Low Tumor miR-369-3p Levels Combined with Plasma Hepatitis B Virus Pre-S2 Gene Deletion Mutation Predict Higher Hepatocellular Carcinoma Recurrence Authors: Ya-Chi Wu, Long-Bin Jeng, Tsai-Chung Li, John Wang, Chiao-Fang Teng Published: 2026 ABSTRACT IntroductionAlthough resection surgery is a curative treatment for hepatocellular carcinoma (HCC), high HCC recurrence leads to poor patient survival. Chronic hepatitis B virus (HBV) infection is an important risk factor for HCC. Deletion mutation in HBV pre-S2 gene results in expression of pre-S2 mutant oncoprotein and represents an independent prognostic biomarker for HCC recurrence. MicroRNAs (miRNAs) are small non-coding RNAs that play key roles in HBV-related HCC.MethodsThis study aimed to identify the miRNAs whose expression levels in tumor tissues were correlated with pre-S2 gene deletion mutation and post-resection HCC recurrence and investigate their potential in combination with pre-S2 gene deletion mutation to predict HCC recurrence in a retrospective cohort of patients.ResultsThe results showed that the expression level of miR-369-3p was decreased in tumor tissues of patients with pre-S2 gene deletion mutation or HCC recurrence. miR-369-3p was identified as a prognostic biomarker for HCC recurrence. Patients with pre-S2 gene deletion mutation combined with a low level of miR-369-3p had a higher risk of HCC recurrence than patients with either one or none of these two biomarkers. The combination of pre-S2 gene deletion mutation and miR-369-3p expression level showed a greater prognostic potential for HCC recurrence than either biomarker alone.ConclusionCollectively, miR-369-3p held exploratory promise in serving as a combination biomarker with pre-S2 gene deletion mutation to provide a better potential in predicting HBV-related HCC recurrence after curative surgical resection. Introduction As a major histological form of primary liver cancer, hepatocellular carcinoma (HCC) contributes to over 90% of total liver malignancies and ranks sixth in morbidity and third in mortality among all human cancers worldwide, leading to approximately 900,000 new cases and 800,000 new deaths each year.1–3 Although curative therapeutic options such as liver transplantation and surgical resection have been developed for the treatment of HCC patients, low donor liver availability and high post-resection HCC recurrence (up to 70% within 5 years after surgery) diminish their survival benefit, respectively.4–6 Moreover, high tumor heterogeneity and drug resistance of HCC limit the treatment efficacy of systemic therapies such as chemotherapy and molecular targeted therapy, resulting in poor patient outcomes.7–9 Therefore, identification of valuable biomarkers for predicting post-resection HCC recurrence is a key goal to allow timely diagnosis and better management of the high-risk patients for improved patient prognosis. As a major etiological factor for HCC development, chronic hepatitis B virus (HBV) infection accounts for more than 50% of total HCC cases worldwide.10 Moreover, HBV is regarded as an important independent risk factor for HCC recurrence after curative surgical resection.11,12 HBV contains a circular genome with four open reading frames (ORFs): C, P, X, and S. The S ORF consists of three sequential segments-pre-S1, pre-S2, and S- and gives rise to three surface proteins of different sizes. The small surface protein is encoded by the S gene alone, the middle surface protein by the pre-S2 and S regions, and the large surface protein by the pre-S1, pre-S2, and S regions.13 Various deletion mutations which naturally occur in the HBV pre-S2 gene have been shown to result in the expression of a mutated form of large surface protein called pre-S2 mutant.14,15 Pre-S2 mutant functions as an important HBV oncoprotein which accumulates in the endoplasmic reticulum (ER) of hepatocytes and stimulates ER stress to initiate multiple signaling pathways involved in cell proliferation, survival, growth, cell cycle progression, glucose and lipid metabolism, and genomic instability to induce HCC tumorigenesis in vitro and in vivo.16,17 Moreover, inhibition of the pre-S2 mutant-activated oncogenic signaling pathways exhibits suppressive effects on HCC tumor formation in a pre-S2 mutant-transgenic mouse model.18,19 The presence of pre-S2 gene deletion mutation in liver tissues or blood of patients with chronic hepatitis B and HBV-related HCC is independently associated with a higher risk of HCC recurrence after curative surgical resection.20–23 Furthermore, patients with the presence of pre-S2 gene deletion mutation together with low serum albumin level, high plasma monocyte chemoattractant protein-1 level, high intratumoral programmed death ligand 1 expression or regulatory T cell infiltration, or high tumor-node-metastasis stage have been shown to have a much higher risk of post-resection HCC recurrence.24–27 Therefore, it would have great promise to combine pre-S2 gene deletion mutation with other biomarkers for developing a novel combination biomarker with superior performance in predicting HCC development and prognosis. As one of the most-studied small non-coding RNAs, microRNAs (miRNAs) play crucial roles in cancer development through repressing the expression of genes related to the malignant phenotypes of tumor cells, such as proliferation, survival, migration, and invasion.28 Based on their target genes as tumor suppressors or oncogenes, miRNAs function as oncogenes or tumor suppressors, respectively. Moreover, alteration of miRNA expression is implicated in different stages of HBV-related HCC tumorigenesis.29,30 HBV X protein has been demonstrated to manipulate the expression of many miRNAs at the epigenetic, transcriptional, and post-transcriptional levels.31 The dysregulated miRNAs in liver tissues and blood circulation (in free or exosome-enclosed forms) can be used as not only potential biomarkers for diagnosis, prognosis, and treatment response prediction but also therapeutic targets for HCC patients.32,33 Although the effect of pre-S2 mutant on regulating miRNA expression has recently been unraveled,34 the clinical and molecular association between pre-S2 mutant and miRNA expression remains largely unclarified. Therefore, to identify the miRNAs whose expression levels are altered in HBV-related HCC patients with pre-S2 gene deletion mutation would be promising to provide new insights into the regulation of miRNA expression by pre-S2 mutant and unravel potential miRNA biomarkers to be combined with pre-S2 gene deletion mutation for better predicting patient prognosis. This study was designed as a hypothesis-generating observational study, conducted in accordance with the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) and Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines,35,36 to evaluate the prognostic potential of pre-S2 gene deletion mutation combined with miRNA biomarkers for the recurrence of HBV-related HCC after curative surgical resection. First, a retrospectively recruited cohort of 70 HBV-related HCC patients who received curative surgical resection were divided into two groups based on the presence or absence of pre-S2 gene deletion mutation in plasma. Next, a total of 380 miRNAs were analyzed to identify the miRNAs whose expression levels in tumor tissues were dysregulated between the two groups of patients and whose expression alteration was correlated with the incidence of post-resection HCC recurrence. Furthermore, the combination between the identified miRNAs and pre-S2 gene deletion mutation was made and the prognostic potential of the combination biomarkers for post-resection HCC recurrence was assessed (Figure S1). Materials and Methods Collection of Patient Specimens The fresh-frozen paired tumor (T) and non-tumor (NT) liver tissues and plasma samples were retrospectively collected from 70 HBV-related HCC patients who underwent curative surgical resection at the China Medical University Hospital (CMUH; Taichung, Taiwan) from March 2004 to September 2016 under the approval of the Human Research Ethics Committee of CMUH (protocol no. CMUH107-REC1-080) (Figure S2). The patients who had chronic HBV infection, suffered from cytologically- or histologically-proved HCC, and received curative surgical resection as primary treatment were included in this study; however, the HCC patients who were co-infected with other types of hepatitis viruses or received other surgical, locoregional, or systemic therapies as combination or adjuvant treatments were excluded. The plasma and tissue samples were obtained from patients immediately before and at surgery, respectively; the former were stored at −80°C and the latter were kept in liquid nitrogen. The T and NT tissues were selected by the pathologists at the Department of Pathology of CMUH who confirmed the tumor cellularity and that NT tissues were free from tumor cells. The clinicopathological data of the patients were retrieved from the electronic medical records at CMUH. The 7th edition of the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system was applied for tumor staging. Recurrence-free survival (RFS) was defined as the interval from the date of surgical resection to the date of confirmed recurrence or death from any cause. Surveillance was performed every 3 months for the first 2 years, and every 6 months thereafter, including contrast-enhanced computed tomography (CT)/magnetic resonance imaging (MRI) and alpha-fetoprotein (AFP) tests. Recurrence was diagnosed via imaging showing classical HCC enhancement patterns or by biopsy. Patients without recurrence were censored at the last surveillance date. No patients died before the last follow-up date. All experiments were conducted in accordance with the principles of the 1975 Declaration of Helsinki and the informed consent was obtained from all patients. Detection of Pre-S2 Gene Deletion Mutation The pre-S2 gene deletion mutation was detected by the next-generation sequencing (NGS)-based platform as described.21,37 First, the viral DNA was extracted from the plasma samples by the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA, USA) and then served as a template for amplification of the pre-S1 and pre-S2 genes by polymerase chain reaction (PCR) with the high-fidelity Platinum SuperFi DNA Polymerase (Invitrogen, Carlsbad, CA, USA). The DNA purity was confirmed by an optical density (OD) ratio at 260/280 nm within the range of 1.7 to 1.9 using the NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The PCR assay was performed in two rounds. The first round was initiated with 1 cycle of 95°C for 1 minute, followed by 35 cycles of 95°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute; the second round was initiated with 1 cycle of 98°C for 1 minute, followed by 35 cycles of 98°C for 1 minute, 58°C for 1 minute, and 72°C for 1 minute. Next, the PCR product was used to prepare the DNA libraries using the KAPA Library Quantification Kit for Illumina Platforms (Kapa Biosystems, Boston, MA, USA). Base calling and quality scoring were performed by an updated implementation of Real-Time Analysis on the NextSeq 500 system (Illumina, San Diego, CA, USA) with a minimum read depth of 500×. The bcl2fastq Conversion Software v2.20 (Illumina) was used to demultiplex data and convert BCL files to standard FASTQ file formats. After trimming the adaptor sequence and masking the low-complexity sequence with a quality score under 20, the sequence reads were compared with the master sequences from the reference sets of HBV genotypes, which were available on the NCBI website (https://www.ncbi.nlm.nih.gov/projects/genotyping/view.cgi?db=2), by using BLAST. The E-value cutoff was set as 1e-5 to reduce false-positive result. Finally, the percentage of DNA fragments which contained deletion mutations in the pre-S1 gene solely, the pre-S2 gene solely, or both pre-S1 and pre-S2 genes was calculated as the counts of reads with deletion divided by the total number of reads and then multiplied by 100. The presence of pre-S2 gene deletion mutation was defined as the percentage of DNA fragments containing deletion mutations in either the pre-S2 gene solely or both pre-S1 and pre-S2 genes above a cut-off of 4.643%, which was pre-specified from previous studies,21,37 derived independently of the current cohort, and applied blinded to recurrence status. To serve as negative and positive controls, a known HBV-negative sample and a sample with known pre-S2 gene deletion mutation percentage were included in the analysis. To ensure unbiased analysis, the laboratory personnel performing the NGS library preparation, sequencing, and bioinformatic analysis were blinded to the clinical outcomes. The sequencing datasets are publicly unavailable due to protection of patient privacy. Measurement of miRNA Expression Levels The miRNAs were measured by reverse transcription quantitative PCR (RT-qPCR) assay. First, total RNA was isolated from the T and NT liver tissues by the mirVana miRNA Isolation Kit (Thermo Fisher Scientific, Waltham, MA, USA), followed by conversion into complementary DNA (cDNA) with the Mir-X miRNA First-Strand Synthesis Kit (Clontech, Mountain View, CA, USA). The RNA purity and integrity were confirmed by an OD ratio at 260/280 nm within the range of 1.8 to 2.0 using the NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies) and a RNA integrity number higher than 8.0 using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), respectively. Next, 2.5 ng of cDNA was used as a template for qPCR assay in a 10-μL reaction volume containing 0.4 μM of each miRNA-specific forward primer derived from the Human SeraMir Exosome RNA Profiling Kit (System Biosciences), 0.4 μM of universal reverse primer derived from the Mir-X miRNA First-Strand Synthesis Kit (Clontech), and 1× Fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) on the StepOnePlus Real-Time PCR System (Applied Biosystems). The Human SeraMir Exosome RNA Profiling Kit (System Biosciences) contains forward qPCR primers for 380 miRNAs, which are most frequently detected in not only human tissues but also blood exosomes. The qPCR program was started with 1 cycle of 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds, 60°C for 1 minute, and 72°C for 1 minute. At the end of each run, a melting curve analysis was performed from 65°C to 95°C to confirm the specificity of assay. A no-template sample was served as the negative control to ensure the absence of background noise. Cycle of quantification (Cq) values of > 35 were excluded from further mathematical calculations. For the samples with a Cq value of > 35, an increased amount of template input was applied in the assay to reduce the Cq value to 35 or below. All samples were amplified in triplicate and the obtained mean Cq value was then used for further analysis. The expression levels of miRNAs were normalized by the endogenous U6 small nuclear RNA, which was detected with a set of forward and reverse primers obtained from the Mir-X miRNA First-Strand Synthesis Kit (Clontech) and showed stable expression across samples (Table S1), and calculated as a T-to-NT (T/NT) ratio by the 2−ΔΔCq method as described.38 Statistical Analysis The difference of miRNA expression levels between patients with and without pre-S2 gene deletion mutation or HCC recurrence was analyzed by the multiple unpaired t-tests, corrected for multiple comparisons using the Holm-Sidak method. The association between miRNA expression levels and pre-S2 gene deletion mutation or HCC recurrence in patients was assessed by the chi-square test. The univariate and multivariate analyses of RFS were performed by the Cox proportional-hazards regression model without violation of proportional hazards assumption. The RFS curves were calculated by the Kaplan-Meier method. Pairwise comparisons between groups were performed by the Log rank test. Adjustment for multiple comparisons was conducted by the Sidak method. The receiver operating characteristic (ROC) curves were constructed to determine the area under the ROC curve (AUC). The differences between the AUCs were compared by the DeLong test. The model calibration was assessed by the Hosmer–Lemeshow test. The model selection was carried out by the Akaike information criterion (AIC) test. The model potential for predicting time-to-event outcomes was examined by time-dependent concordance index (C-index). A P value < 0.05 was considered statistically significant. In addition, a post hoc power analysis was performed based on the HCC recurrence rates reported in previous studies of pre-S2 gene deletion mutation,24–27 indicating that the current sample size had a statistical power of 0.997 at a significant level of 0.05 to detect a large effect size. The Youden index was used to identify the optimal outcome-based cut-off value of miR-369-3p expression level. Results Profile of Clinicopathological Features and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC Patients As shown in Table 1, the median age of all 70 patients was 53 years (range, 26 to 78), among whom 63 were men and 7 were women. All patients had chronic HBV infection, among whom 56 were genotype B and 14 were genotype C, and had a median serum HBV DNA level of 2.0×104 copies/mL (range, 21.5 to 1.5×108). All of 60 patients with available data were positive for HBV surface antigen (HBsAg) and 57 of 66 patients with available data were negative for HBV e antigen (HBeAg) in serum. All patients developed a single tumor with a median tumor size of 4.5 cm (range, 1.1 to 19.5). All patients received curative surgical resection, among whom 39 received antiviral therapy after surgery. The primary follow-up endpoint was RFS after surgery. After a median follow-up of 30.50 months (range, 1.6 to 159.5), the median RFS of all patients was 30.50 months (range, 1.6 to 159.5), among whom 38 experienced HCC recurrence but the other 32 did not. Based on the pre-S2 genotyping, 36 patients were positive for pre-S2 gene deletion mutation in plasma but the other 34 were not.Table 1Profile of Clinicopathological Characteristics and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC PatientsClinical CharacteristicsNo. of PatientsMedian (Range)Age (years)7053 (26–78) ≤ 502643.5 (26–50) > 504459.5 (51–78)Gender (women/men)7/63Smoking (no/yes)41/29Alcohol (no/yes)43/27HBsAg (negative/positive/NA)0/60/10HBeAg (negative/positive/NA)57/9/4HBV genotype (B/C)56/14HBV DNA (IU/mL) (20–1.7×108/< 20)a69/12.0×104 (21.5–1.5×108)e ≤ 1×104308.4×102 (21.5–9.3×103) > 1×104395.3×105 (1.2×104-1.5×108)Albumin (g/dL)703.7 (1.2–4.9) ≤ 3.8443.4 (1.2–3.8) > 3.8264.2 (3.9–4.9)AST (U/L)7059.5 (14–1052) ≤ 341427 (14–34) > 345674.5 (35–1052)ALT (U/L)7055.5 (13–1338) ≤ 402431 (13–40) > 404696.5 (41–1338)AFP (ng/mL) (≤ 54,000/> 54000)b67/350.2 (1.8–36600)e ≤ 4004414.3 (1.8–271) > 400231950 (461.7–36,600)Pathological CharacteristicsNo. of Patients (%)aMedian (Range)Tumor number (single/multiple)70/0Tumor size (cm)704.5 (1.1–19.5) ≤ 5362.4 (1.1–4.5) > 53410 (5.5–19.5)Tumor encapsulation (no/yes/NA)18/39/13Lymph node involvement (no/yes)64/6Portal vein thrombosis (no/yes)66/4Satellite nodule (no/yes)51/19Vascular invasion (no/yes)48/22Distant metastasis (no/yes)63/7Steatosis grade (0/1/2/3/NA)12/8/1/0/49Metavir inflammation score (0/1/2/3/NA)3/32/5/0/30Ishak fibrosis score (0/1/2/3/4/5/6/NA)5/13/9/7/3/4/10/19Child-Pugh cirrhosis score (A/B/C)54/14/2CLIP score (0/1/2/3/4/5/6)31/22/8/8/1/0/0Tumor differentiation grade (1/2/3/4)2/33/34/1BCLC stage (A/B/C/D)35/28/6/1AJCC TNM stage (I/II/IIIA/IIIB/IIIC/IVA/IVB)c39/18/6/5/2/0/0Antiviral therapy after surgery (no/yes)31/39HCC recurrence after surgery (month)d7026.6 (1.5–158.9) No32141.3 (25.4–158.9) Yes3810.7 (1.5–72.3)HBV Pre-S2 GenotypingNo. of PatientsMedian (Range)Pre-S2 gene deletion mutation (absence/presence)34/36miRNA Expression Level (T/NT)No. of PatientsMedian (Range)miR-369-3pf700.65 (0.01–19.44) High (> 0.22)521.41 (0.23–19.44) Low (≤ 0.22)180.09 (0.01–0.22)miR-369-3pg700.65 (0.01–19.44) High (> 0.65)352.44 (0.69–19.44) Low (≤ 0.65)350.22 (0.01–0.62)miR-369-3ph700.65 (0.01–19.44) High (> 2.55)174.93 (2.91–19.44) Low (≤ 2.55)530.38 (0.01–2.44)Notes: aHBV DNA was measured with a detection range of 20 to 1.7×108 IU/mL. bAFP was measured with the highest detection limit of 54000 ng/mL. cThe 7th edition of the AJCC TNM staging system was applied. dShown was the survival time free of recurrence after curative surgical resection within the follow-up period. eOnly data within the detection range were analyzed. fThe expression level of miR-369-3p was divided into high or low level based on the 25th percentile T/NT expression ratio. gThe expression level of miR-369-3p was divided into high or low level based on the median T/NT expression ratio. hThe expression level of miR-369-3p was divided into high or low level based on the 75th percentile T/NT expression ratio.Abbreviations: HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; NA, not available; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Table 1Profile of Clinicopathological Characteristics and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC PatientsClinical CharacteristicsNo. of PatientsMedian (Range)Age (years)7053 (26–78) ≤ 502643.5 (26–50) > 504459.5 (51–78)Gender (women/men)7/63Smoking (no/yes)41/29Alcohol (no/yes)43/27HBsAg (negative/positive/NA)0/60/10HBeAg (negative/positive/NA)57/9/4HBV genotype (B/C)56/14HBV DNA (IU/mL) (20–1.7×108/< 20)a69/12.0×104 (21.5–1.5×108)e ≤ 1×104308.4×102 (21.5–9.3×103) > 1×104395.3×105 (1.2×104-1.5×108)Albumin (g/dL)703.7 (1.2–4.9) ≤ 3.8443.4 (1.2–3.8) > 3.8264.2 (3.9–4.9)AST (U/L)7059.5 (14–1052) ≤ 341427 (14–34) > 345674.5 (35–1052)ALT (U/L)7055.5 (13–1338) ≤ 402431 (13–40) > 404696.5 (41–1338)AFP (ng/mL) (≤ 54,000/> 54000)b67/350.2 (1.8–36600)e ≤ 4004414.3 (1.8–271) > 400231950 (461.7–36,600)Pathological CharacteristicsNo. of Patients (%)aMedian (Range)Tumor number (single/multiple)70/0Tumor size (cm)704.5 (1.1–19.5) ≤ 5362.4 (1.1–4.5) > 53410 (5.5–19.5)Tumor encapsulation (no/yes/NA)18/39/13Lymph node involvement (no/yes)64/6Portal vein thrombosis (no/yes)66/4Satellite nodule (no/yes)51/19Vascular invasion (no/yes)48/22Distant metastasis (no/yes)63/7Steatosis grade (0/1/2/3/NA)12/8/1/0/49Metavir inflammation score (0/1/2/3/NA)3/32/5/0/30Ishak fibrosis score (0/1/2/3/4/5/6/NA)5/13/9/7/3/4/10/19Child-Pugh cirrhosis score (A/B/C)54/14/2CLIP score (0/1/2/3/4/5/6)31/22/8/8/1/0/0Tumor differentiation grade (1/2/3/4)2/33/34/1BCLC stage (A/B/C/D)35/28/6/1AJCC TNM stage (I/II/IIIA/IIIB/IIIC/IVA/IVB)c39/18/6/5/2/0/0Antiviral therapy after surgery (no/yes)31/39HCC recurrence after surgery (month)d7026.6 (1.5–158.9) No32141.3 (25.4–158.9) Yes3810.7 (1.5–72.3)HBV Pre-S2 GenotypingNo. of PatientsMedian (Range)Pre-S2 gene deletion mutation (absence/presence)34/36miRNA Expression Level (T/NT)No. of PatientsMedian (Range)miR-369-3pf700.65 (0.01–19.44) High (> 0.22)521.41 (0.23–19.44) Low (≤ 0.22)180.09 (0.01–0.22)miR-369-3pg700.65 (0.01–19.44) High (> 0.65)352.44 (0.69–19.44) Low (≤ 0.65)350.22 (0.01–0.62)miR-369-3ph700.65 (0.01–19.44) High (> 2.55)174.93 (2.91–19.44) Low (≤ 2.55)530.38 (0.01–2.44)Notes: aHBV DNA was measured with a detection range of 20 to 1.7×108 IU/mL. bAFP was measured with the highest detection limit of 54000 ng/mL. cThe 7th edition of the AJCC TNM staging system was applied. dShown was the survival time free of recurrence after curative surgical resection within the follow-up period. eOnly data within the detection range were analyzed. fThe expression level of miR-369-3p was divided into high or low level based on the 25th percentile T/NT expression ratio. gThe expression level of miR-369-3p was divided into high or low level based on the median T/NT expression ratio. hThe expression level of miR-369-3p was divided into high or low level based on the 75th percentile T/NT expression ratio.Abbreviations: HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; NA, not available; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Profile of Clinicopathological Characteristics and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC Patients Notes: aHBV DNA was measured with a detection range of 20 to 1.7×108 IU/mL. bAFP was measured with the highest detection limit of 54000 ng/mL. cThe 7th edition of the AJCC TNM staging system was applied. dShown was the survival time free of recurrence after curative surgical resection within the follow-up period. eOnly data within the detection range were analyzed. fThe expression level of miR-369-3p was divided into high or low level based on the 25th percentile T/NT expression ratio. gThe expression level of miR-369-3p was divided into high or low level based on the median T/NT expression ratio. hThe expression level of miR-369-3p was divided into high or low level based on the 75th percentile T/NT expression ratio. Abbreviations: HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; NA, not available; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Decreased Expression Level of miR-369-3p in the Tumor Tissues of HBV-Related HCC Patients with Pre-S2 Gene Deletion Mutation or Post-Resection HCC Recurrence To identify the miRNAs whose expression levels in tumor tissues were significantly altered between patients with and without pre-S2 gene deletion mutation or post-resection HCC recurrence, a total of 380 miRNAs were analyzed. As shown in Figure 1A and Table S2, among the miRNAs analyzed, 7 miRNAs (let-7g-5p, miR-30b-5p, miR-148a-3p, miR-302b-3p, miR-335-5p, miR-369-3p, and miR-506-3p) displayed a significantly altered expression level in the tumor tissues between patients with and without pre-S2 gene deletion mutation. The log2-transformed T/NT expression ratios of let-7g-5p, miR-30b-5p, miR-148a-3p, and miR-335-5p were significantly higher in patients with pre-S2 gene deletion mutation than patients without pre-S2 gene deletion mutation (P = 0.0309, P = 0.0078, P = 0.0252, and P = 0.0408, respectively), while those of miR-302b-3p, miR-369-3p, and miR-506-3p lower in patients with pre-S2 gene deletion mutation than patients without pre-S2 gene deletion mutation (P = 0.0239, P = 0.0088, and P = 0.0188, respectively). As shown in Figure 1B and Table S3, 9 miRNAs (miR-101-3p, miR-188-3p, miR-215-5p, miR-302d-3p, miR-367-3p, miR-369-3p, miR-494-3p, miR-498, and miR-517c-3p) in analysis exhibited a significantly altered expression level in the tumor tissues between patients with and without post-resection HCC recurrence. Patients with post-resection HCC recurrence had significantly higher log2-transformed T/NT expression ratios of miR-215-5p and miR-302d-3p (P = 0.0450 and P = 0.0464, respectively) and lower ratios of miR-101-3p, miR-188-3p, miR-367-3p, miR-369-3p, miR-494-3p, miR-498, and miR-517c-3p than patients without post-resection HCC recurrence (P = 0.0401, P = 0.0448, P = 0.0159, P = 0.0394, P = 0.0476, P = 0.0379, and P = 0.0146, respectively). Since miR-369-3p was the only miRNA whose expression levels were associated with both pre-S2 gene deletion mutation and post-resection HCC recurrence, miR-369-3p was chosen for further investigation. Furthermore, when the expression level of miR-369-3p in tumor tissues was divided into high or low level based on the 25th percentile (0.22), median (0.65), or 75th percentile T/NT expression ratio (2.55), the significant association of miR-369-3p expression level with pre-S2 gene deletion mutation and post-resection HCC recurrence was only observed when the 75th percentile rather than 25th percentile or median T/NT expression ratio was used as the cut-off value (Table 2 and Table 3). Up to 31 of 36 (86%) patients with pre-S2 gene deletion mutation and 34 of 38 (89%) patients with post-resection HCC recurrence had a low expression level of miR-369-3p (T/NT ≤ 2.55) but only 22 of 34 (65%) patients without pre-S2 gene deletion mutation and 19 of 32 (59%) patients without post-resection HCC recurrence did (P = 0.0306 and P = 0.0049, respectively) (Table 2 and Table 3). According to these results, the 75th percentile T/NT expression ratio was chosen as the optimal cut-off value for miR-369-3p expression level in the subsequent analyses. Additionally, in consistence with previous studies, the frequency of post-resection HCC recurrence was significantly higher in patients with pre-S2 gene deletion mutation (31 of 36 (86%) patients) than patients without pre-S2 gene deletion mutation (7 of 34 (21%) patients) (P value < 0.0001) (Table 2).Table 2Association Between miR-369-3p Expression Level and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC PatientsmiRNA Expression Level (T/NT)Pre-S2 Mutation Absence (No. of Patients (%))Pre-S2 Mutation Presence (No. of Patients (%))P ValuedmiR-369-3pa34 (100)36 (100)1.0000 High (> 0.22)25 (74)27 (75) Low (≤ 0.22)9 (26)9 (25)miR-369-3pb34 (100)36 (100)0.0262* High (> 0.65)22 (65)13 (36) Low (≤ 0.65)12 (35)23 (64)miR-369-3pc34 (100)36 (100)0.0306* High (> 2.55)12 (35)5 (14) Low (≤ 2.55)22 (65)31 (86)Post-Resection HCC RecurrencePre-S2 Mutation Absence (No. of Patients (%))Pre-S2 Mutation Presence (No. of Patients (%))P ValuedHCC recurrence34 (100)36 (100)<0.0001*** No27 (79)5 (14) Yes7 (21)31 (86)Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. *P value < 0.05; ***P value < 0.001.Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Table 3Association Between miR-369-3p Expression Level and Post-Resection HCC Recurrence in 70 HBV-Related HCC PatientsmiRNA Expression Level (T/NT)HCC Non-Recurrence (No. of Patients (%))HCC Recurrence (No. of Patients (%))P ValuedmiR-369-3pa32 (100)38 (100)0.4139 High (> 0.22)22 (69)30 (79) Low (≤ 0.22)10 (31)8 (21)miR-369-3pb32 (100)38 (100)0.8106 High (> 0.65)17 (53)18 (47) Low (≤ 0.65)15 (47)20 (53)miR-369-3pc32 (100)38 (100)0.0049** High (> 2.55)13 (41)4 (11) Low (≤ 2.55)19 (59)34 (89)Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. **P value < 0.01.Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Figure 1Comparative analysis of the expression level of miR-369-3p in tumor tissues between HBV-related HCC patients with and without pre-S2 gene deletion mutation or post-resection HCC recurrence. The difference of log2-transformed T/NT expression ratio of miR-369-3p between the patients with and without pre-S2 gene deletion mutation (A) or HCC recurrence (B) was analyzed by the multiple unpaired t-tests, corrected for multiple comparisons using the Holm-Sidak method. Thick and thin dashed horizontal lines represented the median and quartile values of miR-369-3p expression level in each group of patients, respectively. *, P value < 0.05; **, P value < 0.001.Image A: Violin plot of miR-369-3p expression by pre-S2 gene deletion mutation status. P = 0.0088 (**). Y-axis: Expression level (log2(T/NT)), ticks at -10, -5, 0, 5, 10. X-axis: miR-369-3p. Two violins: patients without and with pre-S2 gene deletion mutation. Dotted line at 0. Each violin has a thick dashed line (median) and two thin dashed lines (quartiles). Table: Patients 34 and 36. Median (range) 0.49 (-6.64 to 4.28) and -2.93 (-6.64 to 2.94). Image B: Violin plot of miR-369-3p expression by hepatocellular carcinoma recurrence status. P = 0.0394 (*). Y-axis: Expression level (log2(T/NT)), ticks at -10, -5, 0, 5, 10. X-axis: miR-369-3p. Two violins: patients without and with hepatocellular carcinoma recurrence. Dotted line at 0. Each violin has a thick dashed line (median) and two thin dashed lines (quartiles). Table: Patients 32 and 38. Median (range) 0.02 (-6.64 to 4.28) and -2.90 (-6.64 to 2.20).Two violin plots of miR-369-3p expression comparing patient groups, with lower medians in red groups. Table 2Association Between miR-369-3p Expression Level and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC PatientsmiRNA Expression Level (T/NT)Pre-S2 Mutation Absence (No. of Patients (%))Pre-S2 Mutation Presence (No. of Patients (%))P ValuedmiR-369-3pa34 (100)36 (100)1.0000 High (> 0.22)25 (74)27 (75) Low (≤ 0.22)9 (26)9 (25)miR-369-3pb34 (100)36 (100)0.0262* High (> 0.65)22 (65)13 (36) Low (≤ 0.65)12 (35)23 (64)miR-369-3pc34 (100)36 (100)0.0306* High (> 2.55)12 (35)5 (14) Low (≤ 2.55)22 (65)31 (86)Post-Resection HCC RecurrencePre-S2 Mutation Absence (No. of Patients (%))Pre-S2 Mutation Presence (No. of Patients (%))P ValuedHCC recurrence34 (100)36 (100)<0.0001*** No27 (79)5 (14) Yes7 (21)31 (86)Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. *P value < 0.05; ***P value < 0.001.Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Association Between miR-369-3p Expression Level and Pre-S2 Gene Deletion Mutation in 70 HBV-Related HCC Patients Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. *P value < 0.05; ***P value < 0.001. Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Table 3Association Between miR-369-3p Expression Level and Post-Resection HCC Recurrence in 70 HBV-Related HCC PatientsmiRNA Expression Level (T/NT)HCC Non-Recurrence (No. of Patients (%))HCC Recurrence (No. of Patients (%))P ValuedmiR-369-3pa32 (100)38 (100)0.4139 High (> 0.22)22 (69)30 (79) Low (≤ 0.22)10 (31)8 (21)miR-369-3pb32 (100)38 (100)0.8106 High (> 0.65)17 (53)18 (47) Low (≤ 0.65)15 (47)20 (53)miR-369-3pc32 (100)38 (100)0.0049** High (> 2.55)13 (41)4 (11) Low (≤ 2.55)19 (59)34 (89)Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. **P value < 0.01.Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Association Between miR-369-3p Expression Level and Post-Resection HCC Recurrence in 70 HBV-Related HCC Patients Notes: aThe expression level of miR-369-3p above the 25th percentile T/NT expression ratio was defined as high level. bThe expression level of miR-369-3p above the median T/NT expression ratio was defined as high level. cThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. dP value was determined by the chi-square test. **P value < 0.01. Abbreviations: miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; T, tumor; NT, non-tumor. Comparative analysis of the expression level of miR-369-3p in tumor tissues between HBV-related HCC patients with and without pre-S2 gene deletion mutation or post-resection HCC recurrence. The difference of log2-transformed T/NT expression ratio of miR-369-3p between the patients with and without pre-S2 gene deletion mutation (A) or HCC recurrence (B) was analyzed by the multiple unpaired t-tests, corrected for multiple comparisons using the Holm-Sidak method. Thick and thin dashed horizontal lines represented the median and quartile values of miR-369-3p expression level in each group of patients, respectively. *, P value < 0.05; **, P value < 0.001. Association of the Presence of Pre-S2 Gene Deletion Mutation or Low Expression Level of miR-369-3p Alone with a Higher Risk of Post-Resection HCC Recurrence Considering that the expression level of miR-369-3p in tumor tissues was significantly downregulated in patients with pre-S2 gene deletion mutation or post-resection HCC recurrence, the association between miR-369-3p expression level and the risk of post-resection HCC recurrence was assessed. As shown in Table 4, the univariate RFS analysis showed that the presence of pre-S2 gene deletion mutation, low expression level of miR-369-3p (T/NT ≤ 2.55), formation of satellite nodule, development of distant metastasis, and high AJCC TNM stage (IIIA-IIIC) were significantly associated with a higher risk of post-resection HCC recurrence (hazard ratio (HR) (95% confidence interval (CI)), 5.647 (2.436 to 13.094), P value < 0.0001; 2.943 (1.018 to 8.507), P value = 0.0462; 3.472 (1.719 to 7.011), P value = 0.0005; 3.265 (1.332 to 8.003), P value = 0.0097; 6.128 (2.960 to 12.686), P value < 0.0001, respectively). However, when these five biomarkers which exhibited significance in the univariate RFS analysis were further analyzed in the multivariate RFS analysis, only the presence of pre-S2 gene deletion mutation, low expression level of miR-369-3p, and high AJCC TNM stage still showed a significantly negative impact on RFS (HR (95% CI), 6.464 (2.516 to 16.605), P value = 0.0001; 3.745 (1.186 to 11.765), P value = 0.0244; 10.770 (2.931 to 39.584), P value = 0.0003, respectively) (Table 4; Model 1). Furthermore, the Kaplan-Meier analysis revealed that patients with the presence of pre-S2 gene deletion mutation or low expression level of miR-369-3p displayed significantly shorter RFS than patients with the absence of pre-S2 gene deletion mutation or high expression level of miR-369-3p (median RFS (range), 16.45 (1.6 to 158.9) versus 68.50 (2.0 to 159.5) months, log-rank P value < 0.0001; 20.20 (1.6 to 158.9) versus 110.90 (3.0 to 159.5) months, log-rank P value = 0.0021, respectively) (Figure 2A and B).Table 4Univariate and Multivariate RFS Analyses of Pre-S2 Gene Deletion Mutation and/or miR-369-3p Expression Level in 70 HBV-Related HCC PatientsClinical CharacteristicsaUnivariate AnalysisMultivariate AnalysiscModel 1Model 2Model 3HR95% CIP ValuebHR95% CIP ValuebHR95% CIP ValuebHR95% CIP ValuebAge (years) ≤ 50Reference > 500.9860.504–1.9300.9671Gender WomenReference Men1.3510.415–4.4030.6174Smoking NoReference Yes0.8280.428–1.6020.5755Alcohol NoReference Yes0.8820.451–1.7260.7136HBsAgd Negative PositiveHBeAg NegativeReference Positive0.9110.321–2.5850.8612HBV genotype CReferenceReferenceReferenceReference B0.5980.304–1.1760.13620.1770.069–0.4510.0003***0.2010.083–0.4850.0004***0.1900.079–0.4540.0002***HBV DNA (IU/mL) ≤ 1×104Reference > 1×1041.4090.729–2.7240.3078Albumin (g/dL) ≤ 3.8Reference > 3.80.6980.356–1.3680.2953AST (U/L) ≤ 34ReferenceReferenceReferenceReference > 340.6470.314–1.3330.23770.6870.280–1.6880.41320.6680.275–1.6240.37380.6260.257–1.5250.3025ALT (U/L) ≤ 40Reference > 400.7670.400–1.4700.4237AFP (ng/mL) ≤ 400Reference > 4001.6540.866–3.1570.1274Pathological CharacteristicsaTumor numbere Single MultipleTumor size (cm) ≤ 5ReferenceReferenceReferenceReference > 51.5520.819–2.9400.17750.7680.291–2.0260.59320.7170.274–1.8800.49910.7340.296–1.8250.5063Tumor encapsulation NoReference Yes0.5530.266–1.1520.1136Lymph node involvement NoReference Yes0.4310.104–1.7910.2468Portal vein thrombosis NoReference Yes0.7930.191–3.2980.7501Satellite nodule NoReferenceReferenceReferenceReference Yes3.4721.719–7.0110.0005***1.4440.477–4.3690.51541.5710.521–4.7390.42241.8580.658–5.2420.2420Vascular invasion NoReferenceReferenceReferenceReference Yes1.7750.961–3.4410.08931.1320.466–2.7450.78471.1650.470–2.8890.74170.9770.414–2.3040.9571Distant metastasis NoReferenceReferenceReferenceReference Yes3.2651.332–8.0030.0097**2.3220.771–6.9920.13422.3010.731–7.2400.15432.9540.967–9.0170.0572Steatosis gradef 0-1Reference 23.4720.405–29.7260.2561Metavir inflammation scoreg 0-1Reference 20.3910.091–1.6730.2055Ishak fibrosis score 0-3Reference 4-60.8640.408–1.8290.7027Child-Pugh cirrhosis score AReference B-C1.4150.669–2.9950.3639CLIP scoreh 0-3ReferenceReferenceReferenceReference 43.2870.439–24.6330.24683.4250.349–33.6250.29073.3290.337–32.8430.30312.8110.290–27.2910.3727Tumor differentiation grade 1-2ReferenceReferenceReferenceReference 3-41.6300.855–3.1070.13762.2850.944–5.5310.06692.2020.896–5.4140.08542.5521.074–6.0680.0339*BCLC stage A-BReference C-D1.1270.399–3.1830.8216AJCC TNM stagei I–IIReferenceReferenceReferenceReference IIIA-IIIC6.1282.960–12.686< 0.0001***10.7702.931–39.5840.0003***10.2762.687–39.3030.0007***8.2302.437–27.7970.0007***Antiviral therapy after surgery NoReference Yes1.1240.715–1.9660.3462HBV Pre-S2 GenotypingPre-S2 gene deletion mutation AbsenceReferenceReference Presence5.6472.436–13.094< 0.0001***6.4642.516–16.6050.0001***miRNA Expression Level (T/NT)miR-369-3pj High (> 2.55)ReferenceReference Low (≤ 2.55)2.9431.018–8.5070.0462*3.4311.077–10.9300.0370*Combination of Pre-S2 Gene Deletion Mutation and miR-369-3p Expression Level Group 1, Absence and HighReferenceReference Group 2–3, Absence and Low & Presence and High6.0770.768–48.0870.08736.3530.693–58.2150.1019 Group 4, Presence and Low25.2023.388–187.4900.0016**30.3633.735–246.8130.0014** Group 1–3, Absence and High/Low & Presence and HighReferenceReference Group 4, Presence and Low6.1762.950–12.931< 0.0001***7.6143.299–17.573< 0.0001***Notes: aOnly patients with available data were included in the statistical analysis. bP value was determined by the Cox proportional-hazards regression model. cThe factors with a P value < 0.25 in the univariate analysis were included as covariates in the multivariate analysis. Four factors (AFP, lymph node involvement, Metavir inflammation score, and tumor encapsulation) were excluded due to the lack of complete data for 70 patients. dNo patients were HBsAg negative for analysis. eNo patients had more than one tumor for analysis. fNo patients were steatosis grade 3 for analysis. gNo patients were Metavir inflammation score 3 for analysis. hNo patients were CLIP score 5–6 for analysis. iNo patients were AJCC TNM stage IVA-IVB for analysis. jThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. *P value < 0.05; **P value < 0.01; ***P value < 0.001.Abbreviations: RFS, recurrence-free survival; miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HR, hazard ratio; CI, confidence interval; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Figure 2Kaplan-Meier curves for comparing the RFS of HBV-related HCC patients with different statuses of pre-S2 gene deletion mutation and/or miR-369-3p expression level. (A) RFS difference between patients with presence and absence of pre-S2 gene deletion mutation. (B) RFS difference between patients with low (T/NT ≤ 2.55) and high expression level (T/NT > 2.55) of miR-369-3p. (C) RFS difference between a combination of Group 2 and Group 3 of patients (absence of pre-S2 gene deletion mutation and low expression level of miR-369-3p or presence of pre-S2 gene deletion mutation and high expression level of miR-369-3p) (designated Group 2–3) or Group 4 of patients (presence of pre-S2 gene deletion mutation and low expression level of miR-369-3p) and Group 1 of patients (absence of pre-S2 gene deletion mutation and high expression level of miR-369-3p). (D) RFS difference between Group 4 of patients and a combination of Group 1 to Group 3 of patients (designated Group 1–3). The number of patients at risk and the number of patients censored (shown in parenthesis) were listed in the bottom of each plot. **, P value < 0.01; ***, P value < 0.001.Image A: Kaplan-Meier graph shows recurrence-free survival over 150 months. Absence (n=34, solid) vs. Presence (n=36, dashed), P<0.0001. Absence remains higher; Presence drops to ~0.1 by 70-80 months. Image B: Kaplan-Meier graph with Low (n=53, solid) vs. High (n=17, dashed), P=0.0021. High stays ~0.7-0.9; Low declines to 0.3 by 70-80 months. Image C: Kaplan-Meier graph with Group 1 (n=12, solid), Group 2-3 (n=27, dashed), Group 4 (n=31, dash-dot), P<0.0001. Group 1 stays near 1.0; Group 2-3 declines to 0.6; Group 4 drops to 0.1 by 70-80 months. Image D: Kaplan-Meier graph with Group 1-3 (n=39, solid) vs. Group 4 (n=31, dashed), P<0.0001. Group 1-3 remains ~0.7-0.8; Group 4 declines to 0.1 by 70-80 months.Different Kaplan-Meier plots comparing recurrence-free survival after surgery across four group definitions. Table 4Univariate and Multivariate RFS Analyses of Pre-S2 Gene Deletion Mutation and/or miR-369-3p Expression Level in 70 HBV-Related HCC PatientsClinical CharacteristicsaUnivariate AnalysisMultivariate AnalysiscModel 1Model 2Model 3HR95% CIP ValuebHR95% CIP ValuebHR95% CIP ValuebHR95% CIP ValuebAge (years) ≤ 50Reference > 500.9860.504–1.9300.9671Gender WomenReference Men1.3510.415–4.4030.6174Smoking NoReference Yes0.8280.428–1.6020.5755Alcohol NoReference Yes0.8820.451–1.7260.7136HBsAgd Negative PositiveHBeAg NegativeReference Positive0.9110.321–2.5850.8612HBV genotype CReferenceReferenceReferenceReference B0.5980.304–1.1760.13620.1770.069–0.4510.0003***0.2010.083–0.4850.0004***0.1900.079–0.4540.0002***HBV DNA (IU/mL) ≤ 1×104Reference > 1×1041.4090.729–2.7240.3078Albumin (g/dL) ≤ 3.8Reference > 3.80.6980.356–1.3680.2953AST (U/L) ≤ 34ReferenceReferenceReferenceReference > 340.6470.314–1.3330.23770.6870.280–1.6880.41320.6680.275–1.6240.37380.6260.257–1.5250.3025ALT (U/L) ≤ 40Reference > 400.7670.400–1.4700.4237AFP (ng/mL) ≤ 400Reference > 4001.6540.866–3.1570.1274Pathological CharacteristicsaTumor numbere Single MultipleTumor size (cm) ≤ 5ReferenceReferenceReferenceReference > 51.5520.819–2.9400.17750.7680.291–2.0260.59320.7170.274–1.8800.49910.7340.296–1.8250.5063Tumor encapsulation NoReference Yes0.5530.266–1.1520.1136Lymph node involvement NoReference Yes0.4310.104–1.7910.2468Portal vein thrombosis NoReference Yes0.7930.191–3.2980.7501Satellite nodule NoReferenceReferenceReferenceReference Yes3.4721.719–7.0110.0005***1.4440.477–4.3690.51541.5710.521–4.7390.42241.8580.658–5.2420.2420Vascular invasion NoReferenceReferenceReferenceReference Yes1.7750.961–3.4410.08931.1320.466–2.7450.78471.1650.470–2.8890.74170.9770.414–2.3040.9571Distant metastasis NoReferenceReferenceReferenceReference Yes3.2651.332–8.0030.0097**2.3220.771–6.9920.13422.3010.731–7.2400.15432.9540.967–9.0170.0572Steatosis gradef 0-1Reference 23.4720.405–29.7260.2561Metavir inflammation scoreg 0-1Reference 20.3910.091–1.6730.2055Ishak fibrosis score 0-3Reference 4-60.8640.408–1.8290.7027Child-Pugh cirrhosis score AReference B-C1.4150.669–2.9950.3639CLIP scoreh 0-3ReferenceReferenceReferenceReference 43.2870.439–24.6330.24683.4250.349–33.6250.29073.3290.337–32.8430.30312.8110.290–27.2910.3727Tumor differentiation grade 1-2ReferenceReferenceReferenceReference 3-41.6300.855–3.1070.13762.2850.944–5.5310.06692.2020.896–5.4140.08542.5521.074–6.0680.0339*BCLC stage A-BReference C-D1.1270.399–3.1830.8216AJCC TNM stagei I–IIReferenceReferenceReferenceReference IIIA-IIIC6.1282.960–12.686< 0.0001***10.7702.931–39.5840.0003***10.2762.687–39.3030.0007***8.2302.437–27.7970.0007***Antiviral therapy after surgery NoReference Yes1.1240.715–1.9660.3462HBV Pre-S2 GenotypingPre-S2 gene deletion mutation AbsenceReferenceReference Presence5.6472.436–13.094< 0.0001***6.4642.516–16.6050.0001***miRNA Expression Level (T/NT)miR-369-3pj High (> 2.55)ReferenceReference Low (≤ 2.55)2.9431.018–8.5070.0462*3.4311.077–10.9300.0370*Combination of Pre-S2 Gene Deletion Mutation and miR-369-3p Expression Level Group 1, Absence and HighReferenceReference Group 2–3, Absence and Low & Presence and High6.0770.768–48.0870.08736.3530.693–58.2150.1019 Group 4, Presence and Low25.2023.388–187.4900.0016**30.3633.735–246.8130.0014** Group 1–3, Absence and High/Low & Presence and HighReferenceReference Group 4, Presence and Low6.1762.950–12.931< 0.0001***7.6143.299–17.573< 0.0001***Notes: aOnly patients with available data were included in the statistical analysis. bP value was determined by the Cox proportional-hazards regression model. cThe factors with a P value < 0.25 in the univariate analysis were included as covariates in the multivariate analysis. Four factors (AFP, lymph node involvement, Metavir inflammation score, and tumor encapsulation) were excluded due to the lack of complete data for 70 patients. dNo patients were HBsAg negative for analysis. eNo patients had more than one tumor for analysis. fNo patients were steatosis grade 3 for analysis. gNo patients were Metavir inflammation score 3 for analysis. hNo patients were CLIP score 5–6 for analysis. iNo patients were AJCC TNM stage IVA-IVB for analysis. jThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. *P value < 0.05; **P value < 0.01; ***P value < 0.001.Abbreviations: RFS, recurrence-free survival; miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HR, hazard ratio; CI, confidence interval; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Univariate and Multivariate RFS Analyses of Pre-S2 Gene Deletion Mutation and/or miR-369-3p Expression Level in 70 HBV-Related HCC Patients Notes: aOnly patients with available data were included in the statistical analysis. bP value was determined by the Cox proportional-hazards regression model. cThe factors with a P value < 0.25 in the univariate analysis were included as covariates in the multivariate analysis. Four factors (AFP, lymph node involvement, Metavir inflammation score, and tumor encapsulation) were excluded due to the lack of complete data for 70 patients. dNo patients were HBsAg negative for analysis. eNo patients had more than one tumor for analysis. fNo patients were steatosis grade 3 for analysis. gNo patients were Metavir inflammation score 3 for analysis. hNo patients were CLIP score 5–6 for analysis. iNo patients were AJCC TNM stage IVA-IVB for analysis. jThe expression level of miR-369-3p above the 75th percentile T/NT expression ratio was defined as high level. *P value < 0.05; **P value < 0.01; ***P value < 0.001. Abbreviations: RFS, recurrence-free survival; miR-369-3p, microRNA-369-3p; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HR, hazard ratio; CI, confidence interval; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; CLIP, Cancer of the Liver Italian Program; BCLC, Barcelona Clinic Liver Cancer; AJCC, American Joint Committee on Cancer; TNM, tumor-node-metastasis; T, tumor; NT, non-tumor. Kaplan-Meier curves for comparing the RFS of HBV-related HCC patients with different statuses of pre-S2 gene deletion mutation and/or miR-369-3p expression level. (A) RFS difference between patients with presence and absence of pre-S2 gene deletion mutation. (B) RFS difference between patients with low (T/NT ≤ 2.55) and high expression level (T/NT > 2.55) of miR-369-3p. (C) RFS difference between a combination of Group 2 and Group 3 of patients (absence of pre-S2 gene deletion mutation and low expression level of miR-369-3p or presence of pre-S2 gene deletion mutation and high expression level of miR-369-3p) (designated Group 2–3) or Group 4 of patients (presence of pre-S2 gene deletion mutation and low expression level of miR-369-3p) and Group 1 of patients (absence of pre-S2 gene deletion mutation and high expression level of miR-369-3p). (D) RFS difference between Group 4 of patients and a combination of Group 1 to Group 3 of patients (designated Group 1–3). The number of patients at risk and the number of patients censored (shown in parenthesis) were listed in the bottom of each plot. **, P value < 0.01; ***, P value < 0.001. Association of the Presence of Pre-S2 Gene Deletion Mutation Combined with Low Expression Level of miR-369-3p with a Greater Risk of Post-Resection HCC Recurrence Than Either Biomarker Alone Since low expression level of miR-369-3p was significantly associated with the presence of pre-S2 gene deletion mutation and the presence of pre-S2 gene deletion mutation or low expression level of miR-369-3p alone displayed a significant association with the risk of post-resection HCC recurrence, the association of the combination of these two biomarkers with post-resection HCC recurrence was analyzed. Based on the combinations of different statuses of pre-S2 gene deletion mutation (absence or presence) and miR-369-3p expression level (T/NT ≤ 2.55 or T/NT > 2.55), the 70 HBV-related HCC patients were divided into four groups: Group 1, absence of pre-S2 gene deletion mutation and high expression level of miR-369-3p (12 patients); Group 2, absence of pre-S2 gene deletion mutation and low expression level of miR-369-3p (22 patients); Group 3, presence of pre-S2 gene deletion mutation and high expression level of miR-369-3p (5 patients); and Group 4, presence of pre-S2 gene deletion mutation and low expression level of miR-369-3p (31 patients). As shown in Table 4, the univariate RFS analysis revealed that the Group 4 combination rather than a combination of Group 2 and Group 3 (designated Group 2–3) was significantly associated with a higher risk of post-resection HCC recurrence compared with the Group 1 combination (HR (95% CI), 6.077 (0.768 to 48.087), P value = 0.0873; 25.202 (3.388 to 187.490), P value = 0.0016, respectively). When these groups of biomarker combinations and the other three biomarkers with univariate significance (satellite nodule, distant metastasis, and AJCC TNM stage) were further examined in the multivariate RFS analysis, only the Group 4 combination of pre-S2 gene deletion mutation and miR-369-3p expression level and high AJCC TNM stage showed a significant impact on RFS (HR (95% CI), 30.363 (3.735 to 246.813), P value = 0.0014; 10.276 (2.687 to 39.303), P value = 0.0007, respectively) (Table 4; Model 2). Furthermore, when the Group 4 combination was compared with a combination of the other three groups (designated Group 1–3), the Group 4 combination remained associated with a significantly higher risk of post-resection HCC recurrence in the univariate and multivariate RFS analyses (HR (95% CI), 6.176 (2.950 to 12.931), P value < 0.0001; 7.614 (3.299 to 17.573), P value < 0.0001, respectively) (Table 4; Model 3). Additionally, the Kaplan-Meier analysis revealed that patients with the Group 4 combination of pre-S2 gene deletion mutation and miR-369-3p expression level rather than Group 2–3 showed significantly shorter RFS than patients with the Group 1 combination (median RFS (range), 12.60 (1.6 to 158.9) months for Group 4, 34.70 (2.0 to 158.1) months for Group 2–3, 145.20 (26.9 to 159.5) months for Group 1; log-rank P value < 0.0001; Group 4 versus Group 1, adjusted P value < 0.0001; Group 2–3 versus Group 1, adjusted P value = 0.9520; Group 4 versus Group 2–3, adjusted P value < 0.0001) (Figure 2C). When compared with patients with the Group 1–3 combination, patients with the Group 4 combination also displayed significantly shorter RFS (median RFS (range), 12.60 (1.6 to 158.9) versus 68.10 (2.0 to 159.5) months, log-rank P value < 0.0001) (Figure 2D). Better Prognostic Potential of Pre-S2 Gene Deletion Mutation Combined with miR-369-3p Expression Level for Post-Resection HCC Recurrence Than Either Biomarker Alone To evaluate the prognostic potential of pre-S2 gene deletion mutation combined with miR-369-3p expression level for post-resection HCC recurrence, the ROC curve analysis was performed. As shown in Figure 3, the combination of pre-S2 gene deletion mutation and miR-369-3p expression level displayed higher potential in discriminating patients with post-resection HCC recurrence from patients without post-resection HCC recurrence, as indicated by a significantly greater AUC (0.8697, 95% CI 0.7877 to 0.9516), than the pre-S2 gene deletion mutation (AUC (95% CI), 0.8298 (0.7404 to 0.9191); P value = 0.0465) or miR-369-3p expression level (AUC (95% CI), 0.6505 (0.5509 to 0.7501); P value < 0.0001) alone. Moreover, model calibration by the Hosmer–Lemeshow test showed no evidence of poor fit (chi-squared (χ2) value = 0.0672, degrees of freedom (df) = 2, P value = 0.9670), suggesting excellent agreement between the predicted and observed outcomes. In addition, time-dependent C-index was conducted to evaluate the model potential for predicting time-to-event outcomes, indicating that pre-S2 gene deletion mutation combined with miR-369-3p expression level achieved the highest C-index (0.7231, 95% CI 0.649 to 0.798), followed by pre-S2 gene deletion mutation (C-index (95% CI), 0.6895 (0.612 to 0.767) and then miR-369-3p expression level (C-index (95% CI), 0.6162 (0.552 to 0.681) alone. Furthermore, the combination of pre-S2 gene deletion mutation and miR-369-3p expression level had the lowest AIC value (63.182) compared with the pre-S2 gene deletion mutation (67.586) or miR-369-3p expression level (91.720) alone, showing a significant improvement in model fit. Based on Youden index, a T/NT expression ratio of 2.40 was determined as the optimal outcome-based cut-off value for miR-369-3p expression level in the prediction of post-resection HCC recurrence. Using this outcome-based cut-off value for the univariate, multivariate, and Kaplan-Meier RFS analyses yielded highly consistent results with using the 75th percentile T/NT expression ratio (2.55) as cut-off value (Table S4 and Figure S3). Figure 3ROC curves for comparing the potential of pre-S2 gene deletion mutation and/or miR-369-3p expression level in predicting post-resection HCC recurrence. 38 patients with and 32 patients without post-resection HCC recurrence were analyzed. The AUC for pre-S2 gene deletion mutation (solid blue line), miR-369-3p expression level (dotted red line) or a combination of both biomarkers (dotted green line) was shown in the bottom of plot.A line graph with the vertical axis labeled Sensitivity and the horizontal axis labeled 1 minus Specificity. The horizontal axis ranges from 0.00 to 1.00 with labeled ticks at 0.00, 0.25, 0.50, 0.75, 1.00. The vertical axis ranges from 0.00 to 1.00 with labeled ticks at 0.00, 0.25, 0.50, 0.75, 1.00. A diagonal reference line runs from (0.00, 0.00) to (1.00, 1.00). Legend title: AUC. HBV pre-S2 gene deletion mutation (0.8298): (0.00, 0.00), (0.20, 0.82), (1.00, 1.00). miR-369-3p expression level (0.6505): (0.00, 0.00), (0.60, 0.90), (1.00, 1.00). Combination of HBV pre-S2 gene deletion mutation and miR-369-3p level (0.8697): (0.00, 0.00), (0.10, 0.75), (0.70, 0.98), (1.00, 1.00).A line graph showing receiver operating characteristic curves for three biomarkers with area under curve values. ROC curves for comparing the potential of pre-S2 gene deletion mutation and/or miR-369-3p expression level in predicting post-resection HCC recurrence. 38 patients with and 32 patients without post-resection HCC recurrence were analyzed. The AUC for pre-S2 gene deletion mutation (solid blue line), miR-369-3p expression level (dotted red line) or a combination of both biomarkers (dotted green line) was shown in the bottom of plot. Discussion Although surgical resection is established as a curative therapy option for HCC patients, high rate of post-resection HCC recurrence remains a serious problem resulting in poor patient survival.39,40 The presence of HBV pre-S2 gene deletion mutation in liver tissues or blood of HCC patients has been demonstrated as a valuable independent prognostic biomarker for a higher risk of HCC recurrence after curative surgical resection.20,21 This study analyzed a retrospective cohort of patients and discovered that miR-369-3p was an important miRNA whose expression level in tumor tissues was negatively associated with not only pre-S2 gene deletion mutation but also post-resection HCC recurrence. Patients with the presence of pre-S2 gene deletion mutation combined with low expression level of miR-369-3p had a higher risk of post-resection HCC recurrence than patients with either or none of both biomarkers. Moreover, the combination of pre-S2 gene deletion mutation and miR-369-3p expression level displayed better prognostic potential than each biomarker alone in predicting post-resection HCC recurrence. Several studies have evaluated the role of miR-369-3p in the tumorigenesis of different types of cancers. It has been shown that miR-369-3p is downregulated in tumor tissues of thyroid cancer patients and low expression of miR-369-3p is associated with shorter overall survival.41 Overexpression of miR-369-3p suppresses proliferation and promotes apoptosis of human thyroid cancer cells through inhibiting the expression of its target tetraspanin 13 (TSPAN13).41 In human lung cancer cells, miR-369-3p plays an important role in modulating drug resistance against the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor osimertinib or the chemotherapeutic drug cisplatin through direct regulation of protein tyrosine phosphatase non-receptor type 12 (PTPN12) or solute carrier family 35 member F5 (SLC35F5) expression, respectively.42,43 Fibroblast-secreted exosomal miR-369-3p promotes migration and invasion of human lung cancer cells through regulating the expression of its target neurofibromin-1 (NF1).44 In addition, the expression level of miR-369-3p has been found to be decreased in tumor tissues of colon cancer patients.45 Overexpression of miR-369-3p inhibits proliferation, invasion, and oxidative stress of human colorectal cancer cells by targeting transcription factor 4 (TCF4).45 In gastric cancer patients, decreased expression of miR-369-3p in tumor tissues is correlated with tumor progression.46 Overexpression of miR-369-3p suppresses proliferation, migration, and invasion of human gastric cancer cells; in contrast, inhibition of miR-369-3p abrogates the effect of the chemotherapeutic drug azacytidine on cell growth and invasion.46 Moreover, downregulation of miR-369-3p has been observed in tumor tissues of HCC patients and is associated with tumor progression, lymph node metastasis, and poor overall and disease-free survival.47–50 Overexpression of miR-369-3p inhibits viability, proliferation, migration, invasion, and metastasis of human HCC cells through targeted suppression of paired box 6 (PAX6), SRY-box transcription factor 4 (SOX4), Zinc finger E-box binding homeobox 1 (ZEB1), or actinin alpha 4 (ACTN4).47–50 Although these studies show a negative association of miR-369-3p with cancer progression and prognosis, the results of this study revealed that miR-369-3p expression level alone had weak discriminatory potential for post-resection HCC recurrence, suggesting that miR-369-3p is likely part of a panel of biomarkers rather than a primary prognostic driver on its own. This weak discriminatory potential may be partially attributed to the compensation by other miRNAs or regulatory pathways in response to the downregulation of miR-369-3p expression to keep target gene expression in check, mitigating the overall difference between diseased and healthy states. In addition, this study identified a higher risk of post-resection HCC recurrence in patients with the presence of pre-S2 gene deletion mutation combined with low miR-369-3p expression level. Considering that miR-369-3p functions as a tumor suppressor and holds potential as a therapeutic target for cancer, it is worth further investigating the promise of pharmacological targeting of miR-369-3p in the prevention and treatment of pre-S2 mutant-associated HCC recurrence after curative surgical resection. The expression of miR-369-3p has been shown to be regulated by several long non-coding RNAs (lncRNAs) or circular RNAs (circRNAs). Overexpression of lncRNA EGF like and EMI domain containing 1 (EGFEM1P) promotes proliferation and inhibits apoptosis of human thyroid cancer cells via acting as a miR-369-3p sponge to upregulate TCF4 expression.51 LncRNA SOX2 overlapping transcript (SOX2-OT) overexpression accelerates proliferation and migration of human prostate cancer cells by functioning as a miR-369-3p sponge to elevate cofilin 2 (CFL2) expression.52 Silencing of lncRNA deleted in lymphocytic leukemia 2 (DLEU2) inhibits transforming growth factor-beta 1 (TGF-β1)-induced proliferation and migration of human lung cancer cells by upregulating miR-369-3p expression and downregulating tripartite motif containing 2 (TRIM2) expression.53 Overexpression of lncRNA musculin antisense RNA 1 (MSC-AS1) promotes proliferation and suppresses apoptosis of human osteoarthritic chondrocytes by regulating the miR-369-3p/Janus kinase 2 (JAK2)/signal transducers and activators of transcription 3 (STAT3) signaling pathway.54 LncRNA OIP5 antisense RNA 1 (OIP5-AS1) plays a crucial role in regulating radioresistance of colorectal cancer cells through targeting miR-369-3p to upregulate dual specificity tyrosine phosphorylation regulated kinase 1A (DYRK1A) expression.55 Additionally, overexpression of circRNA hsa_circ_0000423 facilitates proliferation, migration, invasion, and immune escape of human colorectal cancer cells through competitive binding of miR-369-3p to increase cyclin D1 (CCND1) expression.56 Overexpression of circRNA hsa_circ_0007042 ameliorates intervertebral disc degeneration by adsorbing miR-369-3p to increase bone morphogenetic protein 2 (BMP2) expression and activate the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt) signaling pathway.57 Moreover, some natural compounds (such as quercetin and tetramethylpyrazine) and environmental pollutants (such as cadmium) have been found to exhibit their effects through regulating miR-369-3p expression.58–60 Further studies to clarify the regulation of miR-369-3p by pre-S2 mutant and explore the underlying molecular mechanisms hold potential to provide new insights into the oncogenic role of pre-S2 mutant in HBV-related HCC development. In addition, this study only analyzed a single-center retrospective small cohort of patients and lacks additional independent external cohort of patients for validation. Although the sample size analyzed in this study reaches sufficient statistical power and the clinicopathological profile of patients recruited in this study appropriately corresponds to the representative profile of a large patient population in Taiwan,61 small sample size may undermine the statistical power and wide 95% CIs for HRs could exaggerate the results. In addition, the absence of independent validation may reduce the reproducibility and generalization of the data. Therefore, a large cohort of patients enrolled at different medical centers are needed to further confirm the preliminary findings of this study and improve their translation into clinical practice. In conclusion, this single-center retrospective discovery study identified a negative association between pre-S2 gene deletion mutation and miR-369-3p expression level and provided preliminary evidence suggesting the combination of these two factors as a novel combination biomarker which had a superior prognostic potential in predicting HBV-related HCC recurrence after curative surgical resection. Further external validation studies are necessary to validate the independent prognostic performance of this combination biomarker.