The association between single nucleotide polymorphisms and ovarian cancer risk: A systematic review and network meta‐analysis

Jia Hu, Zhe Xu, Zhuomiao Ye, Jin Li, Zhinan Hao, Yongjun WangCancer 2022 · DOI

TL;DR

This study conducted a comprehensive systematic review and network meta-analysis to identify single nucleotide polymorphisms (SNPs) associated with ovarian cancer (OC) risk. They found that the Fok1 rs2228570 SNP, particularly under a dominant gene model, is strongly linked to OC risk, suggesting its potential as a predictive marker.

What Problem Does This Paper Address?

Ovarian cancer (OC) is a significant health concern, ranking third in incidence and highest in mortality among gynecological cancers. While several risk factors are known, including lifestyle and genetic predispositions, identifying specific genetic markers for early detection and risk assessment remains a challenge. Single nucleotide polymorphisms (SNPs), common genetic variations, have been implicated in OC risk, but previous studies have often lacked representativeness and yielded inconsistent results. This paper aims to address this by systematically reviewing and analyzing a broad range of SNPs to pinpoint those most reliably associated with OC risk and to determine the most appropriate genetic model for their assessment.

Methods

The researchers employed a multi-faceted approach combining a systematic review, pairwise meta-analysis, network meta-analysis, and diagnostic meta-analysis.

  • Systematic Review & Pairwise Meta-Analysis: They searched multiple databases for studies investigating the association between SNPs and OC risk, published up to April 2021. Inclusion criteria focused on case-control studies with histopathologically confirmed OC. Pairwise meta-analyses were conducted to determine the association between individual SNPs and OC risk using either fixed or random effect models, depending on heterogeneity.
  • Network Meta-Analysis: To compare multiple SNPs and their gene models simultaneously, a network meta-analysis was performed using R software. This allowed for a more comprehensive evaluation of associations.
  • Gene Model Selection: Six candidate gene models (allele, homozygous, heterozygous, dominant, recessive, and additive) were evaluated. The Thakkinstian's algorithm was utilized to select the most appropriate gene model for each SNP, avoiding a priori assumptions about the genetic model.
  • Credibility Assessment: The False Positive Report Probability (FPRP) and Venice criteria were used to assess the credibility and epidemiological evidence of the identified associations.
  • Diagnostic Meta-Analysis: A diagnostic meta-analysis was conducted using Meta-DiSc software to evaluate the diagnostic accuracy of the most promising SNP and its gene model.

Datasets

The study utilized data from 128 published studies, encompassing 155,014 OC patients and 230,366 non-cancer controls. No new datasets or benchmarks were introduced; the study focused on synthesizing existing research.

Key Insights and Ideas

  • Fok1 rs2228570 Emerges as a Strong Candidate: The study identified Fok1 rs2228570 as a particularly significant SNP associated with OC risk, outperforming other identified SNPs.
  • Dominant Gene Model is Key: The dominant gene model for Fok1 rs2228570 was determined to be the most suitable for estimating OC susceptibility, indicating that having at least one copy of a specific allele confers increased risk.
  • Credibility is Crucial: The rigorous application of FPRP and Venice criteria helped to filter out potentially spurious associations and highlight those with stronger epidemiological support.
  • Network Meta-Analysis for Comprehensive Comparison: The use of network meta-analysis allowed for a more robust comparison of multiple SNPs and their genetic models than traditional pairwise meta-analyses.
  • Vitamin D Receptor Link: Fok1 rs2228570 is located in the vitamin D receptor (VDR) gene, suggesting a potential biological pathway through which this SNP influences OC risk, aligning with ongoing research into vitamin D's role in cancer.

Key Results

  • A total of 63 genes and 92 SNPs were included after careful screening.
  • The Fok1 rs2228570 SNP, under a dominant gene model, showed a pooled OR of 1.158 (95% CI: 1.068-1.256) with a FPRP <0.2, indicating moderate credibility.
  • Subgroup analysis revealed that the dominant gene model of Fok1 rs2228570 was associated with OC risk in both Caucasian and Asian populations.
  • The diagnostic meta-analysis for the dominant gene model of Fok1 rs2228570 showed a pooled sensitivity of 0.383 and specificity of 0.651, with an AUC of 0.5270, suggesting it may be more useful as an auxiliary screening tool rather than a standalone diagnostic marker.

Limitations and Future Work

The authors acknowledge several limitations:

  • Small Sample Sizes in Some Studies: Some included studies had relatively small targeted populations, potentially affecting the representativeness and reliability of the findings.
  • Data Quality Variability: The STREGA scores indicated that the data quality of some studies was poor, which could decrease the credibility of the results.
  • Potential for Omissions: The manual screening process for SNPs might have led to some omissions.
  • Limited Asian Population Data: While the study found associations in Asian populations, the limited number of studies in this group warrants further exploration.

Future work suggested includes:

  • Larger, Higher-Quality Studies: The need for more extensive and well-designed studies to confirm the findings.
  • Exploring Synergistic Effects: Investigating the combined effects of genetic factors and other risk factors (e.g., lifestyle) for a more comprehensive prediction of OC risk.
  • Further Mechanistic Studies: Delving deeper into the biological mechanisms by which Fok1 rs2228570 and other identified SNPs influence OC development.
  • Combining Multiple SNPs: Exploring the potential for synergistic outcomes when combining multiple SNPs for OC prediction.

Summary updated

Abstract

Fok1 rs2228570 is closely linked to OC risk, and the dominant gene model is likely the most appropriate model for estimating OC susceptibility.
BibTeX
@inproceedings{hu2022association,
  author = {Jia Hu and Zhe Xu and Zhuomiao Ye and Jin Li and Zhinan Hao and Yongjun Wang},
  title = {{The association between single nucleotide polymorphisms and ovarian cancer risk: A systematic review and network meta‐analysis}},
  booktitle = {Cancer},
  year = {2022},
  doi = {10.1002/cam4.4891},
}
Figure