Introduction
Circular RNAs (circRNAs), characterized by their covalently closed-loop structure formed through back-splicing of introns or exons, represent a novel category of noncoding RNAs.1 Increasing evidence indicates that circRNAs are more stable and abundantly expressed than their linear counterparts and possess potential as diagnostic and prognostic biomarkers for various diseases.1,2 Among the extensively studied circRNAs, circPVT1 (circBase ID: hsa_circ_0001821) has emerged as a key player in cancer biology. Derived from an exon of the plasmacytoma variant translocation 1 (PVT1) gene, circPVT1 is located on chromosome 8q24—a genomic region widely recognized for its association with cancer susceptibility.3 Unlike linear RNAs, circPVT1 forms a covalently closed loop, rendering it highly stable and resistant to RNA degradation. This structural integrity allows circPVT1 to function as a dynamic regulator of gene expression, primarily by acting as a microRNA (miRNA) sponge. By sequestering miRNAs, circPVT1 modulates the activity of downstream target genes, impacting essential cellular processes such as proliferation, apoptosis, and metastatic progression.4
Emerging studies have identified circPVT1 as a promising prognostic marker across multiple cancer types. For instance, in hepatocellular carcinoma, elevated circPVT1 expression is associated with worse overall survival (OS) and disease-free survival, supporting its potential as a clinical outcome predictor.5 Additionally, Wang et al. revealed that circPVT1 is upregulated in breast cancer, where it enhances tumor cell invasion and metastasis by regulating the miR-29a-3p/AGR2/HIF-1α pathway.6 However, conflicting findings have also been reported. Kong et al. discovered that circPVT1 expression is reduced in gastric cancer (GC), with lower levels correlating with deeper tumor invasion and lymph node metastasis.7 These discrepancies highlight the complexity of circPVT1’s role in cancer progression and underscore the necessity for a comprehensive evaluation of its prognostic and clinicopathological significance.
Given the growing interest in circPVT1 as a prospective marker and the inconsistent evidence regarding its clinical predictive value across different cancers, we executed a detailed meta-analysis of available clinical research. This analysis aimed to provide a clear and comprehensive assessment of circPVT1’s prognostic value and its association with clinicopathological features in solid tumors, delivering valuable insights to guide future research and clinical strategies.
Materials and methods
Search strategy
This study was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines.8 An exhaustive literature search was performed by two authors (Menglan Li and Kai Qian) across multiple databases, including PubMed, Web of Science, Embase, the Cochrane Library, and CNKI, with a cutoff date of December 31, 2024. The following search terms were utilized: (‘circPVT1’ or ‘circular RNA PVT1’ or ‘hsa_circ_0001821’) and (‘prognosis’ or ‘prognostic’ or ‘survival’ or ‘characteristic’) and (‘tumor ‘or ‘cancer’ or ‘carcinoma’). During the evaluation process, the reference lists of the selected articles were also meticulously reviewed to identify additional relevant studies. This comprehensive approach ensured the inclusion of all pertinent literature examining the role of circPVT1 in tumor prognosis and clinicopathological features.
Study enrollment criteria
Inclusion criteria: (1) Clinical studies exploring the impact of circPVT1 expression on clinicopathological characteristics or survival outcomes in tumor patients; (2) Articles providing hazard ratios (HRs) and confidence intervals (CIs), or survival curves from which HRs could be indirectly calculated; (3) Full-text articles available for review; (4) Literature published in English or Chinese.
Exclusion criteria: (1) Studies lacking usable or sufficient data on cancer prognosis or clinicopathology; (2) Research focusing solely on the molecular mechanisms of circPVT1 without clinical data; (3) Duplicate publications, review articles, letters, comments, and conference abstracts, to avoid redundancy and ensure depth of analysis.
Data extraction and quality assessment
To ensure accuracy and comprehensiveness, two researchers (Menglan Li and Kai Qian) independently extracted data, with discrepancies resolved through consultation with a third party (Zhixian Zhu). The extracted data included: (1) The first author’s name, publication year, country, cancer type, target microRNA, sample type and size, detection methods, cutoff value, circPVT1 expression status, source of HRs, survival outcomes, and follow-up duration; (2) Clinicopathological variables, including gender, tumor size, grade, lymph node metastasis, distant metastasis, tumor-node-metastasis (TNM) stage, and pathological T stage; (3) HRs with 95% CIs, either directly reported or calculated from survival curves.9 The Newcastle-Ottawa Scale (NOS), a validated tool for evaluating the quality of non-randomized studies, was used for quality assessment.10
Statistical analysis
All statistical analyses were conducted using STATA 12.0 (StataCorp, College Station, TX, USA). To evaluate the prognostic and clinicopathological significance of circPVT1 across various cancer types, pooled HRs and ORs with corresponding 95% CIs were calculated. HRs were either directly extracted from the original studies or derived from Kaplan-Meier curves using Engauge Digitizer software (version 4.1). Statistical heterogeneity was examined using Cochrane’s Q test. A fixed-effect model was applied when heterogeneity was low (I2 < 50% or P > 0.1); otherwise, a random-effects model was used. Subgroup analyses were performed to explore potential sources of variation. Sensitivity analyses were conducted to assess the consistency of the results. For analyses involving six or more studies, Begg’s and Egger’s tests were used to detect potential publication bias.11,12
Results
Study selection and characteristics
A total of 312 articles from PubMed, Web of Science, Embase, Cochrane Library databases, and CNKI were initially recruited (Fig. 1). After applying the predetermined inclusion and exclusion criteria, 216 duplicate and irrelevant articles were excluded. Additionally, 14 reviews, three meta-analyses, four conference abstracts, and four comments were removed. Following a thorough review of titles and abstracts, an additional 35 studies not related to prognosis or clinicopathology aspects of circPVT1 were also excluded. The remaining 36 full-text articles were then meticulously reviewed for relevant data, resulting in the exclusion of nine articles due to inadequate information. Ultimately, 27 articles encompassing 2,219 individuals were included. Among these articles, 23 focused on prognosis5,13–15 and 19 studies mentioned clinicopathology.3,6,16,17 These articles were published between 2017 and 2024, with all but two originating from China. Key information extracted from these studies is summarized in Table 1,3,5–7,13–35 providing a comprehensive overview of circPVT1’s impact on solid tumors.
Table 1Characteristics of included studies in this meta-analysis
| Author | Year | Region | Cancer type | miRNA | Sample type | Sample size | Detection methods | Cutoff value | Expression
| Follow-up (months) | HRs source | Expression tatus | Survival outcome | NOS |
|---|
| low | high |
|---|
| Wang et al.6 | 2020 | China | BC | MiR-29a-3p | tissue | 40 | RT-PCR | median | 20 | 20 | NA | NA | Up | NA | 6 |
| Bian et al.18 | 2020 | China | BC | miR-204-5p | tissue | 99 | RT-PCR | median | 47 | 52 | 60 | Direct | Up | OS | 8 |
| Lu et al.19 | 2020 | China | NSCLC | NA | serum | 96 | RT-PCR | median | 48 | 48 | 100 | Direct | Up | OS | 7 |
| Yan et al.16 | 2020 | China | osteosarcoma | miR-526b | tissue | 48 | RT-PCR | median | 24 | 24 | 60 | Curve | Up | OS | 8 |
| Zheng et al.17 | 2020 | China | LAD | miR-145-5p | tissue | 104 | RT-PCR | mean | 56 | 48 | 60 | Direct | Up | OS | 8 |
| Zhu et al.3 | 2019 | China | HCC | miR-203 | tissue | 70 | RT-PCR | median | 35 | 35 | 60 | Curve | Up | OS | 7 |
| Wang et al.20 | 2019 | China | CRC | miR-145 | tissue | 64 | RT-PCR | median | 32 | 32 | 60 | Curve | Up | OS | 7 |
| Qin et al.21 | 2019 | China | NSCLC | miR-497 | tissue | 70 | RT-PCR | median | 43 | 47 | 60 | Curve | Up | OS | 8 |
| Zhu et al.22 | 2018 | China | osteosarcoma | NA | tissue | 80 | RT-PCR | mean | 50 | 30 | 60 | Curve | Up | OS | 7 |
| Verduci et al.23 | 2017 | Italy | HNSCC | NA | tissue | 106 | RT-PCR | median (X)-σ/2 | 35 | 71 | 70 | Direct | Up | OS | 8 |
| Chen et al.24 | 2017 | China | GC | NA | tissue | 187 | RT-PCR | Youdeng’s index | 80 | 107 | 100 | Direct | Up | OS | 7 |
| Kong et al.7 | 2019 | China | GC | NA | tissue | 80 | RT-PCR | mean | 62 | 18 | NA | NA | Down | NA | 6 |
| Tao et al.25 | 2019 | China | PTC | miR-126 | tissue | 39 | RT-PCR | mean | 18 | 21 | 60 | Curve | Up | OS | 7 |
| Zhou et al.13 | 2024 | China | BLC | NA | tissue | 162 | RT-PCR | mean | 50 | 112 | 97 | Direct | Up | OS | 7 |
| Wang et al.14 | 2022 | China | osteosarcoma | miR-24-3p | tissue | 80 | RT-PCR | median | 40 | 40 | 60 | Curve | Up | OS | 8 |
| Mo et al.15 | 2022 | China | NPC | NA | tissue | 159 | ISH | mean | 31 | 128 | 120 | Direct | Up | OS | 7 |
| Shi et al.26 | 2021 | China | LUSC | miR-30d/e | tissue | 104 | RT-PCR | median | 50 | 54 | 60 | Direct | Up | OS | 8 |
| Chen et al.5 | 2024 | China | HCC | NA | tissue | 96 | RT-PCR | median | 49 | 47 | 50 | Direct | Up | OS | 9 |
| Lyu et al. 27 | 2024 | China | LaC | NA | tissue | 65 | RT-PCR | median | 35 | 30 | 65 | Curve | Up | OS | 7 |
| Zeng et al.28 | 2021 | China | PTC | miR-195 | tissue | 50 | RT-PCR | median | 25 | 25 | 60 | Curve | Up | OS | 6 |
| Wang et al.29 | 2021 | China | GBC | miR-339-3p | tissue | 36 | RT-PCR | median | 17 | 19 | 40 | Direct | Up | OS | 8 |
| Hua et al.30 | 2022 | China | PTC | miR-384 | tissue | 36 | RT-PCR | median | 18 | 18 | NA | NA | Up | NA | 7 |
| Liu et al.31 | 2022 | China | LC | miR-124-3p | tissue | 60 | RT-PCR | median | 30 | 30 | 150 | Curve | Up | OS | 7 |
| Wan et al.32 | 2020 | China | osteosarcoma | miR-423-5p | tissue | 36 | RT-PCR | NA | NA | NA | 50 | Curve | Up | OS | 7 |
| Mai et al.33 | 2019 | China | CRC | NA | plasma | 148 | RT-PCR | NA | 62 | 86 | 60 | Curve | Up | OS | 7 |
| Can et al.34 | 2023 | Turkey | HNCs | NA | tissue | 104 | RT-PCR | median | 53 | 51 | NA | Direct | Up | OS | 7 |
| Qi et al.35 | 2022 | China | ESCC | NA | tissue | 40 | RT-PCR | median | 18 | 22 | NA | NA | Up | NA | 6 |
Associations between circPVT1 expression and OS
Due to the high degree of heterogeneity identified in the studies (I2 = 80.2%, P < 0.001), a random-effects model was employed to calculate the combined HR for overall survival OS. As shown in Figure 2, increased circPVT1 expression levels were strongly linked to poorer OS in solid tumor patients. The analysis revealed a combined HR of 1.68, with the 95% CI spanning from 1.39 to 2.02, indicating a statistically significant association (P < 0.001). Subsequently, we performed the analysis based on different cancer types. As shown in Figure 3, for lung cancer patients, high circPVT1 expression correlated with an HR of 2.08 (95% CI: 1.51–2.88, P < 0.001), indicating nearly a twofold increased risk of adverse OS in this patient population. Similarly, for osteosarcoma patients, high circPVT1 expression was linked to an HR of 1.65 (95% CI: 1.38–1.97, P < 0.001), indicating an approximately 1.7-fold increased risk of poor OS. Similar trends were also observed in hepatocellular carcinoma, colorectal cancer, and papillary thyroid carcinoma. These findings underscore the promising role of circPVT1 as a predictive biomarker across various types of solid tumors, highlighting its value in predicting unfavorable clinical outcomes.
Subgroup analysis
To investigate the source of heterogeneity in the OS outcome, we conducted subgroup analyses based on several factors, including sample size, interaction with miRNA, NOS score, and HR sources. As illustrated in Figure 4, significant heterogeneity was predominantly observed in subgroups with a sample size ≥ 100 (HR = 1.45, 95% CI: 0.98–2.14, I2 = 87.3%), no interaction with miRNA (HR = 1.57, 95% CI: 1.08–2.28, I2 = 89.0%), NOS score ≤ 7 (HR = 1.52, 95% CI: 1.19–1.95, I2 = 85.3%), and OS data sourced directly from articles (HR = 1.63, 95% CI: 1.12–2.37, I2 = 87.1%) (Table 2). In contrast, no heterogeneity was observed in groups with a sample size < 100 (HR = 1.78, 95% CI: 1.60–1.98, I2 = 0.0%), interaction with miRNA (HR = 1.73, 95% CI: 1.54–1.94, I2 = 0.0%), NOS score > 7 (HR = 1.88, 95% CI: 1.54–2.30, I2 = 28.1%), and HRs derived from Kaplan-Meier curves (HR = 1.74, 95% CI: 1.56–1.95, I2 = 0.0%) (Table 2).
Table 2Subgroup analyses of pooled HRs for OS
| Subgroup | Studies (n) | OS
| Heterogeneity
|
|---|
| Pooled HR (95%CI) | P-value | I2 (%) | P-value |
|---|
| Sample size | | | | | |
| <100 | 14 | 1.78 (1.60–1.98) | <0.001 | 0.00 | 0.508 |
| ≥100 | 9 | 1.45 (0.98–2.14) | 0.062 | 83.7 | <0.001 |
| Interacted with miRNA | | | | | |
| miRNA | 13 | 1.73 (1.54–1.94) | <0.001 | 0.00 | 0.784 |
| NA | 10 | 1.57 (1.08–2.28) | 0.019 | 89.0 | <0.001 |
| NOS score | | | | | |
| <7 | 14 | 1.52 (1.19–1.95) | 0.001 | 85.3 | <0.001 |
| ≥7 | 9 | 1.88 (1.54–2.30) | <0.001 | 28.1 | 0.195 |
| HRs source | | | | | |
| Direct | 11 | 1.63 (1.55–2.09) | 0.011 | 87.1 | <0.001 |
| Curve | 12 | 1.74 (1.56–1.95) | <0.001 | 0.00 | 0.878 |
Associations between circPVT1 and clinicopathological characteristics
This association analysis encompassed 19 articles (Table 3). The aggregated findings demonstrated that patients with increased circPVT1 expression levels had a higher risk of larger tumor dimensions (OR = 1.36, 95% CI: 1.11–1.67, P = 0.004), lymph node metastasis (OR = 1.56, 95% CI: 1.22–2.00, P < 0.001), distant metastasis (OR = 1.80, 95% CI: 1.10–2.92, P = 0.017), and advanced tumor TNM stage (OR = 1.84, 95% CI: 1.50–2.25, P < 0.001), indicating its potential as a marker for aggressive clinical pathological features. However, there was no significant evidence showing that abnormal circPVT1 expression was associated with gender, grade, or tumor stage.
Table 3Pooled analysis of circPVT1 expression and tumor clinicopathological characteristics
| Clinicopathological parameters | Articles (n) | Cases (n) | Combined OR (95%CI) | Effects model | P-value | Heterogeneity
|
|---|
| I2 (%) | P-value |
|---|
| Gender (male vs female) | 8 | 667 | 1.06 (0.93–1.21) | Fixed | 0.353 | 3.10 | 0.406 |
| Tumor size (≥3 vs <3) | 14 | 1,230 | 1.36 (1.11–1.67) | Random | 0.004 | 69.8 | <0.001 |
| Grade (high vs low) | 9 | 812 | 0.97 (0.85–1.10) | Fixed | 0.611 | 45.9 | 0.063 |
| Lymph node metastasis (yes vs no) | 16 | 1,345 | 1.56 (1.22–2.00) | Random | <0.001 | 74.8 | <0.001 |
| Distant metastasis (yes vs no) | 6 | 645 | 1.80 (1.10–2.92) | Random | 0.017 | 84.5 | <0.001 |
| TNM stage (III/IV vs I/II) | 15 | 1,284 | 1.84 (1.50–2.25) | Random | <0.001 | 66.3 | <0.001 |
| Tumor stage (III/IV vs I/II) | 6 | 636 | 1.05 (0.71–1.55) | Random | 0.822 | 82.9 | <0.001 |
Sensitivity analysis and assessment of publication bias
The sensitivity analysis indicated that there was no significant impact on the combined HR of OS after systematically eliminating each article, suggesting that these studies were relatively reliable and stable (Fig. 5). To further assess the possibility of publication bias, Begg’s and Egger’s tests were conducted, yielding P-values of 0.082 and 0.063, respectively (Fig. 6). These non-significant P-values indicated that no substantial evidence of publication bias was detected in this meta-analysis, further supporting the validity and generalizability of our findings.
Discussion
Cancer continues to be the primary cause of global mortality, presenting a significant challenge to improving life expectancy worldwide. Given the substantial financial burden associated with cancer care, the identification of dependable prognostic biomarkers is crucial.36 Circular RNAs, known for their circular configuration, constitute an intriguing category of noncoding RNAs. Plentiful investigations have confirmed that circRNAs are abundant, tissue-specific, evolutionarily conserved, and highly stable.37,38 These properties position circRNAs as excellent candidates for cancer prognostic biomarkers.
circPVT1, an endogenous circular RNA derived from the PVT1 gene within the cancer-associated genomic locus 8q24, has garnered significant attention. With a length of 410 nucleotides, circPVT1 is primarily synthesized through an exon circularization mechanism, which relies on complementary sequences in flanking intronic regions and the activity of RNA-binding proteins. Research has demonstrated that circPVT1 is highly expressed in various cancers and influences tumor initiation, progression, and metastasis through multiple mechanisms.39 Its primary functions include acting as a miRNA sponge, regulating gene expression, and modulating cellular processes such as proliferation, apoptosis, migration, and invasion.4 For instance, in non-small cell lung cancer, circPVT1 sequesters miR-124-3p, thereby modulating EZH2 expression, which enhances lung cancer cell proliferation, invasion, and migration.31 In gastric cancer, circPVT1 functions as a molecular sponge for the miR-125 family, restoring the expression of its downstream target gene E2F2 and promoting cell proliferation.24 In osteosarcoma, circPVT1 interacts with miR-423-5p, activating the Wnt5a/Ror2 signaling cascade to facilitate glycolysis and metastasis.32 Additionally, Verduci et al. found that circPVT1 is highly expressed in head and neck squamous cell carcinoma harboring mutant p53 proteins and promotes tumorigenesis and progression through its interaction with the YAP/TEAD complex.23 Studies have also revealed that circPVT1 expression is closely linked to chemoresistance of cancer cells. For example, in osteosarcoma, elevated levels of circPVT1 correlate with resistance to chemotherapy, as it regulates the expression of ABCB1, enhancing the resistance of tumor cells to doxorubicin and cisplatin.22 circPVT1 also holds potential value in tumor diagnosis and prognosis. In gallbladder cancer, circPVT1 expression levels are linked to lymph node metastasis and advanced TNM stage, with higher circPVT1 expression associated with poor patient prognosis.29 Therefore, circPVT1 expression levels might function as a predictive marker to categorize patients into groups at higher or lower risk, guiding more personalized treatment strategies. These findings underscore the critical function of circPVT1 in the advancement and prognostic assessment of various solid cancers. Subsequent investigations ought to center on evaluating the capabilities of circPVT1 as a non-invasive indicator in liquid biopsies for early-stage cancer identification, prognosis prediction, and treatment monitoring. Additionally, developing therapeutic strategies to target circPVT1, such as RNA interference or CRISPR-based approaches, could enhance the efficacy of existing treatments and provide new therapeutic options for cancer patients.
While circPVT1 has been studied extensively,40–42 our systematic review and meta-analysis offer a thorough integration of the existing evidence, particularly focusing on its prognostic and clinicopathological significance across different cancer types. Based on current knowledge, this is the first investigation to quantitatively summarize the prognostic value of circPVT1 across diverse solid tumors, offering a broader perspective on its potential clinical utility. The results from the included studies indicated that upregulated circPVT1 expression was closely linked to unfavorable OS outcomes in cancers, with a pooled HR of 1.68. Moreover, cancers with higher levels of circPVT1 had a greater possibility of larger tumor size, lymph node metastasis, distant metastasis, and advanced tumor TNM stage, indicating that increased circPVT1 levels were a marker of aggressive clinicopathological features. However, there was significant heterogeneity in OS outcomes among the included articles. Although an appropriate effect model was employed during data merging, the origin of heterogeneity among the included studies remained ambiguous. Sensitivity analyses, which typically help in assessing the robustness of the findings and identifying potential outliers, were also inconclusive in explaining the observed heterogeneity. Given these challenges, we carried out subgroup analyses. The results indicated that heterogeneity might exist in the subgroups of sample size ≥ 100, not interacting with miRNA, NOS score ≤ 7, and OS data obtained directly from the articles.
Despite these findings, a number of limitations pertinent to this meta-analysis ought to be highlighted. Firstly, the insufficient number of enrolled studies and patients may have led to conflicting results. For example, Chen et al. showed circPVT1 was enhanced in 187 GC tissues compared with matched normal tissue and exhibited a tumor-promotive function,24 while the results from Kong et al. showed that circPVT1 was significantly downregulated in 80 GC tissues.7 Such discrepancies may be attributed to smaller sample sizes and individual variability. Secondly, the majority of the included studies were conducted in China, with only two papers coming from Italy and Turkey, respectively, which may limit the generalizability of the findings. Therefore, the conclusions should be cautiously interpreted in a broader context. Thirdly, some survival data were calculated from Kaplan-Meier survival curves instead of being directly extracted from the original studies, which might introduce minor discrepancies in the pooled results. Finally, all included studies were retrospective in design, potentially introducing selection bias. Consequently, additional prospective clinical trials are required to confirm the predictive significance of circPVT1 across various tumor types.
Future Directions
To advance the clinical relevance of circPVT1, future studies should prioritize validating its prognostic utility through international, prospective cohorts encompassing diverse ethnicities and cancer subtypes, thereby mitigating current geographical biases. Concurrently, mechanistic investigations are essential to elucidate how circPVT1 drives chemoresistance and metastasis, particularly through its interplay with RNA-binding proteins or immune-modulatory pathways in tumor microenvironments. Translational efforts must focus on optimizing non-invasive circPVT1 detection in liquid biopsies for early diagnosis and real-time monitoring, coupled with developing targeted therapies such as CRISPR-based silencing or antisense oligonucleotides to counteract treatment resistance.
Conclusions
Our meta-analysis indicated that elevated circPVT1 expression levels are closely linked to unfavorable OS in tumor patients. circPVT1 has the potential to act as both a prognostic indicator and a molecular target for solid tumor therapy. Future research should focus on validating these findings through large-scale, multicenter prospective studies to further establish the clinical utility of circPVT1 in cancer management.
Declarations
Funding
This work was supported by Open Project Support for the National Clinical Research Base of Traditional Chinese Medicine (JD2022SZ11), Developing Program for High-level Academic Talent in Jiangsu Hospital of Chinese Medicine (Grant No. y2018rc38), and Natural Science Foundation Project of Nanjing University of Chinese Medicine (No. XZR 2024010).
Conflict of interest
The authors declare no conflict of interest.
Authors’ contributions
Contributed to study concept and design (ML, PL), acquisition of the data (ML, KQ), data analysis (ML, ZZ, KQ), provision of reagents, materials, and analysis tools (ZZ, KQ, YD), and drafting of the manuscript (ML, PL).