Introduction
Hepatocellular carcinoma (HCC) is considered one of the most common and serious forms of liver malignancy globally, primarily occurring in a cirrhotic background. Its incidence has been observed to increase over the past two decades, particularly in Africa and Asia, and it is expected to continue rising until 2030, primarily due to Hepatitis viral infections such as hepatitis B virus (HBV) and hepatitis C virus (HCV), which are associated with end-stage liver disease and subsequently HCC. Additionally, steatotic liver disease has emerged as a growing related aetiology, particularly in individuals with underlying obesity, dyslipidemia, and diabetes.1,2 As is well-known, the diagnostic criteria for HCC mainly rely on contrast-enhanced imaging techniques (multiphase computerized tomography/magnetic resonance imaging) and, in selected cases, biopsy. However, non-invasive serum biomarkers may be beneficial for HCC, particularly in initial diagnosis, prognosis, and monitoring responses to different HCC therapies.1,3 Alpha-fetoprotein (AFP) has been widely used as a blood biomarker in clinical practice but has low sensitivity and specificity. Therefore, developing novel biomarkers with higher sensitivity and specificity is essential.1,3–5 One of the promising biomarkers is circular RNAs (circRNAs), which are long-stranded non-coding RNAs that form a sealed circular structure due to the absence of 5′/3′ ends. These RNAs originate from various genomic fragments and have shown high stability and versatility in several in-vivo studies.6,7 These studies have highlighted the role of circRNAs in gene transcription regulation, expression, and interaction with cell cycle proteins, primarily through microRNA (miRNA) sponging. Significant quantities of circRNAs have been detected in various body fluids (plasma, sera, and saliva).8–10 As a result, various researchers have proposed the potential role of circRNAs as biomarkers in multiple cancers,11,12 including glioma, ovarian, gastric, liver, esophageal, and colon cancers.13–18 Regarding HCC, recent studies have identified several forms of circRNAs (either upregulated or suppressed) in the HCC microenvironment that play vital roles in its carcinogenesis and progression, making them promising markers for HCC.19–23 However, genetic studies on the functional role of different circRNAs in HCC are still scarce and in their infancy, especially compared to other non-coding RNAs (miRNAs/long noncoding RNAs).24,25 Therefore, our current study aimed to analyze the expression levels of serum-derived hsa_circ_101555 in Egyptian HCC patients as a diagnostic/prognostic biomarker and to investigate their correlation with other known standard assessment parameters for liver disease and HCC.
Materials and methods
Study purpose
A controlled cross-sectional study was conducted at the Hepatoma Unit in the Tropical Medicine Department, Ain Shams University Hospitals, over the course of one year. The study aimed to measure the serum expression level of hsa_circ_101555 using real-time (RT) polymerase chain reaction (PCR) in 62 clinically/radiologically diagnosed Egyptian HCC patients at baseline (0 point) and three months after HCC treatment, compared to 30 healthy subjects. The primary objective was to evaluate the potential role of serum-derived hsa_circ_101555 as a diagnostic biomarker for HCC. Additionally, its prognostic significance and its ability to predict the response to therapy were assessed. We excluded non-HCC malignancy cases, individuals with immune-related disorders (such as autoimmune diseases or immunocompromised conditions like AIDS), those on immunosuppressive drugs (e.g., radiotherapy or chemotherapy), patients with sepsis, and those who were uncooperative or refused to participate.
Identification of unique circRNAs involved in HCC
The unique circRNAs involved in HCC were retrieved from circRNA databases using a bioinformatics pipeline that integrates various online tools and databases. The following steps were performed.
Sample preparation and data acquisition
RNA-seq data for HCC and normal liver tissues were obtained as follows:
Download RNA-seq data: RNA-seq data were obtained from publicly available repositories, including the Cancer Genome Atlas, Gene Expression Omnibus, and the European Nucleotide Archive. Both HCC samples and normal liver tissues were included for comparative analysis.
Quality control of RNA-seq data: The quality of RNA-seq data was performed using the FastQC tool to assess the quality of the reads.
circRNA prediction and identification
circRNAs were identified from the RNA-seq data using circRNA prediction tools by:
Mapping RNA-seq reads: The RNA-seq reads were aligned to the reference genome using splice-aware aligners such as BWA, STAR, or Bowtie2.
circRNA prediction: The circExplorer2 tool was used to identify potential circRNAs from the mapped RNA-seq data.
Filtration for high-confidence circRNAs
The identified circRNAs were filtered based on the following criteria: read support, back-splice junction reads, and minimum read counts. Based on the obtained results, hsa_circ_101555 was selected for measurement in the recruited subjects.
Differential expression analysis
Differentially expressed hsa_circ_101555 was identified between HCC and normal liver tissues by:
Quantification of hsa_circ_101555: The expression levels of hsa_circ_101555 were quantified using the predicted back-splice junction reads.
Normalization: The expression data were normalized using DESeq2’s normalization method.
Differential expression analysis: Differential expression analysis was performed using DESeq2 to identify significant differences in hsa_circ_101555 expression between HCC and normal samples.
Database comparison and validation
The identified hsa_circ_101555 was validated by comparing it with existing circRNA databases and published studies by:
Database comparison: The identified hsa_circ_101555 was validated with different databases, including circBase, circAtlas, and circBank.
Literature validation: The literature database “PubMed” was used to ensure that hsa_circ_101555 has been previously reported in HCC.
Protocol for designing primers for hsa_circ_101555
To design the potential primer sequences for hsa_circ_101555, the standard approach of selecting sequences around the back-splice junction was followed:
Sequence retrieval
The nucleotide sequence of hsa_circ_101555 was obtained using the circRNA Database “circBase (http://circbase.org/)” to find the sequence of hsa_circ_101555. The sequence was retrieved around the back-splice junction, including upstream and downstream flanking regions.
Back-splice junction identification: To identify the unique back-splice junction in the circRNA sequence, the following steps were applied:
Locate the junction: The back-splice junction is a unique feature of circRNAs, where the downstream exon is spliced to an upstream exon.
Extract junction sequence: A sequence spanning the back-splice junction was selected, typically 15–20 nucleotides upstream and downstream of the junction.
Primer design: to design the primers that specifically amplify the back-splice junction of hsa_circ_101555, the following steps were applied:
Primer design software
The NCBI Primer-BLAST primer design tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/ ) was used, with the following input parameters:
Target sequence: The sequence containing the back-splice junction.
Primer length: Typically 18–25 nucleotides.
GC content: 40–60% GC content.
Melting temperature: Between 58–62°C, with a difference of no more than 2°C between forward and reverse primers.
Product size: PCR product size of 100–200 bp.
Design forward primer: The forward primer should bind to the upstream exon at the 3′ end of hsa_circ_101555.
Design reverse primer: The reverse primer should bind to the sequence immediately downstream of the back-splice junction of hsa_circ_101555.
Primer-check for specificity and validation in silico
The Primer-BLAST tool was used to check for specificity against the human genome to ensure that the primers do not amplify linear mRNA or other sequences, and to ensure the primers specifically amplify the back-splice junction of hsa_circ_101555.
Secondary structure analysis
The secondary structures (e.g., hairpins, dimers) were checked using tools like OligoAnalyzer from IDT (https://www.idtdna.com/calc/analyzer ).
The designed primer sequences for hsa_circ_101555: Forward primer: 5′-AGTCTGACGTGACGTCGAGT-3′ and Reverse primer: 5′-TCGACGTACTGAGCTCAGTA-3′.
Measurement of circ_101555 gene in serum samples using RT-PCR
Sample collection
Three milliliters of venous blood was collected by venipuncture in a gel vacutainer tube, centrifuged at 4,000 rpm for 15 m at 37°C, and preserved at −80°C until analyzed.
Total RNA and miRNAs extraction and purification
Total mRNA and miRNAs were extracted using the miRNeasy Serum/Plasma Advanced Kit, cat no: 217204 (Qiagen, Hilden, Germany), according to the manufacturer’s protocol.
Reverse transcription
cDNA was synthesized by reverse transcription using the miRCURY LNA RT Kit, cat. no. 339340 (Qiagen, Hilden, Germany). The RT reaction mix (10 µL/tube) was prepared by mixing 2 µL of 5x miRCURY Syber Green RT reaction buffer, 4.5 µL of RNase-free water, 1 µL of 10x miRCURY RT Enzyme Mix, and 2 µL of RNA template. The reaction mix was incubated for 60 m at 42°C, followed by incubation for 5 m at 95°C to inactivate the reverse transcriptase enzyme, and finally cooled at 4°C. The cDNA was stored undiluted at 2–20°C until used for RT-PCR.
circ_101555 gene expression analysis
The circ_101555 level was amplified from mRNA using the designed primer sequences for hsa_circ_101555 [Forward primer: 5′-AGTCTGACGTGACGTCGAGT-3′, Reverse primer: 5′-TCGACGTACTGAGCTCAGTA-3′] and the RT2 SYBR Green qPCR Mastermixes Kit, cat no: 330520 (Qiagen, Germany). The Hs_ACTB_1_SG QuantiTect Primer Assay, cat no: 249900, ID: QT00095431, was used as a housekeeping gene. All samples were analyzed using the 5-plex Rotor Gene PCR Analyzer (Qiagen, Germany). The PCR reaction mix (25 µL) was prepared as follows: 10.5 µL of nuclease-free water, 12.5 µL of RT2 SYBR Green master mix, 1 µL of RT2 lncRNA qPCR assay (10 µM stock), and 1 µL of cDNA.
The PCR cycling protocol was adjusted as follows: Hot start activation step for 10 m at 95°C for HotStarTaq DNA Polymerase activation. Three-step cycling: denaturation for 15 s at 95°C, annealing for 30 s at 55°C, extension for 30 s at 72°C, for 40 cycles.
Calculation of gene expression
The Delta-Delta Ct method compares the expression difference (ΔCt) between the gene of interest and the reference gene (ACTB) in the experimental conditions, and separately in positive controls.
The method then compares the difference between the experimental and positive control samples.
Relative changes in gene expression between the two compared sequences are calculated using the formula: 2−ΔΔCt, where 2 is the efficiency set at 100%.
Statistical analysis
Data were entered, cleaned, and analyzed using IBM SPSS Statistics version 22. Descriptive statistics were presented as means and standard deviations. Categorical variables were summarized as frequencies and percentages. For comparative analyses, the independent sample t-test was used to compare the means of continuous variables between two independent groups, while the paired t-test was applied to assess within-group changes. One-way analysis of variance (ANOVA) was performed to compare means across more than two groups. The chi-square test was used to assess associations between categorical variables, and Fisher’s exact test was applied when expected cell counts were below 5 in 20.0% of the cells or more. Correlation analysis was conducted using Pearson’s correlation coefficient (r) to examine linear relationships between continuous variables. Additionally, receiver operating characteristic curve analysis was performed to evaluate the diagnostic performance of biomarkers, with the area under the curve (AUC) used as an indicator of discriminatory ability. A P-value of <0.05 was considered statistically significant for all tests.
Results
Characteristics of HCC patients
Descriptive statistics of clinical, laboratory, and tumor parameters in our hepatocellular carcinoma patients
A summary of the clinical, laboratory, and tumor characteristics in a cohort of HCC patients is shown in Table 1. The mean age is approximately 62.7 years (standard deviation (SD) = 8.5), and the body mass index (BMI) averages 26.1 (SD = 3.3), reflecting an older and potentially overweight population, with an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ratio > 1 (mean = 1.43, SD = 0.8), suggesting significant hepatocellular damage typical of HCC. There is high variability in AFP (mean = 2,548.5 ng/mL, SD = 7,756.9). The albumin-bilirubin (ALBI) score (mean = −2.31, SD = 0.51) and model for end-stage liver disease (MELD) score (mean = 9.3, SD = 2.6) suggest relatively preserved liver function in most patients. Elevated markers of inflammation, such as a high neutrophil-to-lymphocyte ratio (NLR) (mean = 8.16, SD = 4.4), indicate systemic inflammation commonly observed in cancer. The AST to platelet ratio (APRI) (mean = 0.85, SD = 0.61) and fibrosis-4 (FIB-4) scores (mean = 3.72, SD = 2.65) suggest a high degree of fibrosis, consistent with advanced liver disease. Notably, the circRNA value (mean = 7.66, SD = 3.74) displays variability, potentially reflecting differing patient molecular profiles. Overall, this profile reflects an HCC cohort with substantial liver dysfunction, inflammatory burden, and fibrosis, all critical factors in disease progression and management. Nearly all patients (96.8%) fall into the viral aetiology category; most patients (83.9%) have HCV as the primary aetiology (Table 2). Most patients (54.8%) have a single tumor, with 45.2% presenting with a tumor size greater than 5 cm. The presence of lymph node metastasis (9.7%) and vascular invasion (38.7%) reflect the aggressive nature of HCC, with portal vein thrombosis observed in a notable proportion. Regarding treatment choices, systemic therapy was the most used (32.3%), followed by locoregional therapies. Most patients are in advanced Barcelona Clinic Liver Cancer (BCLC) and Tomur, Node, Metastasis (TNM) staging systems, with 41.9% at BCLC stage C, 35.5% in TNM Stage III, and 42.0% in Stage IV, reflecting a high burden of disease. However, Child-Turcotte-Pugh (CTP) and MELD scores showed good liver functions. Overall, this cohort represents patients primarily with viral-related, advanced-stage HCC, with significant liver function preservation, suggesting a mix of treatment options depending on individual staging and liver function.
Table 1A summary of the clinical, laboratory, and tumor characteristics of hepatocellular carcinoma patients
| Variables | Mean | Standard deviation | Minimum | Maximum |
|---|
| Age | 62.68 | 8.47 | 44.00 | 78.00 |
| BMI | 26.12 | 3.33 | 21.00 | 35.20 |
| TLC_Pre | 5.99 | 2.39 | 1.60 | 12.20 |
| Hb_Pre | 12.23 | 1.30 | 10.30 | 14.40 |
| Platelets_Pre | 177.84 | 94.90 | 43.00 | 428.00 |
| Creatinine_Pre | 0.98 | 0.26 | 0.60 | 1.60 |
| Total bilirubin_Pre | 1.28 | 1.16 | 0.40 | 7.00 |
| Direct Bilirubin_Pre | 0.57 | 1.02 | 0.05 | 6.00 |
| AST_Pre | 49.74 | 37.53 | 14.00 | 180.00 |
| ALT_Pre | 39.06 | 30.17 | 13.00 | 190.00 |
| INR_Pre | 1.14 | 0.19 | 1.00 | 1.80 |
| Albumin_Pre | 3.66 | 0.54 | 2.60 | 4.60 |
| AFP_Pre | 2,548.45 | 7,756.93 | 10.00 | 40,267.00 |
| ALBI score_Pre | −2.31 | 0.51 | −3.16 | −1.20 |
| CRP_Pre | 10.93 | 7.76 | 1.70 | 42.00 |
| NLR_Pre | 8.16 | 4.41 | 3.00 | 18.00 |
| PLR_Pre | 303.15 | 209.54 | 58.57 | 983.33 |
| AST/ALT Ratio_Pre | 1.43 | 0.80 | 0.16 | 4.50 |
| FIB 4_Pre | 3.72 | 2.65 | 0.56 | 11.84 |
| APRI_Pre | 0.85 | 0.61 | 0.20 | 3.10 |
| NFS_Pre | 0.04 | 1.88 | −4.69 | 3.42 |
| circRNA value_Pre | 7.66 | 3.74 | 2.23 | 15.75 |
| Tumor no_Pre | 1.90 | 1.24 | 1.00 | 5.00 |
| MELD_Pre | 9.26 | 2.64 | 6.00 | 17.00 |
Table 2The clinical and tumor key characteristics of hepatocellular carcinoma patients
| Variables | No. | % |
|---|
| Etiology | HCV | 52 | 83.9% |
| HBV | 4 | 6.5% |
| HCV + HBV | 4 | 6.5% |
| SLD | 2 | 3.2% |
| Etiology categories | Viral | 60 | 96.8% |
| Non-viral | 2 | 3.2% |
| Arterial enhancement | No | 0 | 0.0% |
| Yes | 62 | 100.0% |
| Washout | No | 0 | 0.0% |
| Yes | 62 | 100.0% |
| Tumor no categories | Single | 34 | 54.8% |
| 2–3 | 20 | 32.3% |
| >3 | 8 | 12.9% |
| Tumor size | ≤5 | 34 | 54.8% |
| >5 | 28 | 45.2% |
| Lymph nodes invasion | No | 56 | 90.3% |
| Yes | 6 | 9.7% |
| Vascular invasion | No | 38 | 61.3% |
| Main PVT | 11 | 17.7% |
| Branch PVT | 13 | 21.0% |
| Intervention | Ablation (RF/MW) | 14 | 22.6% |
| Locoregional TACE | 14 | 22.6% |
| Combined (Ablation + locoregional) | 6 | 9.7% |
| Systemic ttt | 20 | 32.3% |
| Best supp. ttt | 8 | 12.9% |
| BCLC stage | A | 16 | 25.8% |
| B | 18 | 29.0% |
| C | 26 | 41.9% |
| D | 2 | 3.2% |
| Child-Turcotte-Pugh CTP score | A | 47 | 75.8% |
| Early B | 5 | 8.1% |
| Late B | 10 | 16.1% |
| C | 0 | 0.0% |
| MELD categories | Low risk (<10) | 48 | 77.4% |
| Moderate risk (10–19) | 14 | 22.6% |
| High risk (20–30) | 0 | 0.0% |
| Very high (>30) | 0 | 0.0% |
| TNM stage | Stage I | 2 | 3.2% |
| Stage II | 12 | 19.4% |
| Stage III | 22 | 35.5% |
| Stage IV A | 20 | 32.3% |
| Stage IV B | 6 | 9.7% |
Post-intervention tumor characteristics in our hepatocellular carcinoma cases
The key laboratory and clinical values post-intervention for patients with HCC provide a comprehensive view of liver function, inflammatory markers, and tumor burden (Table 3). AFP levels, with a mean of 3,637 and significant standard deviation (11,892.44), indicate substantial variability, likely reflecting differences in tumor biology and size among patients. The ALBI score (mean = −1.97) indicates moderately impaired liver function. Inflammatory markers, including NLR, show elevated averages (7.52). FIB-4 and APRI scores, indicative of liver fibrosis, are elevated in many patients. The average MELD score of 10.03 reflects relatively mild to moderate liver disease in this cohort. The circRNA value post-intervention, averaging 7.78, may provide insights into tumor-specific RNA markers. Tumor characteristics, with an average of 2.74 tumors per patient, reflect considerable disease burden. Overall, the complex clinical and biochemical landscape of post-intervention HCC patients underscores liver dysfunction, inflammation, and diverse tumor characteristics typical of advanced-stage liver cancer.
Table 3Post-intervention laboratory and clinical metrics in hepatocellular carcinoma Egyptian patients
| Variables | Mean | Standard deviation | Minimum | Maximum |
|---|
| TLC_Post | 5.02 | 1.49 | 3.00 | 8.00 |
| Hb_Post | 11.32 | 1.28 | 9.00 | 15.10 |
| Platelets_Post | 169.42 | 75.49 | 50.00 | 387.00 |
| Creatinine_Post | 1.00 | 0.21 | 0.70 | 1.40 |
| Total_Bilirubin_Post | 1.75 | 1.60 | 0.40 | 8.90 |
| Direct_Bilirubin_Post | 1.12 | 1.35 | 0.20 | 7.40 |
| AST_Post | 42.97 | 15.84 | 17.00 | 81.00 |
| ALT_Post | 39.29 | 23.73 | 18.00 | 155.00 |
| INR_Post | 1.15 | 0.19 | 1.00 | 1.70 |
| Albumin_Post | 3.38 | 0.49 | 2.50 | 4.40 |
| AFP_Post | 3,637.00 | 11,892.44 | 2.03 | 63,000.00 |
| ALBI_score_Post | −1.97 | 0.55 | −2.99 | −0.98 |
| CRP_Post | 14.55 | 18.85 | 3.00 | 100.00 |
| NLR_Post | 7.52 | 3.87 | 3.00 | 16.00 |
| PLR_Post | 316.73 | 174.73 | 100.00 | 725.00 |
| AST/ALT_Ratio_Post | 1.21 | 0.40 | 0.27 | 2.33 |
| FIB 4_Post | 3.46 | 2.58 | 0.51 | 11.04 |
| APRI_Post | 0.81 | 0.54 | 0.20 | 2.30 |
| NFS_Post | 0.16 | 1.62 | −4.60 | 3.26 |
| circRNA value post_1 | 7.78 | 4.02 | 1.34 | 15.97 |
| Tumor no_Post | 2.74 | 1.64 | 1.00 | 5.00 |
| MELD_Post | 10.03 | 4.20 | 6.00 | 22.00 |
Post-intervention tumor characteristics in HCC patients reveal a diverse spectrum of disease burden and response to treatment (Table 4). Most patients (61.3%) exhibit both arterial enhancement and washout patterns, common imaging features of HCC. Tumor multiplicity is notable, with 35.5% having more than three tumors. A slight majority (51.6%) have tumors larger than 5 cm. Vascular invasion was observed in 41.9% of patients, while lymph node metastasis (19.4%) indicated aggressive tumor biology and a poorer prognosis. Staging shows most patients were in intermediate to advanced stages according to the BCLC and TNM staging systems, with 45.2% in BCLC Stage C and 38.7% in Stage IV A, reflecting significant disease spread. On the other hand, the CTP score reveals that liver function is relatively preserved in this group, while MELD scores indicate mild to moderate liver disease severity. Regarding treatment response, as per response evaluation criteria in solid tomurs (RECIST)/ modified-RECIST (mRECIST) criteria, stable disease was seen in 40.7% of patients. In comparison, partial response (PR) and progressive disease were each observed in 29.6%, and no patients achieved complete response (CR). Overall, this cohort displays advanced tumor characteristics, with many patients showing large tumor size, vascular invasion, and later-stage disease, suggesting limitations in treatment efficacy and a generally poor prognosis. The distribution of RECIST/mRECIST response categories across different treatment modalities in hepatocellular carcinoma patients (Table 5) was statistically tested using the Fisher exact test. The results reveal a significant association between treatment modality and response categories (P = 0.000**). No CR was observed in any treatment group. Stable disease was most common in the ablation (RF/MW) group (57.1%) and the systemic treatment group (50%). The PR category was most prominent in the locoregional treatment (42.9%) and combined locoregional ablation + trans arterial chemoembolization (TACE) treatment (100%) groups. These findings suggest that the type of treatment modality significantly impacts the distribution of RECIST/mRECIST response categories, with no significant response (CR) observed across all treatment approaches.
Table 4Post-intervention tumor characteristics in HCC patients
| Variables | No. | % |
|---|
| Arterial_enhancement_Post | No | 24 | 38.7% |
| Yes | 38 | 61.3% |
| Washout_Post | No | 24 | 38.7% |
| Yes | 38 | 61.3% |
| Tumor no 1 categories_Post | Single | 22 | 35.5% |
| 2–3 | 18 | 29.0% |
| >3 | 22 | 35.5% |
| Tumor size_Post | ≤5 | 30 | 48.4% |
| >5 | 32 | 51.6% |
| Lymph nodes mets_Post | No | 50 | 80.6% |
| Yes | 12 | 19.4% |
| Vascular invasion_Post | No | 36 | 58.1% |
| Branch PVT | 4 | 6.5% |
| Main PVT | 22 | 35.5% |
| BCLC stage_Post | A | 16 | 25.8% |
| B | 10 | 16.1% |
| C | 28 | 45.2% |
| D | 8 | 12.9% |
| CTP score_Post | A | 34 | 54.8% |
| Early B | 20 | 32.3% |
| Late B | 1 | 1.6% |
| C | 7 | 11.3% |
| MELD categories_Post | Low risk (<10) | 42 | 67.7% |
| Moderate risk (10–19) | 18 | 29.0% |
| High risk (20–30) | 2 | 3.2% |
| Very high (>30) | 0 | 0.0% |
| TNM stage_Post | Stage I | 6 | 9.7% |
| Stage II | 14 | 22.6% |
| Stage III | 6 | 9.7% |
| Stage IV A | 24 | 38.7% |
| Stage IV B | 12 | 19.4% |
| RECIST/mRECIST_Post | CR | 0 | 0.0% |
| SD | 22 | 40.7% |
| PR | 16 | 29.6% |
| PD | 16 | 29.6% |
Table 5Distribution of RECIST/mRECIST response categories across different treatment modalities
| RECIST/mRECIST | Intervention
| Fisher exact test | P-value |
|---|
Ablation (RF/MW)
| TACE
| Combined locoregional ttt
| Systemic ttt
| Best supp. ttt
|
|---|
| No. | % | No. | % | No. | % | No. | % | No. | % |
|---|
| CR | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 25.626 | 0.000** |
| SD | 8 | 57.1% | 4 | 28.6% | 0 | 0.0% | 10 | 50.0% | 0 | 0.0% | | |
| PR | 4 | 28.6% | 6 | 42.9% | 6 | 100.0% | 0 | 0.0% | 0 | 0.0% | | |
| PD | 2 | 14.3% | 4 | 28.6% | 0 | 0.0% | 10 | 50.0% | 0 | 0.0% | | |
Evaluation of serum circRNA hsa_circ_101555 in our HCC cases and healthy subjects
The role of hsa_circ_101555 in HCC patients and healthy subjects
Diagnostic value of hsa_circ_101555 to differentiate between cases and healthy controls
A comparison of circRNA values among healthy controls and HCC cases (Fig. 1) showed a statistically significant difference (P-value = 0.000). The mean circRNA value was highest in HCC cases (7.66 ± 3.74) and lowest in healthy controls (1.21 ± 0.96). This significant elevation of circRNA values in HCC cases compared to healthy controls suggests a potential role for circRNA as a biomarker for distinguishing HCC from healthy states. Moreover, the circRNA value showed excellent diagnostic accuracy in differentiating HCC patients from healthy controls at the cutoff point of 1.966. The test achieves perfect sensitivity (1.000) and 90% specificity (1 - specificity = 0.100) (Tables 6 and 7), as indicated by an AUC of 0.984, reflecting a highly effective test with strong discriminatory power. The standard error of 0.010 highlights the precision of the estimate, while the statistically significant P-value (0.000) confirms that the result is unlikely due to chance. Furthermore, the 95% Confidence Interval (0.964–1.000) underscores the robustness of the diagnostic performance, with the lower bound still well above the threshold for high accuracy. Overall, these findings suggest that the circRNA value is a promising biomarker for HCC detection at the cutoff point of 1.966, warranting further validation in diverse populations and clinical settings to establish its clinical utility.
Table 6Diagnostic value of circRNA in differentiating between cases and healthy controls
Area under the curve
|
|---|
Test result variable(s): circRNA value
|
|---|
| Area | Std. Errora | Asymptotic Sig.b | Asymptotic 95% confidence interval
|
|---|
| Lower bound | Upper bound |
|---|
| 0.984 | 0.010 | 0.000** | 0.964 | 1.000 |
Table 7Optimal cutoff points and corresponding diagnostic performance of circRNA values for differentiating hepatocellular carcinoma patients from healthy controls
Coordinates of the curve
|
|---|
Test result variable(s): circRNA value
|
|---|
| Positive if greater than or equal toa | Sensitivity | 1 – Specificity |
|---|
| −0.5310 | 1.000 | 1.000 |
| 0.5095 | 1.000 | 0.900 |
| 0.6375 | 1.000 | 0.800 |
| 0.7485 | 1.000 | 0.700 |
| 0.8420 | 1.000 | 0.600 |
| 0.9515 | 1.000 | 0.500 |
| 1.0340 | 1.000 | 0.400 |
| 1.0805 | 1.000 | 0.300 |
| 1.3930 | 1.000 | 0.200 |
| 1.9660 | 1.000 | 0.100 |
| 2.4045 | 0.968 | 0.100 |
| 2.8665 | 0.935 | 0.100 |
| 3.3025 | 0.903 | 0.100 |
| 3.6005 | 0.871 | 0.100 |
| 3.8030 | 0.839 | 0.100 |
| 3.9375 | 0.839 | 0.000 |
| 4.0330 | 0.806 | 0.000 |
| 4.0755 | 0.774 | 0.000 |
| 4.6855 | 0.742 | 0.000 |
| 5.3980 | 0.710 | 0.000 |
| 5.7275 | 0.677 | 0.000 |
| 6.0730 | 0.645 | 0.000 |
| 6.2420 | 0.613 | 0.000 |
| 6.3520 | 0.581 | 0.000 |
| 6.4625 | 0.548 | 0.000 |
| 6.6460 | 0.516 | 0.000 |
| 6.9760 | 0.484 | 0.000 |
| 7.2455 | 0.452 | 0.000 |
| 7.8055 | 0.419 | 0.000 |
| 8.4705 | 0.387 | 0.000 |
| 8.7365 | 0.355 | 0.000 |
| 8.8585 | 0.323 | 0.000 |
| 8.9820 | 0.290 | 0.000 |
| 9.4365 | 0.258 | 0.000 |
| 10.2920 | 0.226 | 0.000 |
| 11.1410 | 0.194 | 0.000 |
| 11.9410 | 0.161 | 0.000 |
| 12.6170 | 0.129 | 0.000 |
| 13.4865 | 0.097 | 0.000 |
| 14.9205 | 0.065 | 0.000 |
| 16.7470 | 0.000 | 0.000 |
Prognostic value of post-intervention hsa_circ_101555 to differentiate between constant response, disease progression, and disease regression in our HCC cases based on RECIST/mRECIST response categories
The prognostic value of post-intervention circRNA value in differentiating between HCC progression and regression in our patients, based on RECIST/mRECIST response categories, showed that the best-chosen cutoff value was 5.1150, which demonstrated a sensitivity of 1.000 and a 1 - specificity of 0.250. The AUC for the circRNA value post-intervention was 0.891, with a standard error of 0.058 (Tables 8 and 9). The 95% confidence interval for the AUC ranges from 0.778 to 1.000. These results suggest that the post-intervention circRNA value is a strong predictor, with high sensitivity (100.0%) and a moderate level of specificity (75.0%). The cutoff point and AUC significantly greater than the null hypothesis value of 0.5 indicate the excellent discriminatory ability of the post-intervention circRNA value in distinguishing between disease progression and regression in these patients.
Table 8The prognostic value of post-intervention circRNA in differentiating between HCC progression and regression in cases based on RECIST/mRECIST response categories
Area under the curve
|
|---|
Test result variable(s): circRNA value (post-intervention)
|
|---|
| Area | Std. Errora | Asymptotic Sig.b | Asymptotic 95% confidence interval
|
|---|
| Lower bound | Upper bound |
|---|
| 0.891 | 0.058 | 0.000** | 0.778 | 1.000 |
Table 9The best cutoff point for circRNA post-intervention to differentiate between disease progression and disease regression in hepatocellular carcinoma patients based on RECIST/mRECIST response categories
Coordinates of the curve
|
|---|
Test result variable(s): circRNA value post_1
|
|---|
| Positive if greater than or equal toa | Sensitivity | 1 - Specificity |
|---|
| 0.3440 | 1.000 | 1.000 |
| 2.0930 | 1.000 | 0.875 |
| 3.2830 | 1.000 | 0.750 |
| 4.0010 | 1.000 | 0.625 |
| 4.3380 | 1.000 | 0.500 |
| 4.4915 | 1.000 | 0.375 |
| 5.1150 | 1.000 | 0.250 |
| 6.1560 | 0.875 | 0.250 |
| 7.3250 | 0.750 | 0.250 |
| 8.2695 | 0.625 | 0.250 |
| 9.5815 | 0.625 | 0.125 |
| 10.7185 | 0.500 | 0.125 |
| 12.1750 | 0.500 | 0.000 |
| 13.7585 | 0.375 | 0.000 |
| 14.7635 | 0.250 | 0.000 |
| 15.7485 | 0.125 | 0.000 |
| 16.9670 | 0.000 | 0.000 |
The prognostic value of post-intervention circRNA value in distinguishing between Constant Response and Disease Regression in patients with Hepatocellular Carcinoma, based on the RECIST/mRECIST response categories, showed that a selected cutoff point of 7.8025 yields a sensitivity of 0.750 and a 1 - specificity of 0.364. The AUC is 0.795, indicating a good level of accuracy in using the circRNA value as a predictor of patient response (Tables 10 and 11). The standard error is 0.071, and the asymptotic significance is 0.002, which is statistically significant at the 0.01 level. The 95% confidence interval for the AUC ranges from 0.655 to 0.936, further supporting the reliability of the circRNA value as a diagnostic marker. This cutoff represents a balance between detecting true positive cases and managing the risk of misclassification, suggesting it may be a practical threshold for clinical decision-making. However, further evaluation and comparison with other clinical parameters would be necessary to determine if this point offers the best overall predictive performance.
Table 10The prognostic value of post-intervention circRNA in distinguishing between constant response and disease regression based on the RECIST/mRECIST response categories
Area under the curve
|
|---|
Test result variable(s): circRNA value (post-intervention)
|
|---|
| Area | Std. Errora | Asymptotic Sig.b | Asymptotic 95% confidence interval
|
|---|
| Lower bound | Upper bound |
|---|
| 0.795 | 0.071 | 0.002** | 0.655 | 0.936 |
Table 11The best cutoff point for circRNA post-intervention to differentiate between constant response and disease regression in hepatocellular carcinoma patients based on RECIST/mRECIST response categories
Coordinates of the curve
|
|---|
Test result variable(s): circRNA value (post intervention)
|
|---|
| Positive if greater than or equal toa | Sensitivity | 1 - Specificity |
|---|
| 1.2610 | 1.000 | 1.000 |
| 3.1825 | 1.000 | 0.909 |
| 4.4585 | 1.000 | 0.818 |
| 4.9680 | 1.000 | 0.727 |
| 5.1590 | 1.000 | 0.636 |
| 5.4200 | 1.000 | 0.545 |
| 5.7045 | 0.875 | 0.545 |
| 6.2155 | 0.875 | 0.455 |
| 7.1625 | 0.750 | 0.455 |
| 7.8205 | 0.750 | 0.364 |
| 8.0110 | 0.625 | 0.364 |
| 9.2500 | 0.625 | 0.273 |
| 10.5340 | 0.625 | 0.182 |
| 11.3540 | 0.500 | 0.182 |
| 12.8105 | 0.500 | 0.091 |
| 13.6620 | 0.375 | 0.091 |
| 13.9005 | 0.375 | 0.000 |
| 14.7635 | 0.250 | 0.000 |
| 15.7485 | 0.125 | 0.000 |
| 16.9670 | 0.000 | 0.000 |
Relationship between pre-/post-intervention circRNA hsa_circ_101555 and our HCC patients’ demographic, clinical and laboratory data
Relationship between pre-/post-intervention hsa_circ_101555 and patients’ demographic data
The relationship between pre-intervention circRNA levels and the demographic and clinical characteristics of HCC patients (Table 12) showed that gender was the only statistically significant variable, with females exhibiting higher circRNA levels (Mean = 10.54, SD = 4.90). Although not significant (P = 0.123), BMI categories showed variation, with obese patients having the highest levels (Mean = 9.47, SD = 4.94). Additionally, patients with comorbidities had higher circRNA levels (Mean = 8.72, SD = 3.93), and regarding aetiology, HBV patients showed the highest circRNA levels (Mean = 9.07, SD = 5.80). These findings suggest that gender significantly influences pre-intervention circRNA levels, while trends in other variables warrant further investigation.
Table 12The relationship between pre-intervention circRNA levels and the demographic and clinical characteristics of hepatocellular carcinoma patients
| Variables | circRNA value Pre
| Independent sample t-test | P-value |
|---|
| Mean | ±SD |
|---|
| Gender | Male | 6.97 | 3.08 | −2.413 | 0.031* |
| Female | 10.54 | 4.90 | | |
| BMI categories | Normal | 8.20 | 3.43 | 2.176# | 0.123 |
| Overweight | 6.74 | 3.48 | | |
| Obese | 9.47 | 4.94 | | |
| Comorbidities | No | 6.99 | 3.50 | −1.806 | 0.076 |
| Yes | 8.72 | 3.93 | | |
| Etiology | HCV | 7.57 | 3.80 | 0.204# | 0.893 |
| HBV | 9.07 | 5.80 | | |
| HCV + HBV | 7.62 | 1.36 | | |
| SLD | 7.15 | 0.00 | | |
| Etiology categories | Viral | 7.68 | 3.80 | 1.081 | 0.284 |
| Non-viral | 7.15 | 0.00 | | |
The relationship between post-intervention circRNA levels and the demographic and clinical characteristics of HCC patients (Table 13) showed no statistically significant differences for gender (P = 0.133), although females had slightly higher post-intervention circRNA levels (Mean = 9.35, SD = 3.09). BMI categories also showed no significant differences (P = 0.123), but obese patients exhibited the highest levels (Mean = 9.97, SD = 4.15). Patients with comorbidities had higher circRNA levels (Mean = 8.92, SD = 4.53), though the difference was not significant (P = 0.077).
Table 13The relationship between post-intervention circRNA levels and the demographic and clinical characteristics of hepatocellular carcinoma patients
| Variables | circRNA value Post
| Independent sample t-test | P-value |
|---|
| Mean | ±SD |
|---|
| Gender | Male | 7.40 | 4.15 | −1.523 | 0.133 |
| Female | 9.35 | 3.09 | | |
| BMI categories | Normal | 8.93 | 4.28 | 2.176# | 0.123 |
| Overweight | 6.27 | 3.24 | | |
| Obese | 9.97 | 4.15 | | |
| Comorbidities | No | 7.06 | 3.54 | −1.802 | 0.077 |
| Yes | 8.92 | 4.53 | | |
| Etiology | HCV | 7.88 | 4.09 | 0.204# | 0.893 |
| HBV | 6.03 | 1.88 | | |
| HCV + HBV | 9.29 | 5.43 | | |
| SLD | 5.65 | 0.00 | | |
| Etiology categories | Viral | 7.85 | 4.07 | 4.202 | 0.000** |
| Non-viral | 5.65 | 0.00 | | |
In terms of aetiology, differences among HCV, HBV, HCV + HBV, and steatotic liver disease groups were not statistically significant (P = 0.893). However, categorizing aetiology into viral and non-viral revealed a significant difference (P = 0.000**), with viral aetiology cases showing substantially higher circRNA levels (Mean = 7.85, SD = 4.07) than non-viral cases (Mean = 5.65, SD = 0.00). These results suggest that while gender, BMI, and comorbidities show trends, they do not significantly influence post-intervention circRNA levels. However, aetiology categorization into viral or non-viral is a significant factor affecting post-intervention levels.
Correlation between pre-/post-intervention hsa_circ_101555 and patients’ Laboratory measurements and non-invasive markers
Correlation between pre-intervention circRNA levels and various laboratory measurements in HCC patients (Table 14) demonstrated that significant negative correlations were observed with aspartate aminotransferase (AST; r = −0.359, P = 0.004**). Similarly, alanine aminotransferase (ALT; r = −0.295, P = 0.020*), neutrophil-to-lymphocyte ratio (NLR; r = −0.331, P = 0.009**), and platelet-to-lymphocyte ratio (PLR; r = −0.290, P = 0.022*) also showed significant negative correlations. Other laboratory parameters, including international normalized ratio (INR), albumin, AFP, ALBI score, and C-reactive protein, did not demonstrate significant correlations with circRNA levels (P > 0.05). Overall, the data suggest a meaningful relationship between pre-intervention circRNA levels and markers of inflammation and liver function, with notable trends warranting further investigation.
Table 14The correlation between pre-intervention circRNA levels and various laboratory measurements in hepatocellular carcinoma patients
| Variables | circRNA value Pre
|
|---|
| r | P-value |
|---|
| TLC_Pre | −0.380 | 0.002** |
| Hb_Pre | −0.046 | 0.723 |
| Platelets_Pre | −0.206 | 0.108 |
| Creatinine_Pre | −0.011 | 0.930 |
| Total bilirubin_Pre | −0.128 | 0.323 |
| Direct Bilirubin_Pre | −0.190 | 0.138 |
| AST_Pre | −0.359 | 0.004** |
| ALT_Pre | −0.295 | 0.020* |
| INR_Pre | −0.109 | 0.398 |
| Albumin_Pre | 0.032 | 0.803 |
| AFP_Pre | −0.014 | 0.913 |
| ALBI score_Pre | −0.049 | 0.705 |
| CRP_Pre | −0.199 | 0.121 |
| NLR_Pre | −0.331 | 0.009** |
| PLR_Pre | −0.290 | 0.022* |
Correlation between pre-intervention circRNA levels and various non-invasive markers in HCC patients (Table 15) showed that no statistically significant correlations were observed (P > 0.05). The AST/ALT ratio showed a weak negative correlation (r = −0.087, P = 0.499). Similarly, the FIB-4 index had a very weak positive correlation (r = 0.041, P = 0.750), while the APRI score displayed a weak negative correlation (r = −0.169, P = 0.189). In summary, pre-intervention circRNA levels do not show significant associations with the non-invasive markers analyzed. These results suggest that circRNA levels might not directly correlate with these markers of liver fibrosis or inflammation in the pre-intervention phase of HCC patients. Further research is needed to clarify these relationships.
Table 15The correlation between pre-intervention circRNA levels and various non-invasive markers in hepatocellular carcinoma patients
| Variables | circRNA value Pre
|
|---|
| r | P-value |
|---|
| AST/ALT Ratio_Pre | −0.087 | 0.499 |
| FIB 4_Pre | 0.041 | 0.750 |
| APRI_Pre | −0.169 | 0.189 |
| NFS_Pre | 0.120 | 0.353 |
Correlation between post-intervention circ-RNA levels and laboratory measurements in HCC patients (Table 16) demonstrated that significant correlations, both positive and negative, were observed for various parameters. Negative correlations were observed with total leukocyte count (r = −0.442, P = 0.000**), platelet count (r = −0.372, P = 0.003**), hemoglobin (r = −0.260, P = 0.042*), and albumin (r = −0.439, P = 0.000**). These findings suggest that higher circRNA levels post-intervention are associated with lower values of these markers, indicating a potential relationship with systemic inflammation or hematological changes. Positive correlations were identified with INR (r = 0.423, P = 0.001**), the ALBI score (r = 0.424, P = 0.001**), NLR (r = 0.410, P = 0.001**), and AFP (r = 0.273, P = 0.032*). These results indicate that higher circRNA levels are associated with increased values of these parameters, possibly reflecting disease severity or liver dysfunction. In summary, significant relationships between post-intervention circRNA levels and markers of systemic inflammation, liver function, and disease severity indicate that negative correlations with total leukocyte count, platelets, Hb, and albumin suggest a decline in these markers with increasing circRNA levels, while positive correlations with INR, ALBI score, NLR, and AFP indicate an association with worsening liver function and disease progression.
Table 16The correlation between post-intervention circRNA levels and various laboratory measurements in hepatocellular carcinoma patients
| Variables | circRNA value post intervention
|
|---|
| r | P-value |
|---|
| TLC_Post | −0.442 | 0.000** |
| Hb_Post | −0.260 | 0.042* |
| Platelets_Post | −0.372 | 0.003** |
| Creatinine_Post | 0.172 | 0.181 |
| Total_Bilirubin_Post | 0.174 | 0.175 |
| Direct_Bilirubin_Post | 0.166 | 0.197 |
| AST_Post | 0.069 | 0.597 |
| ALT_Post | −0.088 | 0.496 |
| INR_Post | 0.423 | 0.001** |
| Albumin_Post | −0.439 | 0.000** |
| AFP_Post | 0.273 | 0.032* |
| ALBI_score_Post | 0.424 | 0.001** |
| CRP_Post | −0.066 | 0.611 |
| NLR_Post | 0.410 | 0.001** |
| PLR_Post | 0.116 | 0.371 |
Correlation between post-intervention circRNA levels and non-invasive markers in HCC patients (Table 17) showed that significant positive correlations were observed for all non-invasive markers analyzed. The AST/ALT ratio showed a statistically significant positive correlation with post-intervention circRNA levels (r = 0.284, P = 0.025*). Stronger positive correlations were observed for FIB-4 (r = 0.501, P = 0.000**), APRI score (r = 0.436, P = 0.000**), and NAFLD fibrosis score (r = 0.530, P = 0.000**). In summary, significant relationships between post-intervention circRNA levels and non-invasive markers of liver fibrosis and function, with particularly strong correlations for FIB-4, APRI, and NFS, suggest that circRNA levels may serve as an indicator of liver disease severity in HCC patients after intervention.
Table 17The correlation between post-intervention circRNA levels and various non-invasive markers in hepatocellular carcinoma patients
| Variables | circRNA value Post
|
|---|
| r | P-value |
|---|
| AST/ALT Ratio_Post | 0.284 | 0.025* |
| FIB 4_Post | 0.501 | 0.000** |
| APRI_Post | 0.436 | 0.000** |
| NFS_Post | 0.530 | 0.000** |
Relationship between pre-/post-intervention hsa_circ_101555 and patients’ clinical and pathological features
A significant association was found between pre-intervention circRNA levels and certain clinical/pathological features in HCC patients (Table 18). Tumor size showed a highly significant relationship, with higher circRNA values observed in patients with tumors ≤5 cm. Similarly, significant variations were noted across BCLC stages (P = 0.000), with the highest values in Stage A. TNM staging also demonstrated a statistically significant association (P = 0.015), with Stage I patients exhibiting the highest circRNA values. While patients without lymph node metastasis and those without vascular invasion had higher circRNA values compared to their counterparts, these differences were not statistically significant (P = 0.250 and P = 0.666, respectively). Additionally, arterial enhancement, washout, CTP score, and MELD categories showed trends in circRNA values, but these were not statistically significant. Overall, circRNA values appear to correlate significantly with key features such as tumor size, BCLC stage, and TNM stage, suggesting their potential utility in assessing disease progression and stratification among HCC patients.
Table 18The relationship between pre-intervention circRNA values and clinical and pathological features of hepatocellular carcinoma patients
| Variables | circRNA value Pre
| Independent sample t-test | P-value |
|---|
| Mean | ±SD |
|---|
| Arterial enhancement (Pre) | No | . | . | | |
| Yes | 7.66 | 3.74 | | |
| Washout (Pre) | No | . | . | | |
| Yes | 7.66 | 3.74 | | |
| Tumor no categories (Pre) | Single | 7.44 | 4.04 | 0.492# | 0.614 |
| 2–3 | 7.53 | 3.64 | | |
| >3 | 8.89 | 2.57 | | |
| Tumor size (Pre) | ≤5 | 9.29 | 3.80 | 18.415 | 0.000** |
| >5 | 5.68 | 2.55 | | |
| LN mets (Pre) | No | 7.84 | 3.84 | 1.347 | 0.250 |
| Yes | 5.98 | 2.02 | | |
| Vasc invasion (Pre) | No | 7.94 | 4.40 | 0.410# | 0.666 |
| Main PVT | 6.78 | 1.65 | | |
| Branch PVT | 7.57 | 2.85 | | |
| BCLC stage (Pre) | A | 10.63 | 4.59 | 8.273# | 0.000** |
| B | 5.57 | 3.13 | | |
| C | 7.60 | 2.13 | | |
| D | 3.45 | 0.00 | | |
| CTP score (Pre) | A | 7.34 | 3.61 | 1.786# | 0.177 |
| Early B | 10.62 | 3.30 | | |
| Late B | 7.68 | 4.21 | | |
| C | . | . | | |
| MELD categories (Pre) | Low risk (<10) | 8.11 | 3.89 | 3.299 | 0.074 |
| Moderate risk (10–19) | 6.09 | 2.70 | | |
| High risk (20–30) | . | . | | |
| Very high (>30) | . | . | | |
| TNM stage (Pre) | Stage I | 15.75 | 0.00 | 3.359# | 0.015* |
| Stage II | 8.50 | 5.15 | | |
| Stage III | 7.26 | 3.41 | | |
| Stage IV A | 7.29 | 2.64 | | |
| Stage IV B | 5.98 | 2.02 | | |
The relationship between post-intervention circRNA values and clinical and pathological features of HCC patients (Table 19) revealed significant associations with several variables. Arterial enhancement and washout both showed higher circRNA values in the presence of these features compared to their absence (P = 0.030). Tumor size was also significantly associated, with larger tumors (>5 cm) having higher circRNA values (P = 0.019). Tumor number categories exhibited a significant relationship (P = 0.031), with the highest values seen in patients with more than three tumors. Vascular invasion was another significant factor (P = 0.030), with main portal vein thrombosis and branch portal vein thrombosis showing elevated circRNA values. BCLC stage demonstrated highly significant differences (P = 0.007), with Stage C patients exhibiting the highest circRNA values. Similarly, the CTP score revealed highly significant associations (P = 0.000), with increasing scores correlating with higher circRNA values. MELD categories also showed a significant trend (P = 0.045), with moderate-risk patients having the highest circRNA values. TNM stage was significantly associated (P = 0.012), with Stage IV A patients having higher circRNA values. Interestingly, lymph node metastasis did not demonstrate a significant association with post-intervention circRNA values (P = 0.831). These findings suggest that post-intervention circRNA values are significantly influenced by several clinical and pathological features, particularly those related to tumor burden, vascular invasion, and liver function status.
Table 19The relationship between post-intervention circRNA values and clinical and pathological features of hepatocellular carcinoma patients
| Variables | circRNA value Post
| Independent sample t-test | P-value |
|---|
| Mean | ±SD |
|---|
| Arterial enhancement (Post) | No | 6.40 | 3.72 | 4.949 | 0.030* |
| Yes | 8.65 | 4.00 | | |
| Washout (Post) | No | 6.40 | 3.72 | 4.949 | 0.030* |
| Yes | 8.65 | 4.00 | | |
| Tumor no categories (Post) | Single | 7.36 | 4.01 | 3.678# | 0.031* |
| 2–3 | 6.23 | 3.03 | | |
| >3 | 9.46 | 4.26 | | |
| Tumor size (Post) | ≤5 | 6.56 | 3.31 | 5.777 | 0.019* |
| >5 | 8.92 | 4.33 | | |
| Lymph nodes invasion (Post) | No | 7.83 | 4.21 | 0.064 | 0.831 |
| Yes | 7.55 | 3.25 | | |
| Vascular invasion (Post) | No | 6.66 | 3.94 | 3.732# | 0.030* |
| Main PVT | 10.20 | 3.02 | | |
| Branch PVT | 9.17 | 3.80 | | |
| BCLC stage (Post) | A | 7.87 | 3.65 | 4.505# | 0.007** |
| B | 3.95 | 2.04 | | |
| C | 8.91 | 4.22 | | |
| D | 8.40 | 3.40 | | |
| CTP score (Post) | A | 5.86 | 2.43 | 9.147# | 0.000** |
| Early B | 9.67 | 4.55 | | |
| Late B | 15.97 | 0.00 | | |
| C | 10.52 | 3.66 | | |
| MELD categories (Post) | Low risk (<10) | 6.95 | 3.62 | 3.269# | 0.045* |
| Moderate risk (10–19) | 9.74 | 4.52 | | |
| High risk (20–30) | 7.66 | 0.00 | | |
| Very high (>30) | . | . | | |
| TNM stage (Post) | Stage I | 9.62 | 4.17 | 3.511# | 0.012* |
| Stage II | 4.79 | 2.52 | | |
| Stage III | 7.65 | 3.46 | | |
| Stage IV A | 9.21 | 4.39 | | |
| Stage IV B | 7.55 | 3.25 | | |
Comparison of post-intervention hsa_circ_101555 levels among different treatment modalities and RECIST/mRECIST response categories
Post-intervention circRNA values across different treatment modalities and RECIST/mRECIST response categories in HCC patients are shown in Table 20. The mean circRNA values varied among the intervention groups, with the highest observed in the systemic treatment group (9.21 ± 4.03) and the lowest in the combined locoregional ablation and TACE group (5.04 ± 2.75). However, the differences among these treatment modalities were not statistically significant (P = 0.181). In contrast, when categorized by RECIST/mRECIST response, there was a statistically significant variation in circRNA values (P = 0.000**). The highest mean value was found in the progressive disease category (11.24 ± 3.95), while the lowest was in the PR category (5.07 ± 3.00). These findings suggest that circRNA expression levels may differ significantly based on tumor response to treatment, as indicated by RECIST/mRECIST criteria, but not necessarily by the type of intervention.
Table 20Comparison of post-intervention circRNA values among different treatment modalities and RECIST/mRECIST response categories
| Variables | circRNA value (Post)
| One way ANOVA | P-value |
|---|
| Mean | ±SD |
|---|
| Intervention | Ablation (RF/MW) | 6.77 | 2.77 | 1.621 | 0.181 |
| Locoregional (TACE) | 7.88 | 5.52 | | |
| Combined (Ablation + TACE) | 5.04 | 2.75 | | |
| Systemic ttt | 9.21 | 4.03 | | |
| Best supp. ttt | 7.84 | 2.44 | | |
| RECIST/mRECIST | CR | . | . | 12.864 | 0.000** |
| SD | 7.21 | 3.51 | | |
| PR | 5.07 | 3.00 | | |
| PD | 11.24 | 3.95 | | |
Discussion
HCC represents a common and serious form of primary liver malignancy in Egypt, and its incidence has been observed to increase globally over the past two decades. Therefore, early diagnosis of HCC is considered the most effective way to prevent cancer progression and related mortality and morbidity. However, detecting the early stages of HCC remains a challenge for many hepatologists.1–3 Recently, circRNAs have emerged as promising diagnostic markers and even therapeutic targets for HCC.19,22,23 In our study, we analyzed the expression level of serum-derived hsa_circ_101555 in 62 Egyptian HCC patients compared to 30 healthy subjects using Real-time PCR to evaluate its potential role as a diagnostic biomarker for HCC. Moreover, it was measured three months later in these patients after receiving HCC treatment to observe its prognostic significance, compare it with other known standard parameters, and predict the response to therapy.
As known, most of the human transcriptome consists of non-coding RNA, and only about 2% of human genomic sequences are protein-coding genes. circRNAs represent a novel class of non-coding RNAs that have been detected in human peripheral blood, tissues, and exosomes through high-throughput sequencing technology.6,7,10 Several studies have reported their potential role in activating the initiation of various diseases, particularly malignancies of different origins, including lung, breast, gastric, colon, pancreatic, and liver cancers. Moreover, abnormal circRNA expression has been associated with tumor progression, as it plays an essential role in cell cycle progression and proliferation. Therefore, circRNAs may serve as diagnostic biomarkers for early detection.13,15,16,18,19 Regarding HCC, an increasing number of upregulated or suppressed circRNAs have been reported to play significant roles in regulating the pathophysiological processes, including hsa_circ_101555, hsa-circ-000996, hsa-circ-101094, hsa-circ-100053, hsa-circ-104760, hsa-circ-0064286, and hsa-circ-0000475.23,26–30
Interestingly, hsa_circ_101555 has been specifically reported to be upregulated and highly expressed in HCC cell lines as well as diseased liver tissues compared to adjacent non-tumor tissues. Various studies have explained the mechanism through which circular RNA hsa_circ_101555 promotes tumor cell proliferation and migration, including sponging miR-145-5p, regulating CDCA3 expression, and competing as an endogenous RNA of miR-597-5p.23,31,32 Their bioinformatic role in diseased liver tissue and HCC was investigated by Gu et al.23 The researchers observed, among the top 25 upregulated DEcircRNAs, that hsa_circ_101555 was conserved and significantly upregulated in all three GSE datasets (GSE7852, GSE94508, and GSE97322, downloaded from the Gene Expression Omnibus database), with its high expression associated with HCC progression. Circular RNA hsa_circ_101555 functions as a tumor promoter by sustaining HCC cell proliferation, migration, and invasion through the miR-145-5p/CDCA3 signaling axis. It acts as a competing endogenous RNA by sponging miR-145-5p, thereby impeding the miRNA’s suppression of the oncogene CDCA3. Moreover, EIF4A3 (RNA-binding protein) was shown to mediate hsa_circ_101555 biogenesis. These findings suggest that hsa_circ_101555 may serve as a potential novel biomarker for liver disease and HCC. However, the definitive mechanisms by which circRNAs regulate carcinogenesis and tumor progression are still unclear,23,31,32 warranting further in vitro and in vivo experiments to elucidate their functional pathways. Based on these findings, our research studied the potential role of hsa_circ_101555 as a non-invasive marker for HCC diagnosis and progression. First, our study showed that the mean circRNA value was highest in HCC cases (7.66 ± 3.74) compared to healthy controls (1.21 ± 0.96), with excellent diagnostic accuracy indicated by an AUC of 0.984 at the optimal cutoff point of 1.966. Additionally, the prognostic role of the post-intervention circRNA value at the cutoff point of 5.1150 in differentiating between disease progression and regression in our HCC cases, based on RECIST/mRECIST response categories, was indicated by an AUC of 0.891 with a standard error of 0.058, demonstrating good accuracy for using hsa_circ_101555 as a diagnostic tool and a predictor of HCC response. Second, our findings observed a significant relationship between post-intervention circRNA levels and non-invasive markers of liver fibrosis and function, with robust correlations for FIB-4, APRI, and NFS, suggesting that it may serve as an indicator of liver disease severity in HCC patients after intervention. Third, post-intervention circRNA values were significantly influenced by several clinical and pathological features, particularly those related to tumor burden, vascular invasion, and liver function status. Lastly, the data on the definitive function and mechanism of serum circRNA hsa_circ_101555 in HCC are still in their infancy. As far as we know, only Gu et al.23 investigated their bioinformatic role in diseased liver tissue and HCC. Our main limitation in this study was the relatively small sample size of the study population owing to the limited availability and high cost of such a tool. However, the scarce data on the role of circular RNA hsa_circ_101555, particularly in Egypt, encouraged our team to further evaluate this biomarker in our HCC patients and assess its potential utility as a promising novel marker in this area. However, further studies on larger scales are needed to support our findings.
Conclusions
Hepatocellular carcinoma is considered one of the most common primary liver malignancies, with high morbidity and mortality, so diagnosing HCC at earlier stages is associated with better prognoses. Since circRNA represents a promising biomarker in this field, our study demonstrated that hsa_circ_101555 was upregulated in Egyptian patients with HCC, showing significant diagnostic accuracy in differentiating HCC cases from healthy subjects. Moreover, its high expression was associated with HCC progression according to the RECIST/mRECIST response categories among our patients, serving as a predictor of disease response and offering the potential for better clinical decision-making in HCC management. Hence, our findings suggest that hsa_circ_101555 may serve as a promising novel biomarker and therapeutic target for HCC. However, further evaluation in larger cohort studies is warranted.
Declarations
Acknowledgement
We would like to express our deepest appreciation to the Tropical Hepatoma Centre, Ain Shams University, Egypt, for giving us the opportunity to become involved in this field.
Ethical statement
Our study was reviewed and approved by the Research Ethics Committee at the Faculty of Medicine, Ain Shams University (FMASU REC, FWA 000017585). All subjects were informed and provided written informed consent, preserving the rights and privacy of patients’ data. The procedures involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments.
Data sharing statement
All data generated or analyzed during this study are included in this published article. If any additional data are needed, it will be available from the corresponding author upon reasonable request.
Funding
None to declare.
Conflict of interest
The author declares that there are no competing interests.
Authors’ contributions
Collected the research data (MS, AT, NB, NK, HD), designed and wrote the manuscript, critical final revision, and editing (NB, AT, MS). All authors have approved the final version and publication of the manuscript.