DOI: https://doi.org/10.21203/rs.3.rs-1398104/v1
Background: Since Epstein-Barr virus (EBV)-encoded small RNAs (EBER) has become the standard for detection of EBV, the prognostic role of EBV infection in gastric cancer (GC) remains controversial. The main aim of this meta-analysis was to determine the correlation between EBV infection and the prognosis of GC.
Methods: We systematically searched PubMed and EMBASE to identify eligible studies to review up to June 23,2021. Pooled hazard ratios with 95% confidence intervals were calculated using the fixed effects model or random-effects model for overall survival (OS). Sensitivity and subgroup analyses were also performed, and publication bias was assessed by using Stata 14.0.
Results: 35 studies comprising 26163 patients with GC were included in this meta-analysis. After pooling all studies, EBV-positive patients were noted to have better OS than EBV-negative patients (HR: 0.72, 95% CI: 0.59-0.84, P < 0.001), though with high heterogeneity (I2: 51.7%). In subgroup analysis, similar conclusion can be drawn for patients in Asian and American, while for patients in Europe EBV infection is not a prognostic indicator. Apart from that, studies conducted in Asia and Europe, statistical method of multivariate analysis, high quality score, all had significant heterogeneity. Pooled data based on sensitivity analyses did not change the conclusion.
Conclusions: EBV infection was associated with better OS in patients with GC. These results differed between studies due to differences in region, which need further study to clarify its prognostic significance.
Prospero registration number: CRD42021262332.
Gastric cancer (GC) is the fifth most common tumor and there were 768,793 deaths due to GC globally[1]. Although the effectiveness of GC treatment has improved in recent years and has prolonged the overall survival (OS) of patients, survival rates for GC remains poor[2]. Therefore, it is necessary to explore the potential indicators to evaluate the prognosis of GC patients.
Epstein-Barr virus (EBV) is a gamma virus that was originally discovered in Burkitt lymphoma in 1964[3]. EBV infection can be seen in nasopharyngeal carcinoma, infectious mononucleosis, GC, etc. EBV infection can be detected by EBV-encoded small RNAs (EBER) in about 10% of GC patients[4]. The Cancer Genome Atlas divides the molecular subtypes of GC into four categories: EBV-positive, microsatellite instable, genomically stable, and chromosomal instable[5]. EBV-positive patients had a unique clinicopathological characteristics, such as more common in men[6]. A previous meta-analysis in 2015 found that EBV-positive patients had a better OS, which included studies using non-EBER detection methods [7]. However, since EBER has become the gold standard for detection[8], the prognostic of EBV infection remains controversial in different studies recently[9].
Therefore, in order to explore prognostic significance of EBV infection in patients with GC, we launched this meta-analysis.
Search strategy and selection criteria
The meta-analysis was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[10]. The PROSPERO registration number was: CRD42021262332. Two databases (PubMed and EMBASE) were used to explore literatures. All the English publications before 23 June 2021 will be searched. Medical Subject Headings words were the following keywords: “Epstein-Barr virus”, “stomach cancer”, “prognosis” and “survival”. We screened carefully to avoid overlapping studies or duplicate data.
The eligibility criteria of the studies were as follows:
1) Diagnosis of GC confirmed by histology
2) A reliable detection method of EBER for EBV;
3) Assess the relationship between EBV infection and OS and provide a hazard ratio (HR) and 95 % confidence interval (CI) or enough data for estimating the HR and 95 % CI.
Duplicate publications, reviews, editorials, meetings, case reports, comments and animal studies were excluded.
Data extraction
Data were collected by two reviewers (Z-ZH and W-ZX) independently. Any disagreement will be resolved by discussion until consensus is reached by consulting a third author. The following data information were extracted from each eligible literature: 1st author, year of publication, countries, number of patients, HR and 95 % CI for OS by univariate and multivariate survival analyses.
Quality assessment
The Newcastle-Ottawa Scale (NOS)[11] was applied to assess the quality of all selected article. The full score of NOS is 9 points, and studies with scores > 6 points were recognized as high-quality. The evaluation of quality was also conducted by 2 independent reviewers (Z-ZH and W-ZX), and disagreements were resolved by another reviewer (L-GH). Our study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University.
Statistical analysis
Fixed effects model was employed and when there was no obvious heterogeneity, random effects model would be performed. Heterogeneity was calculated by Cochrane Q test and I2 statistic, and PQ > 0.05 and I2< 50 % were considered low heterogeneity in the selected researches. Funnel plot and Begg’s test were applied for evaluation of publication bias. Sensitivity analysis was launched to investigate the effect of individual study on the pooled HR of OS. Kaplan-Meier curves of OS were extracted by Engauge Digitizer 4.1. All p values were 2-sided tests and p value less than 0.05 was considered significant difference.
Literature search
1407 studies were retrieved from the databases. 411 articles were excluded because of duplication. After screening the titles and abstracts, 853 articles were excluded and 143 potentially eligible studies were selected for further evaluation. Finally, 35 studies[12-46] were selected for meta-analysis. The literature screening process was presented as PRISMA flowchart (Figure 1).
Study characteristics
Among the 35 studies investigated the influence of EBV infection on patient’s OS, 24 studies were performed in Asia, 7 in Europe, 4 in America. NOS score of 26 studies were ≥7 points while that of other 9 studies were＜7 points. The sample size among the studies varied from 58 to 8819. The baseline characteristics of the studies were shown in Table 1.
Meta-analysis and sensitivity analysis
After pooling all included studies, patients with GC with EBV infection were noted to have better OS than EBV-negative patients (HR: 0.72, 95% CI: 0.59-0.84, P < 0.001) (Figure 2). Significant heterogeneity among the all studies was observed (I2: 51.7%; P < 0.001). Therefore, we conducted sensitivity analyses and subgroup analysis to assess the source of heterogeneity.
Sensitivity analyses
Sensitivity analysis of the OS was launched to determine the effects of each study on the pooling HR. The result of this analysis discovered that the omission of each study did not significantly change the pooled HR of OS, illustrating the robustness of our analysis result (Figure 3).
Subgroup analysis
Series of subgroup analyses were conducted on the basis of region, quality of study, statistical method (Figure 4). The result of stratification showed that studies conducted in Asia and Europe, statistical method of multivariate analysis, high quality score, all had significant heterogeneity, while the other subgroups had low heterogeneity. Studies conducted in Asian and American showed that EBV infection was a protective factor in OS (P < 0.05) but studies conducted in Europe showed no survival difference (P > 0.05) (Table 2).
Publication Bias
The results from Begg’s funnel plot (p=0.712) and Egger’s test (p=0.277) indicated that there was no significant publication bias in our meta-analysis (Figure 5).
Accumulated studies have been performed to assess the correlation between EBV infection and the prognosis of GC[47]. Some studies have reported that EBV infection predicted better OS in GC patients after surgery[27, 48]. On the contrary, some recent studies still draw different conclusions which showed worse prognostic value of EBV infection for GC patients[35, 49] .Thus, the prognostic value of the EBV remains inconclusive in GC.
Usually, the gold standard of EBV detection method is EBER with accurate localization and strong specificity[50]. The positive rate of PCR can be up to 80%[51], which were higher than that of EBER (about 10%). However, PCR or RNA sequencing may be not reliable due to the presence of intratumoral and peritumoral lymphocytes which may have EBV infection and this limitation cannot be solved by microdissection[52]. Therefore, the present meta-analysis was conducted to investigate the prognostic value of EBV infection with detection method of EBER in GC patients.
35 studies involving 26163 patients in our study were included, and the results showed that EBV-positive patients had a better OS than EBV-negative patients (HR: 0.72, 95% CI: 0.59–0.84, P < 0.001), which was similar to the results of a previous international pooled analysis[53], indicating that EBV-positive GC patients have a better prognosis. The better prognosis of EBV-positive patients may be due to tumor microenvironment and chemotherapy[54].
Some studies have shown that there was no survival difference in EBV-positive patients and EBV-negative patients after surgery and chemotherapy, but EBV-positive patients performed better to chemotherapy with 5-fluorouracil and platinum[55, 56]. Although it has not been proven, EBV-related antigens may help enhance effects of chemotherapy-induced neoantigens and improve survival [57].
EBV-positive GC has a unique tumor microenvironment. EBV-positive GC contains abundant CD8 + tumor-infiltrating lymphocytes (TILs), which enhanced anti-tumor immunity by eliminating EBV-positive tumor cells[58]. In addition, the Asian Cancer Research Group (ACRG) classification also supported the good prognosis of EBV-positive GC patients, because EBV-positive GC was more often included in the microsatellite stable/TP53+ (MSS/P53+) subtype, and together with microsatellite unstable (MSI) adenocarcinoma, which showing the best survival rate[59]. Therefore, EBV-positive GC patients seems to have a better prognosis, but the specific mechanism still needs to be studied.
Our research showed that the prognosis of EBV-positive patients was better than EBV-negative patients among Asians and Americans, while no difference was detected for patients in European. Inconsistent with our research, Xuechao Liu’s study found that the prognosis of EBV-positive patients in Americans were similar to the prognosis of EBV-negative patients[7]. The inconsistency of prognostic effects among American studies may be due to the fact that we did not include an article reporting Risk Ratio and an affiliated study as Xuechao Liu’s study reported[7]. So far, the root cause of these regional differences is still unclear. One possible reason is that phylogeographic diversity of EBV strains in genetic difference (especially in the BamHI–I fragment and the XhoI restriction sites) cause different survival effects[60].
Stratified subgroup analyses were also conducted by statistical method. We discovered that the protective effect of EBV infection still existed both in univariate analysis and multivariate analysis, similar to the results of an international pooled analysis of 13 studies[53]. We included 35 studies, strengthening the reliability that patients with EBV infection had a better prognosis.
There were also several limitations in our present study. Firstly, studies published with the English language only were included in this meta-analysis, and thus potential publication bias existed. Secondly, the studies included were mostly retrospective, which might be more susceptible to biases. Thirdly, we calculated HR and 95% CI from the overall survival curves when the article did not provide them, which might be a potential source of bias. Therefore, large-scale multicenter and prospective researches are needed.
All in all, EBV infection was associated with better OS in patients with GC. These results differed between studies due to differences in region, which need further study to clarify its prognostic significance.
Ethics approval and consent to participate: Not applicable.
Consent for publication: Not applicable.
Availability of data and materials: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing interests: The authors declare that they have no competing interests.
Funding: This study was funded by the National Natural Science Foundation of China (Grant Nos. 81602049 and 81802342), the Natural Science Foundation of Guangdong Province, China (Grant No. 2018A030313978) and the Kelin New Star of the First Affiliated Hospital of Sun Yat‐Sen University (Grant Nos. R08011 and R08010).
Authors' contributions: All authors contributed to the study conception and design. ZW designed this study. ZHZ, ZXW and GHL performed the analysis and writing. ZHZ and JHH contributed to data collection. All authors read and approved the final manuscript.
Acknowledgements: Not applicable.