Article Text

Protocol
Neoadjuvant and/or adjuvant chemotherapy for gastric cancer patients with microsatellite instability or deficient mismatch repair: a systematic review and meta-analysis study protocol
  1. Baike Liu1,
  2. Xiaonan Yin1,
  3. Zhaolun Cai1,
  4. Chaoyong Shen1,
  5. Tianxiang Jiang1,
  6. Yihui Han1,
  7. Yuan Yin1,2,
  8. Bo Zhang1
  1. 1 Gastric Cancer Center, Department of General Surgery, West China Hospital, Sichuan University, Chengdu, People's Republic of China
  2. 2 Department of Gastrointestinal Surgery, West China Xiamen Hospital of Sichuan University, Xiamen, People's Republic of China
  1. Correspondence to Dr Bo Zhang; zhangbo_scu{at}scu.edu.cn; Dr Yuan Yin; yinyuan10{at}hotmail.com

Abstract

Introduction Whether gastric cancer (GC) patients with deficient mismatch repair or microsatellite instability-high (dMMR/MSI-H) benefit from perioperative (neoadjuvant and/or adjuvant) chemotherapy is controversial. This protocol delineates the planned scope and methods for a systematic review and meta-analysis that aims to compare the efficacy of perioperative chemotherapy with surgery alone in resectable dMMR/MSI-H GC patients.

Methods and analysis This study protocol is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols-P guideline. PubMed, Embase, Cochrane (CENTRAL), and the Web of Science databases will be searched, supplemented by a secondary screening of relevant records. Both randomised controlled trials and non-randomised studies will be included in this study. The primary and secondary outcomes under scrutiny will be overall survival, disease-free survival and progression-free survival. Two reviewers will independently screen studies, extract data and assess the risk of bias. We will analyse different treatment settings (eg, neoadjuvant or adjuvant or combined as perioperative chemotherapies) separately and conduct sensitivity analyses.

Ethics and dissemination No ethics approval is required for this systematic review and meta-analysis, as no individual patient data will be collected. The findings of our study will be published in a peer-reviewed journal.

Prospero registration number CRD42023494276.

  • SURGERY
  • Gastrointestinal tumours
  • CHEMOTHERAPY
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • This protocol ensures that we are transparent with the whole process for this study to reduce the possibility of duplication as well as potential bias in our study.

  • Several subgroups (stages II and III) and sensitivity (using multivariable-adjusted results) analyses will boost the robustness of our study.

  • The Cochrane Collaboration’s tool and Risk of Bias In Non-randomised Studies of Interventions tool will be used to evaluate the quality of evidence and provide comprehensive references to clinical recommendations.

  • The heterogeneities among chemotherapy regimens may limit the evidence quality.

  • Our results may be limited by the quality of eligible studies included.

Introduction

Gastric cancer (GC) ranks as the fourth leading cause of cancer-related death worldwide, which accounted for ~7 70 000 deaths globally in 2020 and is projected to increase to 1.3 million deaths by 2040.1 Surgical resection is the primary curative treatment for patients diagnosed at an early stage., While for patients with locally advanced GC, the fluoropyrimidine and/or platinum-based perioperative (neoadjuvant and/or adjuvant) chemotherapy has been established as the standard of care in clinical management, which significant prolonged the overall survival (OS) compared with surgery alone.2–6 However, approximately 40–60% of GC patients who undergo curative resection and perioperative chemotherapy still experience relapse or metastasis.7–9 Thus, there is a crucial need for identifying patients who may or may not benefit from perioperative chemotherapy.

Recent advancements in molecular research have significantly enhanced our comprehension of the underlying mechanisms and intrinsic characteristics of GC at the genomic, transcriptomic and protein expression levels. Several molecular subtyping systems have been proposed based on these findings. Notably, the microsatellite instability (MSI) subtype has been independently validated by two large consortiums, the Cancer Genome Atlas10 and Asian Cancer Research Group,11 and account for approximately 10–22% of the total GCs.10 12 MSI arises from deficiency in DNA mismatch repair (dMMR) function, either through germline mutations in the genes encoding the MMR enzymes (MLH1, MSH2, MSH6 or PMS2) or via somatic hypermethylation of the MLH1 promoter.13 14 These results in high tumour mutation burden and genetic hypermutability, known as microsatellite instability-high (MSI-H). In recent years, increasing attention has been focused on the debatable question of whether dMMR/MSI-H could serve as a marker of response to perioperative chemotherapy.

Based on the post hoc analysis of the MAGIC15 and CLASSIC16 randomised trial, GC patients with a dMMR/MSI-H status did not benefit from perioperative and adjuvant chemotherapy, or even with a harmful effect. These results were further validated by an individual patient data (IPD) meta-analysis by combining four randomised trials (MAGIC, CLASSIC, ARTIST and ITACA-S).17 However, a retrospective analysis from Kim et al, comprising of 157 MSI-H GC patients revealed that adjuvant chemotherapy improved disease-free survival (DFS) and OS after adjusting for gender, age and pathological type.18 Similarly, Boyers et al also suggest a better outcome in dMMR/MSI-H gastro-oesophageal cancer patients who received chemotherapy.19 More recently, emerging studies, including one prospective analysis, that compare perioperative chemotherapy with surgery alone in dMMR/MSI-H GC patients have been published.20–22 Meanwhile, a previous meta-analysis only focused on adjuvant chemotherapy in dMMR/MSI-H patients and the number of studies (n=7) included is limited.23

Thus, to summarise the conflicting evidence regarding the efficacy of perioperative chemotherapy in GC patients with dMMR/MSI-H, we aim to conduct a systematic review and meta-analysis (including the stratification analyses of neoadjuvant or adjuvant settings separately). Because only a small fraction of dMMR/MSI-H GC patients in the overall GC population may limit the explanation of studies from a single institute, we believe our study will significantly improve the total sample size of dMMR/MSI-H GC patients in the analysis. The findings derived from this study will provide further evidence for precision clinical management of GC and may contribute to future clinical recommendations.

Methods and analysis

Protocol and registration

The current protocol has been registered with the PROSPERO platform (CRD42023494276) and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P) guidelines24 (PRISMA-P checklist included in online supplemental file 1. This systematic review and meta-analysis will be reported according to the PRISMA 2020 statement and will adhere to the standard methodology recommended by the Cochrane Collaboration.25–27

Supplemental material

Inclusion criteria

The inclusion criteria are described under the population, intervention, comparators, outcomes and study design framework.28 There will be no limitations on language or publication year. The inclusion criteria for studies are as follows:

Population: Patients diagnosed with GC based on a histological examination. All patients will be considered, without restrictions based on country/region, ethnicity, age, gender or occupation.

Intervention: Underwent surgery with curative intent and with perioperative chemotherapy (including neoadjuvant and/or adjuvant chemotherapy). In addition, neoadjuvant or adjuvant chemotherapy will be analysed separately. All regimens of chemotherapy will be considered.

Comparison: Surgery alone.

Outcomes and measurement: The primary outcome is the OS, which is defined as the time to death from any cause. Secondary outcomes are DFS, defined as the time to recurrence of disease (or death) after curative-intent surgery, or progression-free survival (PFS), defined as the time to progression or death. We will present time-to-event data as the HR with 95% CI.

Study design: Suitable study designs will encompass non-randomised studies (NRS) such as case-control studies and cohort studies, as well as randomised controlled trials (RCTs).

Other inclusion criteria: Eligible studies must have a minimum of 24 months of follow-up, reporting survival (time-to-event) outcomes. Additionally, these studies need to provide adequate data to compute or estimate HR and 95% CI.

Exclusion criteria

The exclusion criteria include the following: (1) studies with overlapping populations or results; (2) meeting abstracts, letters, case reports/series, reviews or non-clinical studies lacking available data; (3) patients received the intervention other than chemotherapy, such as chemoradiotherapy; (4) patients received palliative resections; (5) data are missing or cannot be retrieved after reasonable contact with the corresponding author; (6) article full-text cannot be acquired.

Information sources and search strategy

We will systematically search the PubMed, Embase, Cochrane (CENTRAL) and Web of Science databases from their inception up to 1 March 2024. The specific search strategy for each database is listed in online supplemental file 2. Additionally, we will manually review the references of all included articles to identify further studies that meet the eligibility criteria, and their full texts will be obtained.

Study selection and data extraction

All retrieved study records will be imported into Zotero software (https://www.zotero.org/). After removing duplicates, two reviewers will independently evaluate all articles based on the specified eligibility criteria. Initially, the two reviewers will screen titles and abstracts. Then, they will independently reassess the full texts of the identified studies, confirming the rationale for inclusion and exclusion. The screening process will be presented in a PRISMA flow diagram.25

Data extraction for the included studies will be independently conducted by two authors using a standardised electronic data extraction form that has been discussed and agreed upon by all reviewers. The following information will be extracted, such as country/region, chemotherapy regimens, outcomes and median follow-up time. Of note, when multiple studies are conducted on the same subjects and report the same outcomes, only the study with the largest number of cases will be included. Any discrepancies during the process of study selection and data extraction will be resolved through team discussion.

Dealing with missing data

In cases where a study does not provide the HR and its 95% CI, we will reach out to the corresponding author via email to request the missing data. If no response is received within 7 days, we will attempt to estimate some or all of the lnHR, the log-rank observed minus expected events (O-E), the log-rank variance and the variance of the lnHR using indirect methods.29 Of note, when multiple groups were analysed in one study, and the HR for the intervention and comparison group were reported using another group as a reference, the HR and its 95% CI for the intervention group (using the comparison group as a reference) will be calculated as suggested by Woods et al.30 If these indirect methods are also inapplicable, we will estimate HR based on Kaplan-Meier curves using IPDfromKM (https://biostatistics.mdanderson.org/shinyapps/IPDfromKM/)31 or KMtoIPD32 in R software. Generally, if Kaplan-Meier curves and the information of the number at risk were available, IPD would be extracted using graph digitizer software (WebPlotDigitizer V.4.6, https://github.com/ankitrohatgi/WebPlotDigitizer). Next, the output data from WebPlotDigitizer will be uploaded in the IPDfromKM platform or KMtoIPD in R software to estimate the HR and 95%CI by using Cox proportional hazard regression models.33

Risk of bias assessment

Two separate reviewers will appraise the methodological quality and potential bias of the incorporated studies. Any discrepancies will be addressed through team discussions. RCTs will undergo a risk of bias evaluation using the Cochrane Collaboration tool.34 NRS will be evaluated using the Risk of Bias In Non-randomised Studies of Interventions tool.35

Assessment of publication biases

Funnel plots will be applied to investigate potential publication bias if more than 10 studies were pooled in an analysis. To ascertain the statistical significance of publication bias, we will employ Egger’s test, and a p-value <0.05 is indicative of a statistically significant publication bias.

Data analysis

For feasible studies, the collected data will be analysed using RevMan software (V.5.4.1). Time-to-event outcomes will be assessed by pooling HR and their corresponding 95% CI. A two-sided p<0.05 will be considered statistically significant. Unless otherwise specified, HR derived from multivariate analysis will be included by default. In the absence of multivariate values, univariate data will be employed.

Heterogeneity will be examined using the Cochrane Q-test, and the I2 statistic. Significance for heterogeneity was defined as a p-value <0.10 and an I2 statistic >50%. Anticipating a certain degree of heterogeneity among studies, we will employ a random-effects model as a default approach. Our assumption is that the studies to be included are not uniformly estimating the same intervention effect, and these intervention effects are expected to conform to a normal distribution across the studies.36 Furthermore, an analysis of the included studies will consider their design and characteristics. This study will start on 1 March, and the expected end time is 1 July.

A priori subgroup analyses

If there are multiple studies presenting homogeneous outcomes within the specified subgroups, the planned subgroup analyses (perioperative chemotherapy in stage-II and stage-III diseases) for the primary and secondary outcomes will be conducted. Given the chemotherapeutic cycles ranging from 2 to 6 for neoadjuvant or adjuvant chemotherapy,37 and the duration of chemotherapy may influence the efficacy. Thus, we plan to conduct a subgroup analysis based on different therapy cycles (ie, no more than two cycles and more than two cycles of chemotherapy in neoadjuvant and/or adjuvant settings) if relevant data are provided.

Sensitivity analysis

Given the potential impact of confounding factors such as age, gender, histopathological type and tumour stage that may influence the interpretation of results in the individual analysis, a sensitivity analysis will be performed by including only studies reporting the results from multivariable analyses that are adjusted for the aforementioned factors. If the number of studies that need to be estimated is relatively large (>50%), sensitivity analysis will be conducted using only reported data from original studies.

Declarations

Ethics and dissemination

No ethics approval is required for this systematic review and meta-analysis, as no individual patient data will be collected. The findings of our study will be published in a peer-reviewed journal.

Ethics statements

Patient consent for publication

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • BL and XY contributed equally.

  • Contributors BL and YY conceptualised this research. BL, XY, ZC, CS, TJ and YH contributed to the formulation of eligibility criteria, development of the search strategy, design of data extraction methods and the plan for data analysis strategies. BL and XY registered the protocol on the PROSPERO platform and drafted the manuscript. The manuscript underwent review and revision by BZ and YY. The final version of the manuscript received approval from all authors.

  • Funding This study was supported by the National Nature Science Foundations of China (Grant number 82203108), China Postdoctoral Science Foundation (Grant number 2022M722275) and the Key R&D Program of Sichuan Province, China (Grant number 2023YFS0278).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.