A novel miRNA-based model for predicting the 3-year recurrence risk of hepatocellular carcinoma following liver transplantation

Introduction

Liver transplantation (LT) is the optimal curative treatment for patients with early-stage hepatocellular carcinoma (HCC) in theory, but not all the HCC patients receiving LT could be completely free from tumor recurrence, the strongest survival-limiting factor (1). Particularly, the inevitable immunosuppression therapy following LT significantly increases the risk of HCC recurrence in liver recipients (1). In the era of organ shortage, more attention has been emphasized on patient selection to reduce tumor recurrence risk following LT, thereby maximizing the utilization of available organs and promoting the overall prognosis of HCC patients (2). Based on the most recognized patient selection criteria, namely Milan criteria (MC) (single tumor ≤5 cm, or up to three tumors each ≤3 cm, and no vascular invasion and extrahepatic spread) established in 1996, HCC patients receiving LT could achieve an excellent 5-year overall survival (about 75–80%) and a relatively low recurrence rate (about 10–15%) (3). However, MC has been criticized for being excessively stringent and being inexact in reflecting the tumor biology and prognosis in HCC, which may inaccurately enroll some patients who meet MC but have a high risk of recurrence or exclude certain patients who fall outside MC but potentially benefit from LT (4). Thus, it is time to reconsider MC to accurately evaluate tumor burden and precisely identify HCC patients with a high risk of recurrence following LT.

The consensus is that optimization or expansion of MC should not be at the cost of rising tumor recurrence rates. Currently, there are two directions for expanding MC. In contrast with reconsidering the threshold of certain clinical indicators (such as tumor number and diameter), incorporating tumor biological characteristics into patient selection criteria for LT may be the most promising way in the era of precision medicine (5,6). MicroRNAs (miRNAs), a class of small noncoding RNAs, have become one of the most valuable biomarkers for cancer diagnosis and prognosis (including HCC), due to their important biological roles in cancer initiation and progression via post-transcriptionally modulating gene expression (7). Up to now, some recurrence-related miRNAs following LT have been identified in previous relevant studies, including miR-122, miR-126, miR-147, miR-15a, miR-19a, miR-24, miR-223, miR-22, miR-30a and miR-886 (8-10). However, their prognostic accuracy in terms of HCC recurrence following LT is still far from satisfactory, which may result from imperfect study design or statistical methods in the existing studies from the current perspective.

It has been estimated that HCC recurrence is prevalent in the first 2 years (nearly 70%) following LT and gradually moves towards relaxation after 3 years post-LT (11). This may be attributed to that HCC recurrence following LT is mostly caused by the growth of existing occult metastases (12). Of note, it has been reported that HCC patients with early recurrence have worse overall survival following LT (13). Hence, in this study, by making full use of the public GSE30297 dataset in a reasonable way, we aimed to identify more reliable miRNAs that reflected the risk of HCC recurrence within 3 years following LT and to establish an excellent multiple miRNA-based prognostic model. Besides, a meaningful combined model integrating the miRNA-based model and MC was also developed. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-9/rc).

Methods

Data preparation and ethical statement

GSE30297 dataset, shared by Barry et al. in the Gene Expression Omnibus (GEO), provided the normalized miRNA profiles of HCC tissues taken from 69 HCC patients receiving LT, as well as their important clinical characteristics (including MC and recurrence status within 3 years post-LT) (10). According to the statement of data quality evaluation in the original article by Barry et al., the high-quality miRNA profiles of 64 patients were finally enrolled in our study. Among them, 37 patients developed recurrent HCC within 3 years after LT (namely the recurrence group), while 27 patients had no recurrent HCC within 3 years (namely the non-recurrence group). In the process of model construction based on the GSE30297 dataset, a Z-score transformation of expression data was performed for each selected miRNA. Besides, the miRNA-seq data, mRNA-seq data, and complete clinical data [including tumor-node-metastasis (TNM) stage and follow-up data] of 141 HCC patients receiving R0 resection were downloaded from The Cancer Genome Atlas (TCGA)-liver hepatocellular carcinoma (LIHC) dataset. The sequencing data of 50 normal liver tissues (NT) in the TCGA-LIHC dataset were also retrieved as controls. The raw miRNA-seq or mRNA-seq data were preprocessed by the R package “edgeR” for normalization and further log2-transformed. The above public datasets were enrolled in compliance with the ethical standards of GEO and TCGA databases, additional ethical approval was not required. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Recurrence-related miRNAs screening

The differentially expressed miRNAs (DERs) in the recurrence group compared with the non-recurrence group were firstly identified using the GSE30297 dataset. The DERs between HCC tissues and NT were also explored using the TCGA-LIHC dataset. Then, by utilizing overlapping analysis, the miRNAs that were dysregulated in HCC and associated with the 3-year recurrence of HCC following LT were discovered. For validating the prognostic values of selected DERs in HCC, survival analyses were performed to see if the dysregulation of miRNAs was also related to recurrence-free survival (RFS) in HCC patients receiving R0 resection (TCGA-LIHC cohort). The R package “limma” was applied for differential expression analysis. “Fold change (FC) >1.1 with P value <0.05” was deemed as the threshold. Survival analysis was conducted by Kaplan-Meier (KM) methods. For grouping the HCC patients into two groups according to the DERs’ expression, X-tile 3.6.1 software was applied for choosing the best cutoff value.

Model construction and validation

Based on the GSE30297 dataset, least absolute shrinkage and selection operator (LASSO)-logistic regression was carried out in the 64 HCC patients receiving LT using the R package “glmnet”. The best miRNA set among the DERs for model construction was identified according to 10-fold cross-validation and lambda.min. Then, multivariate logistic regression was performed to calculate the coefficients (β values) of the selected miRNAs. The formula was as follows: risk score = β₁ * Z-score of miRNA₁ + β₂ * Z-score of miRNA₂ + ... + β_n * Z-score of miRNA_n. Besides, by using the R package “rms”, a nomogram integrating the miRNA-based model with MC was developed. The predictive ability of the model was effectively evaluated based on comprehensive methods, including decision curve analysis (DCA), receiver operating characteristic (ROC) curve analysis, calibration plot, and concordance index (C-index). The ROC curve analysis and DCA were conducted by using the R package “pROC” and “rmda” separately.

miRNA target gene prediction and enrichment analysis

DIANA-microT-CDS, an online platform for miRNA target prediction, was used for predicting the potential target genes of selected miRNAs (http://diana.imis.athena-innovation.gr/DianaTools/). Of note, the results obtained in the DIANA-microT-CDS database were also validated by another popular database—Targetscan (http://www.targetscan.org). In addition, the dysregulated mRNAs in HCC tissues compared with NT were screened out. Finally, target genes of the core miRNAs were determined via overlapping analysis based on the negative regulatory relationship between miRNAs and mRNAs.

DAVID 6.7, an online functional annotation database, was utilized for Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (https://david-d.ncifcrf.gov/). The GO analysis included three aspects: cellular component (CC), molecular function (MF), and biological process (BP). Relevant results were present in enrichment plots generated by using SangerBox (http://sangerbox.com/).

Statistical analysis

Continuous variables between two groups were compared by Student’s t-test or Mann-Whitney test. The area under the curve (AUC) was applied to evaluate the predictive accuracy of models. A logistic regression model was carried out for analyzing the risk indicators related to HCC recurrence within 3 years after LT. GraphPad Prism 8.0.2, R 3.4.2, and SPSS 22.0 were used for graphing and statistical analysis. A P value <0.05 was regarded as statistically significant.

Results

Identification of recurrence-related miRNAs following LT

The study design was generally described in Figure 1. According to the screening strategy, 21 miRNAs that were dysregulated in HCC and potentially related to HCC recurrence within 3 years after LT were firstly screened out (Figure 2A,2B). By validating their prognostic values in the TCGA-LIHC cohort, 10 key miRNAs that were both related to HCC recurrence following LT or R0 resection were finally identified, including miR-454, miR-1293, miR-22, miR-23b, miR-130a, miR-139, miR-150, miR-152, miR-99a, and miR-199b (Figure 2C,2D). The clinical significances of the 10 miRNAs in HCC were preliminarily analyzed (Table S1). The relative expression levels of the 10 miRNAs in the recurrence group and the non-recurrence group were briefly displayed in the heat map, along with the cluster analysis (Figure 3).

Figure 1 The flowchart describing the overall study design. HCC, hepatocellular carcinoma; LASSO, least absolute shrinkage and selection operator; LT, liver transplantation; miRNAs, microRNAs; NT, normal liver tissues; ROC, receiver operating characteristic; TCGA-LIHC, The Cancer Genome Atlas-liver hepatocellular carcinoma.

Figure 2 Identification of miRNAs related to the recurrence of HCC following LT. (A) Volcano plot representing the screening of DERs in the recurrence group compared with the non-recurrence group, and the screening of DERs in HCC tissues compared with NT. Green dots: downregulated miRNAs; red dots: upregulated miRNAs; black dots: no significance. (B) Identification of the miRNAs that were dysregulated in HCC and potentially related to HCC recurrence within 3 years after LT by overlapping analyses. Green area: upregulated miRNAs in the recurrence group compared with the non-recurrence group; red area: downregulated miRNAs in the recurrence group compared with the non-recurrence group; blue area: upregulated miRNAs in HCC tissues compared with NT; yellow area: downregulated miRNAs in HCC tissues compared with NT. (C,D) By validating their prognostic values in the TCGA-LIHC cohort, 10 core miRNAs that were related to HCC recurrence following LT or R0 resection were identified. **, P<0.01; ***, P<0.001. DERs, differentially expressed miRNAs; HCC, hepatocellular carcinoma; LT, liver transplantation; miRNAs, microRNAs; NT, normal liver tissues; TCGA-LIHC, The Cancer Genome Atlas-liver hepatocellular carcinoma.

Figure 3 Heat map displaying the relative expression levels of the 10 miRNAs in different groups. miRNAs, microRNAs.

Construction and validation of a miRNA-based risk score

Based on the GSE30297 dataset, LASSO regression analysis suggested that the optimal miRNA set for modeling contained five core miRNAs (including miR-454, miR-1293, miR-139, miR-99a, and miR-130a) (Figure 4). The correlations between the five miRNAs expressions and HCC recurrence after LT were verified by univariate logistic regression analysis (all P<0.05). According to the results of multivariate logistic regression, the five miRNA-based risk score was constructed based on the GSE30297 dataset as follows: risk score = 1.232 * Z-score of miR-454 + 1.259 * Z-score of miR-1293 − 1.004 * Z-score of miR-139 − 0.493 * Z-score of miR-99a − 0.449 * Z-score of miR-130a (Table 1).

Figure 4 Selection of the optimal miRNAs for model construction by LASSO-penalized logistic regression model. (A) The optimal miRNA set was chosen by 10-fold cross-validation and lambda.min. (B) LASSO coefficient profile of the miRNAs. LASSO, least absolute shrinkage and selection operator; miRNAs, microRNAs.

As shown in Figure 5, the risk score of the recurrence group (n=37) was significantly higher than that of the non-recurrence group (n=27) (P<0.001). ROC analyses indicated that the risk score had a good prognostic value for HCC recurrence within 3 years following LT (AUC =0.901, P<0.001), which was obviously better than that of MC (AUC =0.629, P=0.08) or that of the five miRNAs alone (miR-454: AUC =0.727, P=0.002; miR-1293: AUC =0.705, P=0.005; miR-139: AUC =0.717, P=0.003; miR-99a: AUC =0.708, P=0.005; miR-130a: AUC =0.716, P=0.003).

Figure 5 Identification of the predictive value of the miRNA-based risk score. (A) Comparison of the risk score between the recurrence group and the non-recurrence group. (B) ROC curves exhibiting the predictive abilities of the miRNA-based risk score, each miRNA and Milan criteria. ***, P<0.001. AUC, area under the curve; miRNA, microRNA; ROC, receiver operating characteristic.

Establishment and validation of a combined model

By performing logistic regression analyses, the miRNA-based risk score and MC were recognized as the risk factors of HCC recurrence within 3 years after LT, and their prognostic values were independent of each other (Table 2). Therefore, a combined model integrating the risk score and MC was established, and a nomogram was correspondingly generated for visualization (Figure 6A).

Figure 6 Establishment and validation of a combined predictive model. (A) The nomogram of the combined model integrating the miRNA-based risk score with Milan criteria. (B-D) The calibration curve, the ROC curves and the decision curve analyses of the nomogram for predicting the 3-year recurrence risk of hepatocellular carcinoma following liver transplantation. AUC, area under the curve; HCC, hepatocellular carcinoma; miRNA, microRNA; ROC, receiver operating characteristic.

Calibration plots displayed excellent consistency between the actual probability of HCC recurrence within 3 years after LT and the nomogram-predicted risk of HCC recurrence (Figure 6B). The C-index (equal to the AUC of ROC curves) of the combined predictive model for HCC recurrence was 0.926 [95% confidence interval (CI): 0.866–0.986], which was remarkably better than that of MC [C-index, 0.629 (95% CI: 0.509–0.749)] (P<0.001) (Figure 6C). Besides, for visually evaluating its clinical application value, DCA curves showed that the combined model had a superior net benefit than that of MC (Figure 6D).

Potential biological roles of the core miRNAs in HCC

A total of 1,457 genes dysregulated in HCC were predicted as the miRNA targets, including 881 upregulated genes and 576 downregulated genes (Table S2). The enrichment analyses suggested that the miRNA targets were closely involved in many pathophysiological processes related to the development and progression of cancer (Figure 7).

Figure 7 Enrichment analyses for exploring the potential biological roles of the target genes of the five miRNAs. (A-C) GO analysis, including BP, MF and CC. (D) KEGG pathway analysis. BP, biological process; CC, cellular component; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; miRNAs, microRNAs.

Discussion

Along with the rising incidence of HCC and the high-proportioned patients with unresectable HCC, there is an enormous demand for LT as radical therapy among HCC patients (14). However, there should be a balance between the expansion of MC for enlarging LT candidates and the potential risk of tumor recurrence to ensure the reasonable allocation of scarce organs (1). Given that most recurrent cases occur in the early years post-LT, which is closely associated with occult metastases of the primary tumor, incorporating biomarkers (particularly miRNAs) reflecting tumor biology into existing clinical selection criteria for LT is undoubtedly of great significance (11,15). Compared with the existing relevant studies, this study has the following features (9,10,16,17). First, to ensure the involvement of the selected miRNAs in the progression of HCC, we introduced the TCGA-LIHC dataset to identify the miRNAs dysregulated in HCC tissues compared with normal controls, rather than directly screening out the miRNAs mathematically associated with HCC recurrence following LT. In addition, we also made full use of the cohort of patients with early-stage HCC who underwent R0 resection to validate the prognostic values of the selected miRNAs to some extent. Besides, in contrast with the simple utilization of a single indicator or the unreasonable choice of multiple miRNAs for modeling in previous studies, a relatively rational statistical approach (LASSO regression model) was applied in this study to choose the optimal miRNAs for establishing the multiple miRNA-based risk score. Furthermore, after identifying the miRNA-based model and MC were both related to the 3-year risk of HCC recurrence after LT, a genomic-clinical model was constructed to strengthen the predictive power. Importantly, to enhance its clinical applicability, a visual nomogram of the combined model was accordingly established and exhibited high predictive accuracy.

Based on the improved study design mentioned above, the five miRNA-based risk score was developed and showed a good efficacy (AUC =0.901) for predicting the risk of HCC recurrence within 3 years following LT, and the predictive accuracy was further improved (AUC =0.926) by integrating the miRNA-based model with MC. In light of the good performance of our model, it could offer necessary guidance for pretransplant or posttransplant decision-making in LT for HCC patients. On the one hand, among the HCC patients receiving preoperative tissue biopsy, the genomic-clinical model makes it possible to assess tumor biological features more comprehensively and precisely before LT in comparison with classical morphologic criteria (like MC), thereby selecting more suitable HCC patients for LT in the context of liver shortage (1,2,5). The patients outside MC but exhibiting favorable tumor biology could be selected as candidates for LT, while the patients within MC but showing a high risk of HCC recurrence post-LT need more rigorous considerations (4,15). For example, it has been reported that allografts suffering severe ischemia-reperfusion injury provide a special environment (including ongoing inflammation, tissue hypoxia, and microvascular dysfunction) contributing to tumor cell reseeding and growth (1,18). Thus, choosing the proper donor type (like standard donor) and advanced organ preservation or repair technique (such as hypothermic oxygenated perfusion) to ensure the quality of allografts may reduce the risk of HCC recurrence for certain patients (1,2,18). In addition, experienced operative skills in avoiding intraoperative tumor spread, as well as reasonable posttransplant management, are also crucial for reducing the probability of HCC recurrence (19). On the other hand, given that liver biopsy is not routinely performed before LT in clinical practice, the utility of the miRNA-based model after LT may be of more practical significance at present (1,19). Once the patient is classified as an individual with a high risk of HCC recurrence following LT, several therapeutic or preventive interventions could be timely taken into consideration for the specific patients, such as the tailored immunosuppressive regimen, the close monitoring strategy, and new adjuvant therapies (1,2).

It has been reported that the five key miRNAs act as either oncogene or antioncogene closely involved in the progression of cancers (including HCC). First, the dysregulation of the five miRNAs (including miR-454, miR-139, miR-99a, miR-130a, and miR-1293) in HCC tissues have been consistently identified in previous studies, which are closely associated with the prognosis of HCC patients (20-24). In addition, compared with tissue biopsy, the benefits of liquid biopsy (particularly noninvasive operation) have been recognized (25). Strikingly, dysregulation of circulating miR-130a/miR-139/miR-99a in HCC patients has been revealed, which suggests that they may serve as a promising noninvasive biomarker in terms of the diagnosis and prognosis of HCC (22,26,27). It is a pity that the alterations of the other two miRNAs (including miR-454 and miR-1293) in the serum of HCC patients remain unknown. Thus, it is worthy of identifying the prognostic values of the five circulating miRNAs in HCC patients receiving LT in the future. Besides, up to now, to elucidate the biological roles of the five miRNAs in cancers (including HCC), some of their target genes (such as SOCS6, CHD5, GDF10, AGO2, and SMAD4) and competitive endogenous RNAs (such as lncRNA-ZFPM2-AS1, Circ-0001175, and LINC00667) have been revealed (28-30). For a better understanding of the biological functions of the five miRNAs in HCC, a total of 1,457 target genes were comprehensively predicted in this study, and the enrichment analyses suggested that they were closely involved in many important pathophysiological processes related to cancer development and progression. This may help reveal the potential mechanisms underlying HCC recurrence following LT and develop novel target therapies in the future.

There are still several limitations in the present study. First, given the sample size of the GSE30297 dataset is relatively small, this cohort could not be ideally divided into a training cohort and an internal validation cohort. In addition, apart from the GSE30297 dataset, no other suitable dataset was retrieved by thoroughly searching the GEO database. Thus, the efficacy of the five miRNA-based model needs to be externally validated based on large-scale cohorts in the future. Besides, except for MC, the detailed clinicopathologic information of HCC patients (including tumor features, specific follow-up information, neoadjuvant treatments, and immunosuppressive regimen) is not shared in the GSE30297 dataset. Therefore, it could not be achieved to comprehensively assess the independent prognostic value of the miRNA-based risk score by performing KM survival analysis or Cox regression analysis, and develop a more promising genomic-clinicopathologic nomogram integrating the miRNA-based model with other excellent selection criteria [such as Hangzhou criteria, University of California, San Francisco (UCSF) criteria and Up-to-seven criteria] or individual prognostic clinicopathological factors [such as alpha-fetoprotein (AFP), AFP-L3, and protein induced by vitamin K absence or antagonist II (PIVKA-II)] in term of HCC recurrence following LT, rather than simply utilizing the MC (31). Finally, in future studies, it is important and meaningful to deeply identify whether the tumor that occurred in the liver graft within 3 years following LT is a “true relapse” or whether it is a “de novo malignancy” from the perspective of tumor biology.

Conclusions

In conclusion, we developed a valuable five miRNA-based risk score for predicting the risk of HCC recurrence within 3 years after LT. The genomic-clinical model combining the miRNA-based model with MC displayed excellent predictive efficacy and clinical applicability. This model may not only contribute to the reasonable selection of LT candidates in HCC patients by accurate risk stratification regarding postoperative HCC recurrence, but also promote personalized postoperative management, such as surveillance strategy, immunosuppressive regimen, and adjuvant therapies.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-9/rc

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-9/prf

Funding: This study was supported by the Fundamental Research Funds for the Central Universities (grant No. 2042023kf0043) and the National Natural Science Foundation of China (grant No. 82403534).

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-9/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.

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. Magyar CTJ, O’Kane GM, Aceituno L, et al. Liver Transplantation for Hepatocellular Carcinoma: An Expanding Cornerstone of Care in the Era of Immunotherapy. J Clin Oncol 2025;43:589-604. [Crossref] [PubMed]
2. Tabrizian P, Abdelrahim M, Schwartz M. Immunotherapy and transplantation for hepatocellular carcinoma. J Hepatol 2024;80:822-5. [Crossref] [PubMed]
3. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693-9. [Crossref] [PubMed]
4. Singal AG, Kudo M, Bruix J. Breakthroughs in Hepatocellular Carcinoma Therapies. Clin Gastroenterol Hepatol 2023;21:2135-49. [Crossref] [PubMed]
5. van der Meeren PE, de Wilde RF, Sprengers D, et al. Benefit and harm of waiting time in liver transplantation for HCC. Hepatology 2025;82:212-31. [Crossref] [PubMed]
6. Terrault NA, Francoz C, Berenguer M, et al. Liver Transplantation 2023: Status Report, Current and Future Challenges. Clin Gastroenterol Hepatol 2023;21:2150-66. [Crossref] [PubMed]
7. Nemeth K, Bayraktar R, Ferracin M, et al. Non-coding RNAs in disease: from mechanisms to therapeutics. Nat Rev Genet 2024;25:211-32. [Crossref] [PubMed]
8. Zeng Q, Cao J, Niu Y, et al. Identification of Recurrence-Related mRNAs and Noncoding RNAs in Hepatocellular Carcinoma Following Liver Transplantation. Turk J Gastroenterol 2023;34:394-405. [Crossref] [PubMed]
9. Dundar HZ, Aksoy F, Aksoy SA, et al. Overexpression of miR-21 Is Associated With Recurrence in Patients With Hepatitis B Virus-Mediated Hepatocellular Carcinoma Undergoing Liver Transplantation. Transplant Proc 2019;51:1157-61. [Crossref] [PubMed]
10. Barry CT, D’Souza M, McCall M, et al. Micro RNA expression profiles as adjunctive data to assess the risk of hepatocellular carcinoma recurrence after liver transplantation. Am J Transplant 2012;12:428-37. [Crossref] [PubMed]
11. Norman JS, Li PJ, Kotwani P, et al. AFP-L3 and DCP strongly predict early hepatocellular carcinoma recurrence after liver transplantation. J Hepatol 2023;79:1469-77. [Crossref] [PubMed]
12. Sapisochin G, Bruix J. Liver transplantation for hepatocellular carcinoma: outcomes and novel surgical approaches. Nat Rev Gastroenterol Hepatol 2017;14:203-17. [Crossref] [PubMed]
13. Foerster F, Hoppe-Lotichius M, Vollmar J, et al. Long-term observation of hepatocellular carcinoma recurrence after liver transplantation at a European transplantation centre. United European Gastroenterol J 2019;7:838-49. [Crossref] [PubMed]
14. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12-49. [Crossref] [PubMed]
15. Lucey MR, Furuya KN, Foley DP. Liver Transplantation. N Engl J Med 2023;389:1888-900. [Crossref] [PubMed]
16. Liese J, Peveling-Oberhag J, Doering C, et al. A possible role of microRNAs as predictive markers for the recurrence of hepatocellular carcinoma after liver transplantation. Transpl Int 2016;29:369-80. [Crossref] [PubMed]
17. Xie QY, Almudevar A, Whitney-Miller CL, et al. A microRNA biomarker of hepatocellular carcinoma recurrence following liver transplantation accounting for within-patient heterogeneity. BMC Med Genomics 2016;9:18. [Crossref] [PubMed]
18. Garcia KB, Hussein A, Satish S, et al. Machine Perfusion as a Strategy to Decrease Ischemia-Reperfusion Injury and Lower Cancer Recurrence Following Liver Transplantation. Cancers (Basel) 2024;16:3959. [Crossref] [PubMed]
19. European Association for the Study of the Liver. EASL Clinical Practice Guidelines on liver transplantation. J Hepatol 2024;81:1040-86. [Crossref] [PubMed]
20. Zhong F, Wang Y. YY1-regulated lncRNA SOCS2-AS1 suppresses HCC cell stemness and progression via miR-454-3p/CPEB1. Biochem Biophys Res Commun 2023;679:98-109. [Crossref] [PubMed]
21. Hu T, Gu J, Tan L, et al. Identification and validation of an immune-related miRNA signature for predicting prognosis of hepatocellular carcinoma. Int Immunopharmacol 2025;146:113850. [Crossref] [PubMed]
22. Cave MC, Pinkston CM, Rai SN, et al. Circulating MicroRNAs, Polychlorinated Biphenyls, and Environmental Liver Disease in the Anniston Community Health Survey. Environ Health Perspect 2022;130:17003. [Crossref] [PubMed]
23. Meng H, Li R, Xie Y, et al. Nanoparticles Mediated circROBO1 Silencing to Inhibit Hepatocellular Carcinoma Progression by Modulating miR-130a-5p/CCNT2 Axis. Int J Nanomedicine 2023;18:1677-93. [Crossref] [PubMed]
24. Zhu B, Gu S, Wu X, et al. Bioinformatics analysis of tumor-educated platelet microRNAs in patients with hepatocellular carcinoma. Biosci Rep 2021;41:BSR20211420. [Crossref] [PubMed]
25. Lehrich BM, Zhang J, Monga SP, et al. Battle of the biopsies: Role of tissue and liquid biopsy in hepatocellular carcinoma. J Hepatol 2024;80:515-30. [Crossref] [PubMed]
26. Ali HEA, Abdel Hameed R, Effat H, et al. Circulating microRNAs panel as a diagnostic tool for discrimination of HCV-associated hepatocellular carcinoma. Clin Res Hepatol Gastroenterol 2017;41:e51-62. [Crossref] [PubMed]
27. Mourad L, El-Ahwany E, Zoheiry M, et al. Expression analysis of liver-specific circulating microRNAs in HCV-induced hepatocellular Carcinoma in Egyptian patients. Cancer Biol Ther 2018;19:400-6. [Crossref] [PubMed]
28. Lei Z, Tang X, Si A, et al. microRNA-454 promotes liver tumor-initiating cell expansion by regulating SOCS6. Exp Cell Res 2020;390:111955. [Crossref] [PubMed]
29. He H, Wang Y, Ye P, et al. Long noncoding RNA ZFPM2-AS1 acts as a miRNA sponge and promotes cell invasion through regulation of miR-139/GDF10 in hepatocellular carcinoma. J Exp Clin Cancer Res 2020;39:159. [Crossref] [PubMed]
30. Qin Z, Liu X, Li Z, et al. LncRNA LINC00667 aggravates the progression of hepatocellular carcinoma by regulating androgen receptor expression as a miRNA-130a-3p sponge. Cell Death Discov 2021;7:387. [Crossref] [PubMed]
31. Costentin CE, Bababekov YJ, Zhu AX, et al. Is It Time to Reconsider the Milan Criteria for Selecting Patients With Hepatocellular Carcinoma for Deceased-Donor Liver Transplantation? Hepatology 2019;69:1324-36. [Crossref] [PubMed]