Research question

Is pembrolizumab, used alone or associated with bevacizumab, surgery, radiotherapy, and chemotherapy, as effective and safe as the standard treatments for recurrent and metastatic cervical cancer?

Search strategy

Study investigators and a medical librarian collaboratively developed the search strategy to identify primary studies that examined whether pembrolizumab is effective and safe, either alone or combined with a standard treatment, in the management of cervical cancer.

Eight databases will be searched from inception until July 2024: Pubmed, Embase, Scopus, Web of Science, Cancerlit, The World Health Organization (WHO), International Clinical Trials Registry Platform (ICTRP), Clinical Trials Registry Platform and Cochrane Library. Grey literature will be searched using the following sources: Clinicaltrials.gov, Google Scholar and Opengrey. Additionally, there will be a search for abstracts of randomized clinical trials presented in major international congresses in the last five years. The reference list of included articles and relevant reviews identified through the search will be scanned to determine any potential article that is relevant to this review.

The search strategy will be designed with a combination of MESH/EMTREE terms and synonyms, being adapted for each database. There will be no restrictions regarding publication date and language.

The Pubmed search strategy will be developed in collaboration among all authors and a health sciences librarian with expertise in systematic review searching. Subsequently, the search strategy will be peer-reviewed by an external health sciences librarian in accordance with the PRESS standard [19]. After acceptance of the Pubmed search strategy, it will be translated to the other databases. The draft of the Pubmed search strategy is as follows: (“Pembrolizumab” OR “SCH-900475” OR “lambrolizumab” OR “MK-3475” OR “Keytruda”) OR (“Bevacizumab” OR “Mvasi” OR “Bevacizumab-awwb” OR “Avastin”) AND (“surgery” OR “radiotherapy OR “chemotherapy”) AND (“uterine cervical neoplasms” OR “cervical neoplasm” OR “cancer of the uterine cervix” OR “cervical cancer” OR “uterine cervical cancer” OR “cervix cancer”) AND (“randomized clinical trial” OR “controlled clinical trial” OR “Trials, Randomized Clinical” OR “intervention study” OR “clinical study” OR “clinical studies”).

After validation of the proposed search strategy for PubMed by the librarian, an adaptation of this strategy was used for EMBASE as follows: (’Pembrolizumab’ OR ’SCH-900475’ OR ’lambrolizumab’ OR ’MK-3475’ OR ’Keytruda’) OR (’Bevacizumab’ OR ’Mvasi’ OR ’Bevacizumab-awwb’ OR ’Avastin’) AND (’surgery’/exp OR ’radiotherapy’/exp OR ’chemotherapy’/exp) AND (’uterine cervical neoplasms’/exp OR ’cervical neoplasm’/exp OR ’cancer of the uterine cervix’/exp OR ’cervical cancer’/exp OR ’uterine cervical cancer’/exp OR ’cervix cancer’/exp) AND (’randomized controlled trial’/exp OR ’controlled clinical trial’/exp OR ’clinical trial’/exp OR ’intervention study’/exp OR ’clinical study’/exp).

Eligibility criteria

In the systematic review, studies that meet the following criteria will be included: randomized clinical trials, which evaluated the efficacy and safety of pembrolizumab, used alone or combined with chemotherapy, radiotherapy, bevacizumab or surgery, in all adult women diagnosed with cervical cancer that received pembrolizumab treatment, from IB2 tumoral stage according to the classification of the International Federation of Gynecology and Obstetrics (FIGO) [20]. There will be no restrictions regarding the follow-up time and setting of the studies.

Only original studies will be eligible for inclusion in this study, including completed and ongoing clinical trials with preliminary results. Studies that involve mixed populations (e.g., cervical cancer patients along with other cancer types), case reports, preprint, narrative and systematic reviews and observational studies will be excluded.

Studies will be selected based on the following eligibility criteria:

1. P(Population): Adult women diagnosed with advanced and high-risk stage IB2 cervical cancer onwards who had been diagnosed and previously treated, or not, with pembrolizumab alone, or in combination with bevacizumab, surgery, radiotherapy, and chemotherapy.
2. I(Intervention): Pembrolizumab used either alone or combined with Bevacizumab, chemotherapy, surgery and other therapies.
3. C(Comparison): Surgery, radiotherapy, platinum-based chemotherapy, bevacizumab and placebo.
4. O(Outcomes): Mortality rate as primary outcome. Overall survival (OS), disease-free survival (DFS), complete remission (CR), partial response (PR), stable disease (SD), disease progression (DP), overall response rate (ORR), disease control rate (DCR), median progression free survival (mPFS), median OS (mOS), and adverse events associated with the use of pembrolizumab will be considered as secondary outcomes.
5. S (Study design): Randomized clinical trials.

Definition of measures

Mortality rate: number of deaths, with cancer as the main cause of death, in a population affected by cervical carcinoma.

Overall survival: time interval between the diagnosis of cervical carcinoma and death from any cause.

Disease-free survival: period of time after the end of primary treatment for cervical cancer, during which the patient survived without any signs or symptoms of this cancer.

Complete remission: all signs and symptoms of cancer have disappeared, although cancer still may be in the patient’s body.

Partial response: decrease in the tumor size, or in the extent of cancer in response to treatment.

Stable disease: when the disease is neither increasing nor decreasing in extent or severity.

Disease progression: increase in the sum of maximum tumor diameters of at least 20%, together with the development of any new lesions or an unequivocal increase in non-measurable malignant disease.

Overall response rate: percentage of patients whose cancer shrinks or disappears after treatment.

Disease control rate: percentage of patients with advanced cancer whose therapy has led to a complete or partial response or stable disease.

Progression free survival: length of time during and after the treatment of the cancer in which the patient lives with the disease without getting worse.

Adverse events: any abnormal clinical finding temporarily associated with the use of therapy for cervical carcinoma, according to CTCAE V5 [21].

Data collection and analysis

The retrieved studies will be exported to Rayyan software (Mourad Ouzzani, University of Oxford, UK). All titles and abstracts will be screened independently by two reviewers (DMD and ATBS). Disagreements will be resolved through discussion with a third reviewer (ACAS). To ensure uniform interpretation of the eligibility criteria, a pilot screening of 200 articles from the Pubmed search will be performed before the initiation of the title and abstract screening phase.

After selecting the studies by titles and abstracts, three independent reviewers (DMD, ATBS, ACAS) will read the articles in their entirety and select the studies according to the eligibility criteria, with disagreements being resolved by consensus with the participation of a fourth reviewer (RNC). After this selection, they will independently extract data from eligible studies using a predefined extraction scheme.

The data from each study will be extracted for the following categories: author, year of publication, country, categories of participants, sample size, study design, histological type, cancer staging, type of intervention, follow-up period, mortality rate, overall survival, disease-free survival, adverse events, and main findings. If additional details are required, the authors of the study will be contacted to provide all necessary unreported data. The data not provided from those studies in which three attempts were made to contact the authors by email, will be excluded from the meta-analysis, with information from these studies remaining only in the narrative synthesis of the systematic review.

The PRISMA 2020 flow diagram (Fig 1) will be used to present the study identification, screening and inclusion.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow diagram for systematic review and meta-analysis.

https://doi.org/10.1371/journal.pone.0312004.g001

Bias risk assessment

The bias risk of the selected articles will be assessed independently by two reviewers using the updated Cochrane risk-of-bias tool for randomized trials (RoB2) [22]. The specific criteria that will be evaluated include random sequence generation to assess selection bias, allocation concealment to further evaluate selection bias, blinding of participants and personnel to address performance bias, analysis of incomplete outcome data to determine attrition bias, examination of selective reporting to identify reporting bias, and the identification of any other potential sources of bias. In case of insufficient details reported in the study, authors will be contacted. Discrepancies in regard to the assessment of risk of bias will be solved through discussion with a third author (RNC).

Randomized clinical trials retrieved from the grey literature will have their risk of bias rigorously assessed by this tool. In cases of studies with incomplete data that compromise the rigorous assessment of the risk of bias, three attempts will be made to contact the authors via email, and only the RCTs whose data were recovered will be included in the meta-analysis.

Whenever possible, funnel plots will be used to assess the potential existence of publication bias, complemented by Egger’s weighted correlation and Begg’s regression intercept at a 5% significance level [23].

Certainty of evidence

The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework will be used to grade the quality of evidence for each intervention as high, moderate, low or very low [24]. Overall, upgrading or downgrading the quality of evidence depends on the risk of bias, imprecision, inconsistency, and indirectness of the findings. Any disagreement regarding the assessment of certainty of evidence will be resolved through consensus among the authors, which may guarantee that the conclusions provide a balanced view of the available evidence.

Summary of Findings (SoF) table will be created and will present the following information for primary and secondary outcomes: absolute risks for the treatment and control, estimates of Odds Ratio (OR), Hazard Ratio (HR), Relative Risk (RR), or Incidence Rate Ratio (IRR), directness, heterogeneity, precision, and risk of publication bias of the results.

All evidence graded as low or very low certainty will be considered with careful caution and will be specifically highlighted in the conclusion of the future paper. This will guide the clinical decision-making regarding the efficacy and safety of pembrolizumab, emphasizing that lower quality evidence should be interpreted more carefully in clinical recommendations.

Statistical analysis

A quantitative synthesis will be conducted using the ‘meta’ package of R software, version 4.3.1. For dichotomous outcomes, the Odds Ratio (OR), Hazard Ratio (HR), Relative Risk (RR), or Incidence Rate Ratio (IRR), along with their 95% Confidence Intervals (CI 95%), will be extracted or calculated for each study. The Mean Difference (MD) or Standardized Mean Difference (SMD) will be computed for continuous data.

In cases of high heterogeneity (I²≥50%), the random-effects model will be employed to combine the studies and to calculate the OR and CI 95% using the DerSimonian-Laird algorithm. We will use random-effects models to analyze survival data. In cases of missing data, we will contact the authors of the studies to obtain the missing information for more accurate inference. Additionally, we will conduct a sensitivity analysis to assess the impact of the missing data on our results.

In cases where the I² statistic is substantially high, we will discuss its potential implications for the overall interpretation of the results. Additionally, only clinical trials whose data are recovered after requesting from the primary study authors will be considered for quantitative synthesis, due to the fact the literature does not recommend statistically estimating unavailable data as this can compromise the quality of the meta-analysis results.

In the case of a few clinical trials with high heterogeneity and very high risk of bias, a meta-analysis will not be conducted, and the results of these studies will be presented in a qualitative synthesis. Synthesis Without Meta-analysis (SWiM) guidelines will be used when preparing and conducting the narrative synthesis whenever data are insufficient to conduct meta-analysis and to estimate an effect [25].

Subgroup analysis

Subgroup analysis will be conducted by separating the doses of pembrolizumab and combining groups that received the same doses of just this drug, or those that received combination therapy with bevacizumab, with the aim of verifying if there are differences in primary and secondary outcomes. Additionally, analysis will be performed according to the follow-up time of the clinical trials and the way adverse effects were assessed.

Analysis by age group of the women and by the stage of the cervical cancer, as well as by classifying the cancer as recurrent or metastatic and by its histological type (squamous, non-squamous) will be assessed by comparing outcomes after treatment with pembrolizumab alone or in combination according to the stage/type of cancer.

Finally, the groups will also be divided according to the expression levels of immune checkpoint molecules, such as PD-1 and PD-L1, with the purpose of assessing whether certain molecular signatures are associated with the efficacy of pembrolizumab treatment.

Sensitivity analysis

The robustness of the meta-analysis findings will be tested through sensitivity analysis, employing a one-by-one elimination method to pinpoint the sources of heterogeneity. Sensitivity analysis will also be conducted for testing the impact of the inclusion of low-quality studies.

Additionally, sensitivity will be tested by combining clinical trials with the same follow-up time, the same method of adverse effects assessment, and standardized outcome measures, increasing the ability to identify potential causes of heterogeneity and to deliver more reliable quantitative results.

Missing data

In instances where data of interest are missing or unclear, the research team will proceed to contact the corresponding author via email to obtain the necessary information. If these efforts are unsuccessful after 3 attempts, the data will be excluded from the quantitative analysis. This limitation will be acknowledged and discussed in the discussion section.