunohistochemistry

Standard immunohistochemistry techniques were followed. Human liver tissue slices (20 µm) were treated with the following antibodies: METTL4 (Proteintech, rabbit anti-human, 1:100) and METTL5 (Proteintech, rabbit anti-human, 1:100). After incubating with primary antibodies for 2 hours at room temperature, secondary antibodies were incubated for 1 hour. The slides were then mounted using ProLong Anti-fade reagent (Invitrogen). To account for processing variations, immunohistochemical tests were performed on all tissue sections in each group at the same time. We used ImageJ to calculate the H-score for each sample, which is based on the proportion of positively stained areas with varying intensities. Subsequently, we classified samples with a score greater than 150 as high expression and those with a score of 150 or less as low expression.

2.3. Data collection and follow-up

Clinical data obtained included patient gender, age, cirrhosis presence, AFP levels, and albumin (ALB) levels. Preoperative imaging data were used to capture tumor parameters such as maximum diameter, vascular invasion, tumor thrombus, and degree of differentiation. The critical values for tumor features and serological markers were similar with earlier literature results. Monitoring strategies and treatment regimens at both medical sites were identical to those documented in the literature. Patients were seen in outpatient clinics on a regular basis, and recurrence monitoring procedures included serum AFP, ultrasound, and computed tomography (every three months in the first year, every six months in the second year, and annually after that).

2.4. Statistical analysis

The Shapiro-Wilk test was performed to determine the normality of variables, with a p-value <0.05 indicating non-normal distribution. Normally distributed variables were expressed as mean ± standard deviation, non-normally distributed variables as median (25%-75%), and categorical variables as count (%). The key prognostic indicators were defined as follows: Recurrence-free survival (RFS) was commonly defined as the time between tumor recurrence, death from any cause, or the last follow-up. Overall survival (OS) was defined as the time from the beginning of follow-up to death from any cause. Survival curves were evaluated using the Kaplan-Meier method, and group comparisons were conducted using the log-rank test. A stepwise approach was used to identify independent prognostic factors using univariate and multivariate Cox proportional hazards regression models. A numerical risk prediction model was developed using the proportionate coefficients of independent prognostic factors from the multivariate Cox regression model. The model’s effectiveness was determined using receiver operating characteristic (ROC) curves. Finally, a nomogram was created to visually represent the Cox regression model. Statistical analyses were conducted using IBM SPSS Statistics, GraphPad Prism, and R-studio, with p-values <0.05 indicating statistical significance.

3. Results

3.1. Association of METTL4 and METTL5 levels with clinical outcomes in HCC patients

We analyzed the transcriptomic data of hepatocellular carcinoma from the TCGA database and conducted differential expression analysis of m6A methylation-related genes (Figure S2). Among these, METTL4 and METTL5 showed significant differential expression between tumor and normal tissues, with notably higher expression in tumor tissues (Figure 1(a,b)). Although a trend was observed where higher pathological T stage of the tumor corresponded to increased levels of METTL4 and METTL5 in tumor tissues, this trend did not reach statistical significance (Figure 1(c,d)). Higher levels of METTL4 (p = 0.003) and METTL5 (p = 0.003) expression, however, were substantially linked to decreased rates of progression-free survival (Figure 1(e–h)). Similarly, elevated levels of METTL4 (p = 0.023) and METTL5 (p = 0.002) related to reduced OS rates (Figure 1(f,i)). Furthermore, high expression of both METTL4 (p = 0.009) and METTL5 (p < 0.001) was also associated with lower disease-specific survival rates (Figure 1(g–j)). Furthermore, correlation analysis showed that the expressions of METTL4 and METTL5 in patients with HCC correlated positively (r = 0.467, p < 0.001) (Figure 1(k)).

Figure 1.

Elevated METTL4, METTL5 expression is linked to worse survival of patients with HCC. (a) METTL4 expression in normal tissues and tumor tissues from TCGA database; (b) METTL5 expression in normal tissues and tumor tissues from TCGA database; (c) association of METTL4 levels with tumor pathologic T stage from TCGA database; (d) association of METTL5 levels with tumor pathologic T stage from TCGA database; (e) analysis of METTL4 and progression-free interval of patients in the TCGA database; (f) analysis of METTL4 and OS of patients in the TCGA database; (g) analysis of METTL4 and disease-specific survival of patients in the TCGA database; (H) analysis of METTL5 and progression-free interval of patients in the TCGA database; (i) analysis of METTL5 and OS of patients in the TCGA database; (j) analysis of METTL5 and disease-specific survival of patients in the TCGA database; (k) analysis of the correlation between METTL4 and METTL5 from TCGA database.

3.2. Summary of clinical information for 132 hCC patients

The baseline characteristics of 132 patients with complete clinical data from two medical institutions are summarized in Table 1. The training set comprised 67 hCC patients, including 58 males and 9 females, while the validation set consisted of 65 hCC patients, including 49 males and 16 females, with an average age of 54 and 59 years, respectively. In the training and validation groups, cirrhosis was present in 44.78% and 64.62% of patients, respectively. Additionally, 40.30% in the training group and 13.85% in the validation group had AFP levels above 400 ng/ml. Vascular invasion was observed in 31.34% of the training group and 58.46% of the validation group. Moreover, 43.28% of patients in the training cohort and 15.38% of patients in the training cohort presented with tumor thrombus.

Table 1.

Demographics and tumor characteristics.

	Training set (n = 67)	Validation set (n = 65)
Receptor gender
Male	58(86.57%)	49(75.38%)
Female	9(13.43%)	16(24.62%)
Recipient age (years)	54 ± 12	59 ± 9
Liver cirrhosis	30 (44.78%)	42 (64.62%)
AFP* (ng/ml)
<400	37 (55.22%)	56 (86.15%)
＞400	27 (40.30%)	9 (13.85%)
ALB (mg/dL)	43.9 ± 4.0	42.9 ± 4.3
Maximum diameter (cm)	8.3 (6.2–10)	3.5 (2–5.55)
Vascular invasion	21 (31.34%)	38 (58.46%)
Tumor thrombus	29 (43.28%)	10 (15.38%)
Differential grade
Low	28 (41.79%)	16 (24.62%)
High	39 (58.21%)	49 (75.38%)
RFS rate	26.87%	36.92%
RFS-D (days)	255 (66–1258)	1029 (441–1235)

*The total number of training set samples is less than 67 due to partial data loss.

Abbreviations: AFP, α-fetoprotein; ALB, Albumin; RFS, Recurrence-Free Survival; RFS- D, Recurrence-Free Survival-Death.

3.3. Expression of METTL4 and METTL5 in tumor tissues is associated with HCC recurrence

To investigate the relationship between the expression levels of METTL4 and METTL5 and patient prognosis in HCC, we analyzed the expression levels of METTL4 and METTL5 in both cancerous and adjacent non-cancerous tissues from various patients. Expression levels of METTL4 and METTL5 were visually represented, with low and high expressions of METTL4 depicted in Figure 2(a,b), respectively, and for METTL5 in Figure 2(c,d). Based on the analysis from ImageJ, the average H-score for the METTL4 high expression group (n = 48) and low expression group (n = 19) was 220.4 ± 45.3 and 97.3 ± 35.9, respectively, while the average H-score for the METTL5 high expression group (n = 37) and low expression group (n = 30) was 214.7 ± 51.4 and 103.6 ± 32.5, respectively (Table S2). Kaplan-Meier survival analysis indicated that higher expression levels of METTL4 (p = 0.0016) and METTL5 (p = 0.0035) in cancer tissues were significantly associated with lower RFS rates (Figure 2(e,f)). Additionally, an elevated expression of METTL4 in cancer tissues was also correlated with lower OS rates (p = 0.0371) (Figure 2(g)), while the expression of METTL5 in cancer tissues did not show a significant association with patient OS (Figure 2(h)). Interestingly, high expression of METTL4 in adjacent non-cancerous tissues was associated with poorer RFS (p = 0.0101), whereas the expression of METTL5 in adjacent non-cancerous tissues did not demonstrate a significant relationship with RFS. Furthermore, when comparing the relative expression levels of METTL4 and METTL5 in cancer tissues to those in adjacent non-cancerous tissues, no correlation with RFS and OS was observed.

Figure 2.

Expression of METTL4 and METTL5 in tumor and adjacent non-tumorous tissues and their correlation with RFS and OS. (a) Immunohistochemical staining image of low expression of METTL4 protein in cancer tissue; (b) Immunohistochemical staining image of high expression of METTL4 protein in cancer tissue; (c) Immunohistochemical staining image of low expression of METTL5 protein in cancer tissue; (d) Immunohistochemical staining image of high expression of METTL5 protein in cancer tissue; (e) Kaplan – Meier analysis for RFS in HCC patients with high or low METTL4 expression; (f) Kaplan – Meier analysis for RFS in HCC patients with high or low METTL5 expression; (g) Kaplan – Meier analysis for OS in HCC patients with high or low METTL4 expression; (h) Kaplan – Meier analysis for OS in HCC patients with high or low METTL5 expression.

3.4. METTL4 and METTL5 in tumor tissues as independent prognostic factors for HCC recurrence

For figuring out the risk of HCC recurrence, we used Cox proportional hazards regression analysis. A univariate Cox regression analysis of 67 hCC patients in the training cohort found that METTL4 (HR = 3.073, 95% CI 1.487–6.352, p = 0.002) and METTL5 (HR = 2.405, 95% CI 1.312–4.41, p = 0.005) expression levels were substantially linked with higher tumor recurrence risk. Furthermore, maximal tumor diameter, vascular infiltration, and pathological differentiation degree were all strongly linked with HCC recurrence risk (Table 2). A subsequent multivariate Cox regression analysis revealed that METTL4 (HR = 2.743, 95% CI 1.174–6.41, p = 0.02), METTL5 (HR = 2.495, 95% CI 1.217–5.115, p = 0.013), and maximal nodule diameter (HR = 1.182, 95% CI 1.082–1.291, p = 0.001) were independent predictive variables for recurrence risk. In comparison to the low METTL4 group, patients with high METTL4 levels had a higher proportion of AFP ≥400 ng/mL (p = 0.026) and bigger nodule diameters (p = 0.037). More information is available in Table S1.

Table 2.

Univariate and multivariate cox analyses of variables related to RFS.

	Univariate analysis			Multivariate analysis
	p value	HR	95.0%CI	p value	HR	95.0%CI	β coefficient	points
METTL4-CA-expression	0.002	3.073	1.487–6.352	0.02	2.743	1.174–6.41	1.009	6
METTL5-CA-expression	0.005	2.405	1.312–4.41	0.013	2.495	1.217–5.115	0.914	5
Maximum diameter (cm)	0	1.16	1.083–1.243	0.001	1.182	1.082–1.291	0.167	1
Vascular invasion	0.013	2.108	1.172–3.791
Differential grade	0.031	0.536	0.303–0.946

Abbreviations: AFP, α-fetoprotein; CI, confidence intervals; HR, hazard ratio.

3.5. Predictive model based on METTL4 and METTL5

Next, we developed a simpler scoring system based on the independent prognostic factors identified by Cox regression analysis – METTL4 level, METTL5 level, and maximal tumor diameter in cancer tissues. Patients were awarded scores based on the relative weights of these prognostic factors, with low, medium, and high-risk groups. The low-risk group had superior RFS and OS than the other two groups (Figure 3(a,b)). The 3-year RFS rates for the low, medium, and high-risk groups were 70.59%, 31.51%, and 0%, respectively (p ≤0.0001). The ROC curve examination confirmed the model’s effectiveness (Figure 3(c,d)).

Figure 3.

The establishment of the risk score model in predicting postoperative tumor recurrence in the training set. (a) Risk score-based RFS survival analysis in patients; (b) Risk score-based OS survival analysis in patients; (c) ROC curve and AUC of RFS in the training cohort; (d) ROC curves and AUC of OS in the training cohort.

3.6. Validation of the risk score model in an independent external HCC cohort

In an external validation cohort of 65 hCC patients, the expression of METTL4 and METTL5 in cancer tissues was likewise strongly linked with patient RFS, with greater expression levels corresponding to lower RFS rates (Figure 4(a,b)). In this validation cohort, our model performed well, with 65 hCC patients categorized into low, medium, and high-risk groups using the aforementioned scoring method. The low-risk group had superior OS and RFS compared to the medium and high-risk groups. The 3-year RFS rates for low, medium, and high-risk groups were 93.75%, 69.70%, and 18.75%, respectively (p < 0.0001, Figure 4(c)). The model once again displayed outstanding predictive ability, as measured by the ROC curve, with an area under the curve (AUC) value of 0.8613 (Figure 4(d)).

Figure 4.

Performance of the risk score model in the validation cohort. (a) Kaplan – Meier analysis for RFS in HCC patients with high or low METTL4 expression in the validation cohort; (b) Kaplan – Meier analysis for RFS in HCC patients with high or low METTL5 expression in the validation cohort; (c) Kaplan – Meier analysis for RFS after risk scoring of the validation cohort; (d) ROC curves and AUC of RFS in the validation cohort.

3.7. Construction and validation of the predictive nomogram

Based on the independent prognostic markers for liver cancer recurrence, we developed a nomogram (Figure 5(a)) to predict the likelihood of post-surgical tumor recurrence in HCC patients. The training cohort’s 1-, 2-, and 3-year AUC curve values were 0.835, 0.834, and 0.881, respectively (Figure 5(b)). In the validation cohort, the 1-, 2-, and 3-year AUC values of the ROC curve were 0.859, 0.846, and 0.827, respectively (Figure 5(c)), demonstrating our model’s high predictive capacity. Based on the calibration curves for the training and validation cohorts (Figure 5(d,e)), we were able to make more accurate predictions about 1-year, 2-year, and 3-year liver cancer recurrences. Decision curve analysis (DCA) indicated favorable net clinical benefit (Figure S1). Collectively, these findings affirm the nomogram’s robust discriminatory and calibration capabilities.

Figure 5.

Nomogram based on the independent prognostic factors. (a) A constructed nomogram for prognostic prediction of postoperative HCC patients; (b) 1-, 2-, and 3-year ROC curves and AUC of RFS in the training set; (c) 1-, 2-, and 3-year ROC curves and AUC of RFS in the validation set; (d) The calibration curves for the training set; (e) The calibration curves for the validation set.

4. Discussion

Although early-stage HCC can be radically treated through surgical resection, liver transplantation, and ablation, the high recurrence and mortality rates persist. The high mortality rate of HCC is mainly due to tumor metastasis and recurrence [16]. Despite the ongoing development and application of various targeted anti-tumor drugs, such as sorafenib, lenvatinib, and ramucirumab, a significant proportion of patients still experience tumor recurrence. As a result, it is critical to investigate biomarkers that can accurately predict HCC recurrence. In this work, we found two m6A-related genes that can help predict HCC recurrence and created a tumor recurrence risk prediction model based on m6A methylation-related genes to accurately predict early HCC recurrence and guide early therapeutic interventions.

m6A methylation-related genes are critical in the proliferation and metastasis of various tumors. In the tumor immune microenvironment of HCC, the methylation and demethylation of key RNAs can alter the expression of immune-related proteins, aiding in tumor immune evasion [17]. m6A methylation-related genes could potentially serve as predictors of HCC recurrence. Previous studies have found that m6A-related genes, such as METTL16 and ALKBH5, are associated with the prognosis of HCC patients [18]. However, the correlation between METTL4, METTL5, and HCC recurrence remains to be explored.

In this study, we used bioinformatics to reveal that METTL4 and METTL5 expression is substantially linked with HCC patient prognosis. While METTL4 and METTL5 May 5 May not exhibit the most extreme expression changes in HCC, their potential regulatory roles in tumorigenesis and their interactions with other m6A-related factors offer a unique angle for investigation. This study aims to elucidate their specific contributions to HCC prognosis, which has been underexplored in existing research. METTL4, along with METTL3 and WTAP, forms a highly conserved mRNA methyltransferase complex, and current research suggests that METTL4 plays a crucial role in tumors [19]. Shen et al. [20]found that in cervical squamous cell carcinoma, METTL4 enhances the tumor epithelial-mesenchymal transition process by promoting the expression of related lncRNA and downstream miRNA. METTL5, in contrast to METTL4, exhibits metabolic stability by forming a heterodimer complex with its coactivator TRMT112 and is involved in the m6A modification of 18S rRNA [21]. The absence of METTL5 in cardiomyocytes leads to hypertrophy [22], and recent reports indicate that METTL5 regulates the initiation of translation in breast and lung adenocarcinomas and has immune-related functions [23,24]. Therefore, we hypothesize that METTL5 might also increase the probability of HCC recurrence after surgery by regulating the initiation of translation in HCC. This relationship needs further verification, and the specific mechanisms by which METTL4 and METTL5 influence post-surgical HCC recurrence require further study. Interestingly, our comparison of METTL4 and METTL5 expression revealed consistency between the two genes, suggesting that METTL4 and METTL5 May 5 May function as paired genes, mutually influencing each other and jointly affecting the prognosis of HCC patients.

Regardless of tumor aggressiveness, increased tumor diameter is associated with a worse patient prognosis [25,26]. A total tumor diameter higher than 8 cm, pre-transplant AFP levels greater than 1000 ng/mL, and intrahepatic tumors with a maximum standard uptake value greater than 5 are independent risk factors for HCC recurrence and overall survival after liver transplantation in Vp2–3 Portal Vein Tumor Thrombus patients [27]. The positive correlation between HCC tumor diameter and adverse patient outcomes suggests that tumor diameter could also be an indicator of HCC recurrence post-surgery.

In our investigation of 67 hCC patients, immunohistochemical staining of HCC and surrounding non-cancerous tissues agreed with bioinformatics analysis. Tumor tissues showed high levels of METTL4 and METTL5. Further study of clinical data revealed that METTL4 was associated with tumor size, AFP, and overall survival. Both METTL4 and METTL5 were associated with tumor recurrence. Thus, we believe that METTL4 and METTL5 have potential predictive value for tumor recurrence, and their combination can improve the accuracy of prediction. Multivariate Cox regression analysis revealed that METTL4, METTL5, and maximal tumor diameter were independent predictors of tumor recurrence. Based on these findings, we rated patients’ tumor recurrence risk and divided them into high, medium, and low risk groups. There were substantial variations in recurrence rates across the three groups, with lower scores indicating reduced recurrence probabilities. The ROC curve of this model showed good performance, with high specificity and sensitivity. Compared to single diagnostic indicators, our model is more reliable. Immunohistochemical results from HCC tissues of 65 patients from another center showed similar outcomes and validated our model for predicting tumor recurrence risk, demonstrating its reliable predictive efficacy.

In this research, we established a risk prediction model for HCC recurrence based on m6A methylation-related genes and validated it in an external cohort. Our model incorporates METTL4, METTL5, and maximum tumor diameter, providing high accuracy and reliable indicators for predicting early HCC recurrence, aiming to guide early clinical interventions and improve the five-year survival rate of patients. However, some limitations of this study must be addressed. First, the sample size of the included cohorts was limited, with 67 samples in the training set and 65 samples in the validation set, necessitating larger sample sizes in future studies to further verify the reliability of the research. Secondly, the specific mechanisms by which METTL4 and METTL5 influence HCC recurrence were not investigated. Future research should focus on these mechanisms. Additionally, differences in HCC surgical methods, clinical management, and measurement methods across the two centers might introduce selection bias, requiring the inclusion of more medical centers to avoid such bias.

Overall, the study shows that METTL4 and METTL5 expression levels in cancer tissues of hepatocellular carcinoma patients serve as excellent postoperative recurrence predictors that are correlated. The risk assessment model that includes METTL4 and METTL5 can stratify patients depending on their recurrence risk, further optimizing postoperative recurrence therapy in patients with hepatocellular carcinoma.

5. Conclusion

Our study identifies METTL4 and METTL5 as significant biomarkers for predicting postoperative recurrence in hepatocellular carcinoma (HCC) patients. The risk model developed, which incorporates METTL4, METTL5, and tumor diameter, effectively stratifies patients by recurrence risk, providing a valuable tool for guiding early clinical interventions and improving long-term survival rates. These findings highlight the potential of m6A methylation-related genes in HCC management. However, further studies with larger cohorts are necessary to confirm these results and extend the model’s applicability across diverse clinical settings.

Funding Statement

This work was funded from the Zhejiang Provincial Natural Science Foundation Public Welfare Project [TGY24H160072], Zhejiang Provincial Health Commission, Zhejiang Health Science and Technology Plan Project [2022KY245].

Article highlights

• METTL4 and METTL5, two m6A methylation-related genes, are identified as key predictors of hepatocellular carcinoma (HCC) recurrence after surgery.

• Bioinformatics analysis from the TCGA database revealed higher expression levels of METTL4 and METTL5 in HCC tissues compared to adjacent non-cancerous tissues, correlating with poorer patient prognosis.

• Immunohistochemical analysis of 67 hCC patients confirmed the high expression of METTL4 and METTL5 in tumor tissues, which was associated with shorter recurrence-free survival.

• Multivariate Cox regression analysis identified METTL4, METTL5, and maximum tumor diameter as independent predictors of HCC recurrence.

• A predictive model based on METTL4, METTL5, and tumor diameter was developed, effectively classifying patients into high, medium, and low recurrence risk groups.

• The risk model demonstrated high specificity and sensitivity, outperforming individual biomarkers in predicting early HCC recurrence.

• Validation in an external cohort of 65 patients confirmed the model’s reliability and predictive accuracy.

• The findings underscore the clinical utility of METTL4 and METTL5 as biomarkers for HCC management, offering potential for improved postoperative treatment strategies.

• Future research should focus on the mechanisms by which METTL4 and METTL5 influence HCC recurrence and validate the model across diverse clinical settings.

• The study suggests the need for larger sample sizes in future research to confirm the generalizability of the findings and refine the risk prediction model.

Abbreviations

AFP: α-Fetoprotein, ALB: Albumin, AUC: Area Under the Curve, CI: Confidence Interval, DCA: Decision Curve Analysis, HCC: Hepatocellular Carcinoma, m6A: N6-Methyladenosine, OS: Overall Survival, PLC: Primary Liver Cancer, RFS: Recurrence-Free Survival, ROC: Receiver Operating Characteristic, WTAP: Wilms Tumor 1 Associated Protein.

Author contributions

Jialing Zhao: Data curation, Formal analysis, Investigation, Software, Validation, Visualization, Writing-original draft. Ruiqi Sun: Data curation, Formal analysis, Methodology, Visualization, Writing-original draft, Writing-review & editing. Liang Zhi: Methodology, Supervision, Writing-review & editing. Danjing Guo: Data curation, Software, Validation, Visualization. Sunbin Ling: Methodology, Resources. Xiangnan Liang: Resources. Jianhui Li: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Writing-review & editing. Changku Jia: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Supervision, Writing-review & editing.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Ethical declaration

This study was reviewed and approved by both the Second Affiliated Hospital of Dalian Medical University Institutional Review Board and Clinical Application of Medical Technology and Scientific Research of Hangzhou First People’s Hospital, with approval number 2022–166 and KY-20211110-0103-01. All participants provided written informed consent, and the study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14796694.2024.2442296