Serum 5’-tiRNA-iMet-CAT is a novel biomarker for the diagnosis and early postoperative assessment of hepatocellular carcinoma

Abstract

Background

Hepatocellular carcinoma (HCC) is one of the most common digestive system malignancies and a leading cause of cancer-related mortality worldwide. Alpha-fetoprotein (AFP), although widely used, shows limited sensitivity in early HCC, leading to missed diagnoses. Novel molecular markers that complement existing biomarkers are urgently needed. tRNA-derived small RNAs (tsRNAs), a newly recognized class of noncoding RNAs generated from transfer RNAs, are abundant in cancer cells and body fluids and have emerging roles in tumourigenesis. This study aimed to identify dysregulated tsRNAs in HCC and evaluate their clinical value as biomarkers.

Methods

We performed high-throughput sequencing on 3 pairs of HCC tissues and their matched adjacent normal tissues to profile 5’-tiRNA-iMet-CAT, validating its serum levels by quantitative reverse transcription PCR (qRT-PCR). Molecular features and stability were confirmed by electrophoresis, sequencing, and freeze-thaw assays. We assessed clinical correlations, diagnostic potential, and predicted target genes via multiple databases with Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis.

Results

Serum 5’-tiRNA-iMet-CAT was significantly elevated in HCC patients versus healthy controls. Its high expression correlated with poor differentiation, larger tumour size, and advanced Barcelona Clinic Liver Cancer stage (all P < 0.05). Receiver operating characteristic analysis confirmed strong diagnostic performance, with levels markedly decreasing postoperatively (P < 0.0001).

Conclusion

Serum 5’-tiRNA-iMet-CAT is stable, specific, and clinically relevant. It may serve as a promising noninvasive biomarker for HCC diagnosis and evaluating short-term therapeutic response during the early postoperative phase.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12957-026-04480-7.

Introduction

Liver cancer is one of the leading causes of cancer-related mortality worldwide. Among its subtypes, hepatocellular carcinoma (HCC), a malignant tumour originating from hepatocytes, accounts for approximately 75% to 80% of primary liver cancers [1]. The development of HCC is believed to involve the dedifferentiation of mature hepatocytes into hepatic progenitor cells [2]. The key contributors to HCC development include chronic hepatitis B and C infections, along with alcoholic and nonalcoholic fatty liver disease [3]. In addition, aflatoxin exposure, prolonged heavy alcohol consumption, tobacco smoking, diabetes, and obesity have also been identified as potential contributors to HCC pathogenesis [4, 5]. Despite considerable progress in the prevention and treatment of HCC over the past decades, which has moderately improved patient survival [6], HCC remains the most common primary liver cancer and a major contributor to cancer-related mortality worldwide. Early detection and accurate diagnosis are therefore crucial for improving overall patient outcomes. However, current diagnostic approaches still present significant limitations. Imaging techniques, including magnetic resonance imaging (MRI) and computed tomography (CT), play important roles in HCC detection but show markedly reduced sensitivity for detecting tumours smaller than 2 cm, resulting in a high rate of missed early diagnoses [7, 8]. Moreover, conventional serum biomarkers such as Alpha-fetoprotein (AFP) are widely used in clinical practice and provide certain prognostic information; however, their sensitivity for early-stage HCC is limited, and their specificity can be inconsistent, leading to high false-negative rates and restricting their utility in screening and early diagnosis [9]. This situation causes many patients to miss the optimal window for treatment. Consequently, there is an urgent need to identify novel tumour biomarkers that could complement existing tools and, when combined with traditional markers, increase the sensitivity and specificity of early HCC detection.

Noncoding RNAs (ncRNAs), which do not encode proteins, play essential roles in regulating gene expression and cellular physiology [10]. Their aberrant expression is linked to cancer, and many, such as microRNAs, are stable in biofluids, highlighting their potential as diagnostic biomarkers [11–13]. A recently recognized and functionally important subclass of small ncRNAs is tRNA-derived small RNAs (tsRNAs), also called tRNA-derived RNAs (tDRs) or tRNA fragments (tRFs) [14], which can be classified into two main types according to their length and cleavage sites. The first type, tRNA-derived stress-induced RNAs (tiRNAs), are 28–36 nucleotides (nts) in length and result from specific cleavage within the anticodon loop of mature tRNAs, typically under stress conditions. tRFs, representing the second type, are short (14–30 nts) RNA fragments produced from precursor or mature tRNAs. Two tiRNA subtypes, 5′-tiRNAs and 3′-tiRNAs, have been identified, both of which are generated via cleavage within or adjacent to the anticodon loop [15]. Angiogenin, a member of the RNase A superfamily, can cleave tRNAs in vitro and in vivo to produce tiRNAs [16]. tRFs, in contrast, are classified into four subtypes on the basis of their origin within the tRNA molecule: 5′-tRF, 3′-tRF, 1-tRF, and 2-tRF [15]. 5′-tRFs and 3′-tRFs are derived from the respective ends of mature tRNAs and are typically generated under stress conditions by nucleases such as angiogenin (ANG) and RNase T2 [17], with cleavage sites located within the D arm and TψC arm, respectively [18]. Moreover, 1-tRFs and 2-tRFs arise from the 5′ leader and 3′ trailer sequences of precursor tRNAs, respectively, through endonucleolytic cleavage mediated by RNase P, RNase Z, and related enzymes during tRNA maturation [19].

In recent years, tsRNAs have attracted considerable attention for their potential roles in cancer diagnosis. Increasing evidence supports their diagnostic value in multiple cancer types, including pancreatic, lung, breast, and gastric cancers [20–23]. For example, Meilin Xue et al. reported that the combined detection of tsRNA-ValTAC-41, tsRNA-MetCAT-37, and tsRNA-ThrTGT-23 markedly improved the sensitivity and specificity for diagnosing pancreatic ductal adenocarcinoma [24]. Similarly, Li et al. identified tRF-29-R9J8909NF5JP as a potential biomarker for gastric cancer [25]. In addition to diagnostics, mechanistic studies have shown that dysregulated tsRNAs can modulate cancer progression. Veronica Balatti et al. suggested that aberrant tsRNAs may interfere with gene expression regulation by associating with Argonaute proteins, thereby contributing to tumourigenesis [26]. In colorectal cancer (CRC), tsRNA-GlyGCC has been reported to promote tumour growth and resistance to 5-fluorouracil (5-FU) by activating SPIB and regulating the JAK1/STAT6 signalling pathway [27]. Furthermore, LeuCAG3′ tsRNA enhances the translation of RPS28 and RPS15 mRNAs, encoding the ribosomal proteins S28 and S15, respectively, to sustain ribosome biogenesis, promote tumour cell proliferation and survival, and thus represents a potential therapeutic target [28].

In this study, high-throughput sequencing revealed significant upregulation of 5’-tiRNA-iMet-CAT in HCC tissues. Its expression was further validated in serum samples from 139 HCC patients, 55 hepatitis patients, and 112 healthy donors. The results demonstrated that 5’-tiRNA-iMet-CAT could effectively distinguish HCC patients from healthy individuals. Clinical correlation analyses and evaluation of its diagnostic performance further supported its potential as a novel biomarker for HCC.

Materials and methods

Clinical serum samples

All the clinical specimens were collected at the Affiliated Hospital of Nantong University and included 139 HCC patients, 55 hepatitis patients, and 112 healthy controls. The primary inclusion criteria for all HCC cases were as follows: (1) pathologically confirmed HCC; (2) availability of qualified preoperative serum samples; (3) no prior anticancer treatments (e.g., radiotherapy, chemotherapy, or transarterial chemoembolization) before surgery; and (4) availability of complete clinicopathological and follow-up data. For the HCC cohort, we systematically collected the following detailed clinicopathological variables: age, sex, hepatitis B virus infection status, cirrhosis, and comprehensive tumor characteristics. The tumor characteristics included tumor size, tumor number, metastasis status, microvascular invasion, histological differentiation grade, and tumor stage according to both the tumor-node-metastasis (TNM) and Barcelona Clinic Liver Cancer (BCLC) staging systems (Table 1). For comparison with the novel biomarker, serum levels of the conventional hepatocellular carcinoma markers alpha-fetoprotein (AFP) and protein induced by vitamin K absence or antagonist-II (PIVKA-II) were analyzed. Critically, all measurements (AFP, PIVKA-II, and 5’-tiRNA-iMet-CAT) were performed on the same batch of preoperative serum samples, ensuring identical timing and sample handling conditions. AFP and PIVKA-II levels were quantitatively determined in the clinical laboratory of our hospital using standardized electrochemiluminescence immunoassays (ECLIA) on a Roche Cobas e801 analyzer (Roche Diagnostics), strictly following the manufacturer’s protocols and the laboratory’s standard operating procedures. Serum samples were collected after an overnight fast and collected in RNase- and DNase-free EP tubes and immediately stored at − 80 °C. This study was approved by the Human Research Ethics Committee of the Affiliated Hospital of Nantong University (Ethics approval number: No. 2018-L055), and written informed consent was obtained from all participants.

Table 1.

Baseline clinicopathological characteristics of the HCC patient cohort (N = 139)

Characteristic	Category	Overall Cohort, n (%) or Median (Range)
Sex	Male / Female	96 (69.1) / 43 (30.9)
Age	—	median age: 69 (37–86)
HBV infection	Positive / Negative	69 (49.6) / 70 (50.4)
Liver cirrhosis	Presence / Absence	94 (67.6) / 45 (32.4)
Tumor size	< 5 cm / ≥5 cm	92 (66.2) / 47 (33.8)
Metastasis	Presence / Absence	57 (41.0) / 82 (59.0)
Histological differentiation	Well / Poor-Moderate	53 (38.1) / 86 (61.9)
TNM stage	I-II / III-IV	75 (54.0) / 64 (46.0)
BCLC stage	0-A / B-D	72 (51.8) / 67 (48.2)
Tumor number	Single / Multiple	122 (87.8) / 17(12.2)
Microvascular invasion	Presence / Absence	20 (14.4) / 119 (85.6)

High-throughput sequencing

Total RNA was extracted via TRIzol reagent, and RNA quality and concentration were assessed via a Qubit^® 2.0 fluorometer and an Agilent 2200 TapeStation system, respectively. Small RNA libraries were constructed using 1 µg of total RNA with the NEBNext^® Multiplex Small RNA Library Prep Set for Illumina, following the manufacturer’s instructions. Sequencing was performed on the Illumina HiSeq 2500 SE50 platform. Clean reads obtained were mapped to the MINTbase database via MINTmap software; unmapped reads were subjected to novel tsRNA prediction. Differential expression analysis of the tsRNAs between groups was conducted via the EdgeR software package. The raw sequencing data have been deposited in the NCBI Gene Expression Omnibus (GEO) database under accession number GSE282472. (The data can be accessed via https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE282472)

Cell culture

Human HCC cell lines (Huh7, HCCLM3, Hep3B, MHCC97-H, PLC-PRF-5, and SNU-182) and the normal human liver cell line THLE-2 were purchased from the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). All the cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Corning, Manassas, VA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, Waltham, MA, USA) and 1% penicillin‒streptomycin (HyClone, Logan, UT, USA) at 37 °C in a humidified incubator with 5% CO₂.

Total RNa extraction, cDNA synthesis, and qRT‒PCR

Total RNA was extracted from the cell samples via TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The RNA was purified from the serum samples via a spin column–based kit (BioTeke, Wuxi, Jiangsu, China). 5’-tiRNA-iMet-CAT was selected as the target molecule, with RUN6B (U6) serving as the internal reference. Reverse transcription was performed via the RevertAid RT Kit (Thermo Fisher Scientific, Waltham, MA, USA) and specific primers (Ribobio Corporation, Guangzhou, China) in a 10 µL reaction mixture containing primers for 5’-tiRNA-iMet-CAT and U6. The mixture was incubated at 42 °C for 1 h, followed by enzyme inactivation at 70 °C for 5 min, after which it was stored at 4 °C–− 20 °C for long-term preservation. Subsequently, quantitative real-time PCR (qRT‒PCR) was performed on an ABI QuantStudio 5 system via a 20-µL reaction mixture containing 10 µL of ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd., China), 5 µL of cDNA, 1 µL of forward primer, 1 µL of reverse primer (Ribobio Corporation), and 3 µL of nuclease-free water. The expression levels of 5’-tiRNA-iMet-CAT were calculated via the 2^−ΔΔCT method, where ΔCT = Ct(target) − Ct(reference) and ΔΔCT = ΔCT (experimental) – ΔCT (control). The procedures were conducted in strict accordance with the manufacturer’s instructions.

Serial dilution assay

RNA was extracted from randomly selected HCC cells and reverse-transcribed. The resulting cDNA was serially diluted 10, 10², 10³, 10⁴, and 10⁵ times, followed by qRT‒PCR to assess amplification efficiency.

Room temperature incubation and repeated freeze‒thaw stability assays

Twenty serum samples were randomly mixed and divided equally into 20 aliquots. Aliquots from one set were incubated at 25°C for time intervals of 0, 6, 12, 18, and 24 hours. The remaining aliquots were subjected to successive freeze–thaw cycles (0, 1, 3, 5, or 10 cycles) alternating between − 80°C and room temperature. RNA was extracted from each aliquot, and 5’-tiRNA-iMet-CAT expression levels were quantified via qRT‒PCR to evaluate stability.

Bioinformatics analysis

Putative target mRNAs of 5’-tiRNA-iMet-CAT were predicted using four algorithms: miRanda (default parameters, total score ≥ 90 and total free energy < − 17 kcal/mol), RNAhybrid (minimum free energy < − 21 kcal/mol), PITA (ddG < − 10 kcal/mol), and TargetScan (default settings). The 3’ UTR sequences of target genes were extracted from the human reference genome hg38. The intersection of the four prediction results was retained as high‑confidence target genes. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using clusterProfiler (v4.7.1.3) with all annotated human genes as background. Significance was determined by Fisher’s exact test with a raw P < 0.05, and the false discovery rate (FDR) was controlled using the Benjamini–Hochberg method (reported as Q‑value). Enriched terms with Q < 0.05 were considered significant, and the top 30 terms ranked by P value were visualized using ggplot2 (v3.4.4).

Statistical analysis

Statistical analyses were performed using SPSS 26.0 (IBM SPSS Statistics, Chicago, USA) and GraphPad Prism 9.0 (GraphPad Software, CA, USA). Normality was assessed by Shapiro–Wilk test. For comparisons of two groups of continuous variables, Student’s t-test was used for normally distributed data and Mann–Whitney U test for non‑normally distributed data. For multiple group comparisons, one‑way analysis of variance (ANOVA) with Bonferroni post‑hoc correction was applied for normally distributed data, and Kruskal–Wallis test for non‑normally distributed data. Categorical variables were analyzed by chi‑square test or Fisher’s exact test. Receiver operating characteristic (ROC) curves were generated, areas under the curve (AUC) were compared by DeLong’s test, and the optimal cut‑off value was determined by the Youden index. Survival curves were estimated by the Kaplan–Meier method and compared by the log‑rank test. Multivariable Cox proportional hazards regression (enter method) was performed, including sex, age, HBV status, cirrhosis, BCLC stage, tumour size, metastasis, and 5’‑tiRNA‑iMet‑CAT as covariates. All categorical variables were entered with the first category as reference. Variables with potential prognostic relevance in clinical practice were incorporated into the multivariate model. Differentially expressed tsRNAs from sequencing data were identified using EdgeR with thresholds of log₂-fold change > 1 and P < 0.05. GO and KEGG enrichment analyses were performed using clusterProfiler, with a significance threshold of P < 0.05 and Benjamini–Hochberg correction. All experiments were conducted in triplicate, and results are expressed as mean ± standard deviation (SD). Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Results

Expression levels of 5’-tiRNA-iMet-CAT in serum and cell lines

To identify differentially expressed tsRNAs in HCC tissues, we performed high-throughput sequencing on 3 pairs of HCC tissues and their matched adjacent normal tissues (Fig. 1a, b). We further screened three significantly upregulated tsRNAs—tRF-17-EXEY0V5, tRF-17-EBXNWD5, and tRF-31-FP18LPMBQ4NKD (log₂-fold change > 1, p < 0.05) (Fig. 1c). Next, serum samples from 20 HCC patients were collected, and qRT‒PCR was performed to validate the differences in expression, revealing that tRF-31-FP18LPMBQ4NKD was significantly elevated in the serum of HCC patients (Fig. 1d). Additionally, compared with that in the normal liver cell line THLE-2, the expression of 5’-tiRNA-iMet-CAT in six HCC cell lines was significantly greater, especially in MHCC-97 H, PLC-PRF-5, and HCCLM3 cells (Fig. 1e). Therefore, tRF-31-FP18LPMBQ4NKD was selected for further comprehensive studies.

Fig. 1.

Expression levels of 5’-tiRNA-iMet-CAT in serum and cell lines. a Heatmap of differential tsRNAs in three pairs of tumour and matched adjacent normal tissues. b Volcano map of the tsRNAs in three pairs of tumour and matched adjacent normal tissues. c Three upregulated genes identified via tsRNA-seq in HCC vs. matched adjacent normal tissues (log2-fold change > 1, P < 0.05). d Relative expression of three highly expressed tsRNAs in HCC patients (n = 20) and healthy donors (n = 20). e Statistical significance of the expression of 5’-tiRNA-iMet-CAT in HCC cells and normal human THLE-2 cells (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05); ns, no significance

Database analysis confirmed that 5’-tiRNA-iMet-CAT is a tRNA-derived fragment (tiRNA)

Using the University of California, Santa Cruz (UCSC) Genome Browser, we determined that tRF-31-FP18LPMBQ4NKD originates from chromosome 6 (chr6: 26,286,754–26,286,784), with a length of 31 bp (Fig. 2a). Data from MINTbase v2.0 (https://cm.jefferson.edu/MINTbase/) further indicated that tRF-31-FP18LPMBQ4NKD is a 31-nucleotide 5′-tiRNA (5′-AGCAGAGTGGCGCAGCGGAAGCGTGCTGGGC-3′) derived from tRNA-Met-CAT (Fig. 2b, c). The molecule was named 5’-tiRNA-iMet-CAT according to the MINTbase nomenclature rules. Agarose gel electrophoresis (AGE) of the qRT‒PCR products revealed a single distinct band of approximately 80 bp (Fig. 2d). Sanger sequencing confirmed that the product contained the full-length 5’-tiRNA-iMet-CAT sequence, which was consistent with the annotation in MINTbase v2.0 (Fig. 2e).

Fig. 2.

Database analysis confirming that 5’-tiRNA-iMet-CAT is a tRNA-derived fragment (tiRNA). a 5’-tiRNA-iMet-CAT is located on Chr6 with coordinates 26,286,754–26,286,784. b MINTbase analysis revealed that the length of 5’-tiRNA-iMet-CAT was 31 nt. c The cleavage site of 5’-tiRNA-iMet-CAT was located above the anticodon loop in the Transfer RNA Database. d The product of qRT‒PCR was a single electrophoresis band of approximately 80 bp in size on a 2.5% agarose gel. e qRT‒PCR products containing the primer sequence for 5’-tiRNA-iMet-CAT were subjected to Sanger sequencing

Methodological evaluation of 5’-tiRNA-iMet-CAT detection

To evaluate the clinical potential of 5’-tiRNA-iMet-CAT, we conducted a systematic methodological validation. The assay demonstrated excellent precision, with intra-assay and interassay coefficients of variation of 1.57% and 1.93%, respectively (Table 2). Stability tests showed that 5’-tiRNA-iMet-CAT expression levels remained stable in serum samples incubated at 25 °C for 24 h or subjected to multiple freeze-thaw cycles (Fig. 3a, b). Specificity validation revealed distinct amplification profiles and single-peak melting curves (Fig. 3c, d). Serial dilution experiments confirmed strong linear correlations for both the target and reference genes (Fig. 3e, f). Furthermore, no significant differences in 5’-tiRNA-iMet-CAT expression were found in patients with breast, lung, or thyroid cancer compared to healthy donors (Fig. 3g), whereas levels were significantly elevated in HCC patients, indicating good specificity for hepatocellular carcinoma. Collectively, these results demonstrate that 5’-tiRNA-iMet-CAT possesses ideal characteristics as a biomarker for HCC.

Table 2.

Intra-assay and interassay coefficients of variation of 5’-tiRNA-iMet-CAT

	5’-tiRNA-iMet-CAT	U6
Intra assay
Mean ± SD	17.443 ± 0.275	20.590 ± 0.382
CV(%)	1.57	1.85
Inter assay
Mean ± SD	18.748 ± 0.346	21.926 ± 0.431
CV(%)	1.93	2.11

CV Coefficient of variation, SD Standard deviation

Fig. 3.

Methodological evaluation of 5’-tiRNA-iMet-CAT detection. a, b Stability of 5’-tiRNA-iMet-CAT during incubation at room temperature and repeated freeze–thawing cycles. c, d Amplification curve and melting curve of 5’-tiRNA-iMet-CAT. e, f Linearity of 5’-tiRNA-iMet-CAT and U6. g The expression level of 5’-tiRNA-iMet-CAT was not significantly different in the other tumours. *Indicates statistical significance (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05); ns, no significance

Clinical significance and prognostic value of 5’-tiRNA-iMet-CAT in HCC serum

We collected serum samples from 139 HCC patients, 55 hepatitis patients, and 112 healthy individuals to evaluate the expression of 5’-tiRNA-iMet-CAT. qRT‒PCR analysis revealed that serum 5’-tiRNA-iMet-CAT levels were significantly higher in HCC patients than in both healthy controls (P < 0.0001) and hepatitis patients (P < 0.0001), whereas hepatitis patients presented slightly higher levels than healthy controls (Fig. 4a). These results suggest that 5’-tiRNA-iMet-CAT has discriminative power for separating HCC patients from those with hepatitis and from normal donors. On the basis of the median value of expression (3.135), HCC patients were classified into a high-expression group (> 3.135, n = 69) and a low-expression group (≤ 3.135, n = 70). Chi-square analysis demonstrated that higher 5’-tiRNA-iMet-CAT expression was significantly associated with poor differentiation (P < 0.0001), advanced BCLC stage (P = 0.0009), advanced TNM stage (P < 0.0001), the presence of metastasis (P = 0.0098), and larger tumour size (P = 0.0072) but not with sex, age, cirrhosis, or HBV infection (Table 3). These findings indicate that elevated 5’-tiRNA-iMet-CAT expression may reflect aggressive tumour progression. To investigate whether serum 5’-tiRNA-iMet-CAT levels are associated with tumour burden and to preliminarily explore its early postoperative changes, we analysed its early dynamic changes after tumour resection. We collected paired pre- and postoperative serum samples from 41 HCC patients (with postoperative samples typically drawn within days to weeks after surgery). qRT‒PCR analysis revealed that the serum 5’-tiRNA-iMet-CAT level decreased in most patients after surgery (Fig. 4b), but the difference from that in healthy donors was not statistically significant (P = 0.0566) (Fig. 4c).

Fig. 4.

Clinical significance and prognostic value of 5’-tiRNA-iMet-CAT in HCC serum. a Expression levels of 5’-tiRNA-iMet-CAT were detected in the serum of HCC patients (n = 139), hepatitis patients (n = 55), and healthy donors (n = 112). b Pre- and postoperative serum 5’-tiRNA-iMet-CAT expression levels in 41 HCC patients. c 5’-tiRNA-iMet-CAT expression levels in serum from postoperative HCC patients and healthy donors. d In 53 well-differentiated and 86 poorly differentiated/moderately differentiated HCC patients, 5’-tiRNA-iMet-CAT expression levels were compared. e 5’-tiRNA-iMet-CAT serum expression levels in HCC patients stratified by tumour size. f HCC patient serum 5’-tiRNA-iMet-CAT expression levels in 57 patients with and 82 patients without metastases. g 5’-tiRNA-iMet-CAT expression levels in HCC patients and healthy donors with different degrees of tumour invasion. (T1-T2: n = 75; T3–T4: n = 64; healthy donors: n = 112). h 5’-tiRNA-iMet-CAT expression levels in HCC patients and healthy donors with different Barcelona Clinic Liver Cancer subtypes. (0/A: n = 72; C/D: n = 67; healthy donors: n = 112). i Kaplan-Meier curve analysis of the relationship between the expression level of 5’-tiRNA-iMet-CAT and the survival of HCC patients. *Indicates statistical significance (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05); ns, no significance

Table 3.

Clinicopathological analysis of 5’-tiRNA-iMet-CAT

Parameter	No. of patients	high	low	P value
Sex
Male	96	43	53	0.1008
Female	43	26	17
Age(year)
≤ 64	54	32	22	0.0706
> 64	85	37	48
HBV infection
Negative	70	37	33	0.4449
Positive	69	32	37
Cirrhosis of the liver
Absence	45	22	23	0.9024
Presence	94	47	47
Tumour size, cm
<5	92	38	54	0.0072**
≥ 5	47	31	16
Metastasis
Negative	82	33	49	0.0098**
Positive	57	36	21
Histological differentiation
Well	53	10	43	< 0.0001****
Poor/Moderate	86	59	27
TNM stage
Ⅰ-Ⅱ	75	24	51	0.0001***
Ⅲ-Ⅳ	64	45	19
BCLC stage
0, A,	72	26	46	0.0009***
B, C, D	67	43	24

^*p< 0.05, ^**p< 0.01, and^***p< 0.001 were considered significant

We subsequently stratified HCC patients on the basis of clinicopathological parameters that were significantly associated with each other and compared the serum 5’-tiRNA-iMet-CAT levels across these groups. Patients with poorly differentiated tumours presented significantly higher 5’-tiRNA-iMet-CAT expression than did those with moderately differentiated or well-differentiated tumours (Fig. 4d). The serum 5’-tiRNA-iMet-CAT levels also increased markedly with increasing tumour size (Fig. 4e), and patients with tumours ≥ 5 cm in size presented greater expression than did those with tumours < 5 cm in size. Among these patients, the expression levels were slightly higher in those with metastasis compared to those without ((Fig. 4f). Patients were categorized into stages I–II or III–IV according to their TNM stage. The expression levels of 5’-tiRNA-iMet-CAT differed significantly between the two groups (P < 0.0001), and both groups presented significantly greater expression than healthy controls did (P < 0.0001) (Fig. 4g). Similarly, patients with BCLC stage B/C/D disease presented higher expression than patients with stage 0/A disease did (P = 0.028), with both groups exceeding the levels observed in healthy controls (P < 0.0001) (Fig. 4h). The Kaplan–Meier analysis indicated that patients with high levels of preoperative serum 5’-tiRNA-iMet-CAT had a lower survival rate than those with low levels (Fig. 4i). To further determine whether 5’-tiRNA-iMet-CAT is an independent prognostic factor, a multivariable Cox proportional hazards regression analysis was performed, adjusting for key clinical confounders including sex, age, HBV status, cirrhosis, BCLC stage, metastasis, and tumor size. As shown in Table 4, a high level of 5’-tiRNA-iMet-CAT remained significantly associated with worse overall survival (Hazard Ratio [HR] = 1.887, 95% Confidence Interval [CI]: 1.141–3.121, P = 0.013). A total of 75 deaths (46.0% censoring rate) occurred among the 139 HCC patients during follow-up.

Table 4.

Multivariable Cox regression analysis of factors associated with overall survival in HCC patients

Parameters	HR (95% CI)	P value
Sex	1.127 (0.677–1.877)	0.645
Age	0.992 (0.968–1.016)	0.488
HBV	0.866 (0.479–1.563)	0.632
Cirrhosis	1.050 (0.565–1.949)	0.878
BCLC Stage	1.157 (0.578–2.316)	0.681
Metastasis	1.038 (0.561–1.919)	0.905
Tumor Size	1.030 (0.556–1.908)	0.926
5’-tiRNA-iMet-CAT	1.887 (1.141–3.121)	0.013*

HR Hazard ratio

^*p< 0.05 was considered significant

These findings indicate that elevated serum 5’-tiRNA-iMet-CAT could serve as a promising biomarker for early-stage HCC, offering potential for timely and accurate detection.

Diagnostic performance of serum 5’-tiRNA-iMet-CAT in HCC

Given the limitations of commonly used HCC biomarkers, such as alpha-fetoprotein (AFP) and protein induced by vitamin K absence or antagonist-II (PIVKA-II), we assessed the potential diagnostic value of 5’-tiRNA-iMet-CAT. ROC curve analysis was performed on the serum data from 139 HCC patients and 112 healthy controls to compare the diagnostic performance of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II. Compared with AFP [0.715 (95% CI: 0.652–0.777)] and PIVKA-II [0.803 (95% CI: 0.750–0.856)], 5’-tiRNA-iMet-CAT demonstrated superior diagnostic performance, with an AUC of 0.909 (95% confidence interval (CI): 0.876–0.943). (Fig. 5a). We further evaluated the diagnostic efficacy of these biomarkers in combination. As shown in Fig. 5b, combining two improved the AUC compared with each marker alone. Notably, the combination of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II achieved the highest diagnostic efficiency, with an AUC of 0.960 (95% CI: 0.940–0.980). We subsequently compared single-marker and combined-marker models on the basis of sensitivity (SEN), specificity (SPE), overall accuracy (ACCU), positive predictive value (PPV), and negative predictive value (NPV). The Youden index is a commonly used measure for assessing the effectiveness of diagnostic tests in binary classification. It is calculated as the Youden index = sensitivity + specificity – 1 and serves to identify the optimal cut-off point. A higher value indicates a better trade-off between sensitivity and specificity, with the maximum value representing the threshold at which the test achieves the most balanced and accurate diagnostic performance. For 5’-tiRNA-iMet-CAT, the optimal cut-off value was 1.678090, yielding the highest Youden index (0.652), with an SEN of 84.9%, an ACCU of 82.9%, and an NPV of 81.1%, outperforming AFP and PIVKA-II (Table 5). These findings indicate that 5’-tiRNA-iMet-CAT could act as a novel biomarker for HCC, with its diagnostic value further improved when used alongside established tumour markers.

Fig. 5.

Diagnostic performance of serum 5’-tiRNA-iMet-CAT in HCC. a ROC curve analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II in the independent diagnosis of HCC patients and healthy donors. b ROC curve analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II in the joint diagnosis of HCC patients and healthy donors. c ROC analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II independent of patients with early HCC and healthy donors. d ROC analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II in the joints of patients with early HCC and healthy donors. e ROC analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II in the independent diagnosis of patients with HCC and hepatitis. f ROC analysis of 5’-tiRNA-iMet-CAT, AFP, and PIVKA-II in the joint diagnosis of patients with HCC and hepatitis. g AFP/PIVKA-II double-negative HCC subgroup. h ROC analysis of serum 5’-tiRNA-iMet-CAT in AFP-negative HCC patients. i ROC analysis of serum 5’-tiRNA-iMet-CAT in PIVKA-II–negative HCC patients. j ROC analysis of 5’-tiRNA-iMet-CAT of patients with early-stage HCC and hepatitis. (Supplementary Figure S1) ROC curve analysis of serum 5’-tiRNA-iMet-CAT for differentiating early-stage hepatocellular carcinoma from hepatitis

Table 5.

Diagnostic value of serum 5’-tiRNA-iMet-CAT, serum AFP and PIVKA-II for HCC

Biomarker	SEN (%)	SPE (%)	ACCU (%)	PPV (%)	NPV (%)
5’-tiRNA-iMet-CAT	84.9(118/139)	80.4(90/112)	82.9(208/251)	84.3(118/140)	81.1(90/111)
AFP	44.6(62/139)	82.1(96/112)	61.4(154/251)	75.6(62/82)	54.4(92/169)
PIVKA-II	54.0(75/139)	89.3(100/112)	69.7(175/251)	86.2(75/87)	61.0(100/164)
5’-tiRNA-iMet-CAT + AFP	92.1(128/139)	73.2(82/112)	83.7(210/251)	81.0(128/158)	88.2(82/93)
5’-tiRNA-iMet-CAT+ PIVKA-II	92.8(129/139)	74.1(83/112)	84.5(212/251)	81.6(129/158)	89.2(83/93)
5’-tiRNA-iMet-CAT + AFP+PIVKA-II	95.0(132/139)	67.9(76/112)	82.9(208/251)	78.6(132/168)	91.6(76/83)

SEN Sensitivity, SPE Specificity, ACCU Overall accuracy, PPV Positive predictive value, NPV Negative predictive value

Owing to the absence of highly sensitive biomarkers in clinical settings, many patients are often diagnosed only at advanced stages. To evaluate the diagnostic potential of 5’-tiRNA-iMet-CAT in early HCC, we collected data from 72 early-stage HCC patients (BCLC 0/A) and 112 healthy individuals. ROC curve analysis revealed that the AUC of 5’-tiRNA-iMet-CAT was 0.894 (95% CI: 0.849–0.939), outperforming those of AFP (0.685, 95% CI: 0.604–0.767) and PIVKA-II (0.746, 95% CI: 0.674–0.819) (Fig. 5c). With a cut-off point of 1.397757 (Youden index = 0.617), 5’-tiRNA-iMet-CAT reached an SEN of 90.3%, an ACCU of 78.8%, and an NPV of 92.0%, outperforming AFP and PIVKA-II across these metrics. (Table 6). When biomarkers were combined, the AUC for identifying early HCC among healthy donors was greater than that of any single marker alone. The combination of all three biomarkers achieved the highest AUC of 0.942 and increased the sensitivity to 97.2% (Fig. 5d).

Table 6.

Diagnostic value of serum 5’-tiRNA-iMet-CAT, serum AFP and PIVKA-II for HCC

Biomarker	SEN (%)	SPE (%)	ACCU (%)	PPV (%)	NPV (%)
5’-tiRNA-iMet-CAT	90.3(65/72)	71.4(80/112)	78.8(135/184)	67.0(65/97)	92.0(80/87)
AFP	40.3(29/72)	85.7(96/112)	67.9(98/184)	64.4(29/45)	69.1(96/139)
PIVKA-II	43.1(31/72)	89.3(100/112)	71.2.(131/184)	72.1(31/43)	70.9(100/141)
5’-tiRNA-iMet-CAT + AFP	95.8(69/72)	65.2(73/112)	77.2(142/184)	63.9(69/108)	96.1(73/76)
5’-tiRNA-iMet-CAT+ PIVKA-II	95.8(69/72)	66.1(74/112)	77.7(143/184)	64.5(69/107)	96.1(74/77)
5’-tiRNA-iMet-CAT + AFP+PIVKA-II	97.2(70/72)	60.7(68/112)	75.0(138/184)	61.4(70/114)	97.1(68/70)

SEN Sensitivity, SPE Specificity, ACCU Overall accuracy, PPV Positive predictive value, NPV Negative predictive value

To address the clinical challenge of differentiating HCC from benign liver diseases, we evaluated the ability of 5’-tiRNA-iMet-CAT to distinguish 139 HCC patients from 55 individuals with hepatitis. It achieved an AUC of 0.779 (95% CI: 0.707–0.850), outperforming AFP (AUC 0.661; 95% CI: 0.585–0.737) and demonstrating similar performance to PIVKA-II (AUC 0.797; 95% CI: 0.727–0.867) (Fig. 5e). When combined with other markers, 5’-tiRNA-iMet-CAT achieved a maximum AUC of 0.859 (Fig. 5f) and an increased sensitivity of 94.2% (Table 7). Given the clinical similarity of early HCC and hepatitis symptoms, the ability of serum 5’-tiRNA-iMet-CAT to distinguish early HCC patients from hepatitis patients is especially important. In terms of differential diagnosis, 5’-tiRNA-iMet-CAT could differentiate the 72 early-stage HCC patients from the 55 hepatitis patients with an AUC of 0.741 (95% CI: 0.654–0.828) (Supplementary Fig. S1).

Table 7.

Diagnostic value of serum 5’-tiRNA-iMet-CAT, serum AFP and PIVKA-II for HCC

Biomarker	SEN (%)	SPE (%)	ACCU (%)	PPV (%)	NPV (%)
5’-tiRNA-iMet-CAT	84.2(117/139)	56.4(31/55)	76.3(148/194)	83.0(117/141)	58.5(31/53)
AFP	47.5(66/139)	80.0(44/55)	56.7(110/194)	85.7(66/77)	37.6(44/117)
PIVKA-II	54.0(75/139)	85.5(47/55)	62.9(122/194)	90.4(75/83)	42.3(47/111)
5’-tiRNA-iMet-CAT + AFP	91.4(127/139)	52.7(29/55)	80.4(156/194)	83.0(127/153)	70.7(29/41)
5’-tiRNA-iMet-CAT+ PIVKA-II	92.1(128/139)	54.5(30/55)	81.4(158/194)	83.7(128/153)	73.2(30/41)
5’-tiRNA-iMet-CAT + AFP+PIVKA-II	94.2(131/139)	50.9(28/55)	82.0(159/194)	82.9(131/158)	77.8(28/36)

SEN Sensitivity, SPE Specificity, ACCU Overall accuracy, PPV Positive predictive value, NPV Negative predictive value

To address whether 5’-tiRNA-iMet-CAT could fulfill the unmet need in patients where standard biomarkers fail, we performed subgroup analyses in HCC patients with negative serum levels of AFP (< 20 ng/mL) and/or PIVKA-II (< 40 mAU/mL). Notably, in the clinically most challenging subgroup of patients who were double-negative for both AFP and PIVKA-II (n = 51), 5’-tiRNA-iMet-CAT still achieved an AUC of 0.893 (95% CI: 0.840–0.947) in distinguishing HCC from healthy controls (Fig. 5g). In the AFP-negative subgroup (n = 81), the AUC was 0.897 (95% CI: 0.854–0.940) (Fig. 5h), while in the PIVKA-II-negative subgroup (n = 71), it was 0.900 (95% CI: 0.855–0.944) (Fig. 5i).

These data highlight the potential of serum 5’-tiRNA-iMet-CAT as a diagnostic biomarker for early-stage HCC, demonstrating superior performance when combined with conventional markers. Notably, it can effectively differentiate HCC patients from hepatitis patients, and its diagnostic value is further enhanced when it is integrated with established tumour markers.

Prediction of downstream target genes of 5’-tiRNA-iMet-CAT

This study is the first to identify serum 5’-tiRNA-iMet-CAT as a novel biomarker for the diagnosis and postoperative assessment of HCC. To explore its potential mechanisms, we performed bioinformatics analysis to predict its downstream target genes, identifying a total of 1,702 potential candidates (Fig. 6b). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses indicated that these predicted target genes were significantly enriched in pathways associated with tumor invasion and metastasis, such as tight junctions and the calcium signalling pathway (Fig. 6a, c). Based on these findings, the predicted target genes were enriched in pathways related to cell junctions and calcium signalling. This provides a direction for future mechanistic investigations of 5’-tiRNA-iMet-CAT in HCC (e.g., verifying its effects on junction protein localization, calcium flux, and cell migration) and also offers in silico support for the biological rationality of this biomarker.

Fig. 6.

Prediction of downstream target genes of 5’-tiRNA-iMet-CAT. a Bubble plot of the KEGG analysis of 5’-tiRNA-iMet-CAT downstream target genes. b 5’-tiRNA-iMet-CAT target gene prediction from Miranda, PITA, TargetScan, and RNAhybrid databases. c Bubble plot of the GO analysis of 5’-tiRNA-iMet-CAT downstream target genes

Discussion

HCC accounts for 75–85% of primary liver cancer cases and remains a leading cause of cancer-related mortality globally, partly due to the lack of sensitive early diagnostic biomarkers [29, 30]. Current serum biomarkers like AFP lack sufficient sensitivity and specificity for early-stage detection [31], underscoring the urgent need for novel biomarkers.

In recent years, with the advancement of high-throughput sequencing technology, a novel class of non-coding RNAs derived from precursor or mature tRNAs—tsRNAs—has been widely discovered. They are highly enriched in blood and, owing to nucleotide modifications and their association with serum proteins, exhibit high stability in serum, demonstrating significant potential for liquid biopsy applications [32]. Numerous studies have confirmed that tsRNAs can serve as biomarkers for various diseases. For example, a signature composed of six tsRNA molecules can be used for early detection, treatment response monitoring, and prognosis prediction in diffuse large B-cell lymphoma (DLBCL) [33]. In gastric cancer, the expression level of serum i-tRF-AsnGTT is closely associated with cancer progression, indicating its diagnostic and prognostic monitoring value [34]. Moreover, differential expression profiles of ribosomal RNA‑derived (risRNA) and tsRNA in circulating extracellular vesicles can effectively distinguish early-stage gastric cancer patients from healthy individuals [35]. Beyond their role as biomarkers, tsRNAs also possess diverse biological functions. Current research has revealed several key mechanisms of action: some tsRNAs drive cancer progression by mediating hormonal responses, such as the 5’ - and 3’ -SHOT‑RNAs that promote cell proliferation in breast and prostate cancers [36]. According to Liu et al., i-tRF-LeuCAA may influence glioma progression by indirectly regulating TPM4 expression and modulating epithelial‒mesenchymal transition [37]. In terms of signaling pathway regulation, hypoxia-induced 5′‑tiRNA‑His GTG promotes colorectal cancer progression by activating the Hippo signaling pathway [38]. These findings establish the theoretical foundation for tsRNAs as critical regulatory molecules. However, the specific mechanisms and clinical translational value of tsRNAs in HCC remain to be further explored.

This study conducted an initial, comprehensive analysis of the potential of 5’-tiRNA-iMet-CAT as an HCC biomarker. We found that the elevated serum level of 5’-tiRNA-iMet-CAT exhibits good stability and specificity, suggesting its promise as a valuable biomarker for HCC. We screened the most significantly differentially expressed 5’ tiRNA molecule and named it 5’-tiRNA-iMet-CAT according to the Mintbase naming rules. Through molecular characterization, we identified its chromosomal localization and validated the PCR product via AGE and Sanger sequencing. Experiments confirmed that this molecule has good stability and specificity. Through large‑scale clinical validation of serum samples from 139 HCC patients, 55 hepatitis patients and 112 healthy individuals, we found significant differences in 5’-tiRNA-iMet-CAT expression among HCC patients, hepatitis patients and healthy donors with a gradient (HCC > hepatitis > healthy). We also found significant correlations between 5’-tiRNA-iMet-CAT expression and clinicopathological features (such as tumour differentiation, tumour size and TNM stage), suggesting its potential to predict HCC progression. More importantly, ROC analysis comparing serum 5’-tiRNA-iMet-CAT with the common HCC diagnostic markers AFP and PIVKA‑II revealed greater diagnostic efficiency for 5’-tiRNA-iMet-CAT. In combined diagnosis (5’-tiRNA-iMet-CAT + AFP + PIVKA‑II), 5’-tiRNA-iMet-CAT had the highest sensitivity (SEN = 0.849), and the three‑marker panel achieved the highest AUC (0.960), further improving diagnostic performance. Encouragingly, we evaluated the diagnostic performance of serum 5’-tiRNA-iMet-CAT in the early (BCLC 0/A) stage and found that its AUC and sensitivity exceeded those of common clinical HCC biomarkers. In addition, we quantitatively show that 5’-tiRNA-iMet-CAT achieves an AUC of 0.741 in discriminating early-stage HCC (BCLC 0/A) from hepatitis patients. This result provides a clear and previously unreported moderate performance benchmark for a tsRNA in this challenging diagnostic niche.

Importantly, preliminary evidence suggests that 5’-tiRNA-iMet-CAT demonstrated promising diagnostic efficacy not only in the AFP-negative subgroup (AUC = 0.897), but also in PIVKA-II-negative patients (AUC = 0.900) or even those double-negative for both markers(AUC = 0.893), highlighting its promise in complementing current diagnostic workflows to reduce diagnostic blind spots.

We further comprehensively assessed the potential of 5’-tiRNA-iMet-CAT as an HCC marker: analysis of postoperative dynamic changes suggested its potential for reflecting early postoperative changes, with expression changes correlated with disease progression. These results suggest promising prospects for serum 5’-tiRNA-iMet-CAT as a diagnostic and prognostic biomarker for HCC. Additionally, GO and KEGG enrichment analyses revealed that this tiRNA is significantly associated with biological processes such as cell junction, signal transduction regulation, ion/small molecule transmembrane transport, and cytoskeleton remodelling, suggesting its potential role in maintaining cellular homeostasis, regulating intercellular communication, and modulating motility.

Notwithstanding our findings, this study has several inherent constraints that merit careful consideration. Primarily, the single-center, retrospective design without an external validation cohort limits the generalizability of our conclusions and may introduce selection bias. The modest sample size of the discovery phase also necessitates caution in interpreting the screening results. Methodologically, diagnostic performance metrics, including ROC-derived cut-off values, were generated from the same cohort, a process potentially susceptible to optimism bias. Therefore, the reported sensitivity, specificity, and cut-off values should be regarded as preliminary findings, and subsequent independent external cohort studies are required to validate them and assess their reliability for clinical application. The exploratory analysis of postoperative dynamics, while promising, is based on a limited subset of patients with paired samples and requires confirmation in a dedicated prospective monitoring study. Finally, the functional implications of 5’-tiRNA-iMet-CAT are presently inferred from preliminary in silico analyses; these computationally generated hypotheses are valuable for guiding future work but await experimental substantiation. Collectively, prospective, multi-center studies with larger, well-annotated cohorts are imperative to rigorously validate the clinical utility of 5’-tiRNA-iMet-CAT and to elucidate its precise biological role in HCC pathogenesis.

Conclusion

In this study, we identified a novel tsRNA with high expression in HCC, which may serve as a valuable biomarker for early-stage HCC diagnosis and for assessing short-term efficacy in the early postoperative period. In combination with other tumor markers, it demonstrated higher sensitivity and specificity than individual biomarkers, highlighting its potential to improve the accuracy of HCC diagnosis and prognosis evaluation. Bioinformatic analyses preliminarily suggested its involvement in HCC progression, indicating that this tsRNA could be a promising target for future therapeutic research.

Abbreviations

HCC

Hepatocellular carcinoma

AFP

Alpha- fetoprotein

tRNA

Transfer RNA

tsRNAs

tRNA-derived small RNAs

qRT-PCR

Quantitative reverse transcription PCR

ROC

Receiver operating characteristic

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

MRI

Magnetic resonance imaging

CT

Computed tomography

ncRNAs

Non-coding RNAs

lncRNAs

Long non-coding RNAs

MiRNAs

MicroRNAs

piRNAs

PIWI-interacting RNAs

SiRNAs

Small interfering RNAs

tDRs

tRNA-derived RNAs

tRFs

tRNA fragments

tiRNAs

tRNA-derived stress-induced RNAs

ANG

Angiogenin

CRC

Colorectal cancer

5-FU

5-fluorouracil

TNM

Tumor-node-metastasis

BCLC

Barcelona Clinic Liver Cancer

DMEM

Dulbecco’s Modified Eagle Medium

FBS

Fetal bovine serum

U6

RUN6B

SD

Standard deviation

UCSC

University of California, Santa Cruz

AGE

Agarose gel electrophoresis

CV

Coefficients of variation

PIVKA-II

Protein Induced by Vitamin K Absence or Antagonist-II

HR

Hazard ratio

CI

Confidence Interval

SEN

Sensitivity

SPE

Specificity

ACCU

Accuracy

PPV

Positive predictive value

NPV

Negative predictive value

JNCC

Journal of the National Cancer Center

ICC

Intrahepatic cholangiocarcinoma

cHCC-CCA

Combined hepatocellular-cholangiocarcinoma

DLBCL

Diffuse large B‑cell lymphoma

RisRNA

Ribosomal RNA‑derived

AUC

Area Under the Curve

Authors’ contributions

Y.H. and S.Q. performed the methodology, investigation, data curation, and wrote the original draft. R.D. conducted validation, investigation, and reviewed and edited the manuscript. X.L. provided resources, supervision, and acquired funding. S.J. was responsible for conceptualization, resources, supervision, project administration, and reviewed and edited the manuscript. All authors commented on previous versions and approved the final manuscript.

Funding

This project was supported by grants from the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX25_3807, KYCX24_3582, KYCX25_3806), the National Natural Science Foundation of China (No. 82472356, No. 82272411), the Jiangsu Province Medical Key Discipline (Laboratory) (ZDXK202240), the Science and Technology Project of Jiangsu Province (BE2023741), and the Foundation of Jiangsu Province Research Hospital (YJXYY202204-XKB16).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study conformed to the Helsinki Declaration principles, and the local hospital ethics committee approved the study (Ethical Review Report No. 2018-L055).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.