Development and application of new statistical methods for meta-analysis to evaluate treatment effectiveness and surrogate relationships in biomarker subgroups

Lead Investigator: Lorna Wheaton, University of Leicester
Title of Proposal Research: Development and application of new statistical methods for meta-analysis to evaluate treatment effectiveness and surrogate relationships in biomarker subgroups
Vivli Data Request: 9826
Funding Source: None
Potential Conflicts of Interest: Sandro Gsteiger is an employee of F. Hoffmann-La Roche Ltd. Therefore, he will not be involved in any data analysis activity or be accessing any individual level data requested through Vivli. He supports the project through general scientific advice, guidance on the research questions, and helps in the interpretation of results providing an industry perspective. For this purpose, the research team from the University of Leicester will exchange only analysis results and summary level data with Sandro Gsteiger. This conflict of interest will also be declared in any publications resulting from this data request.

Summary of the Proposed Research:

Lung cancer is the third most common cancer in the UK, representing 13% of all new cancer cases. 80-85% of lung cancer cases are in non-small cell lung cancer (NSCLC). Ten-year survival for lung cancer in England is 9.5% and in the UK lung cancer is the most common cause of death due to cancer.

Traditional treatment for NSCLC includes chemotherapy which aims to eradicate or reduce the tumour. However, chemotherapy treatments are not specific to cancer cells and can have a toxic effect on normal cells. Alternatively, targeted therapies target proteins that determine how cancer cells grow, divide and spread. Currently, in the UK, advanced NSCLC patients are routinely tested for a selection of genetic biomarkers (which are unique features of a patients’ DNA) to inform the best targeted treatment for the patient.

Every new treatment which is developed is investigated to assess whether it is safe and effective. Ideally, a randomized controlled trial (RCT) will be conducted where patients are randomly assigned to receive the new treatment or a comparator and are followed up over time to assess their outcomes. As treatments become more targeted, they often become more effective but are only suitable for smaller populations of patients. This results in smaller RCTs which take longer to conduct and generates greater uncertainty around estimates of effectiveness, i.e., how well the new treatment works.

When there is high uncertainty around effectiveness estimates obtained from a single RCT, it can be beneficial to combine data from multiple RCTs to estimate the overall treatment effect from all available studies. This is done using a statistical method known as meta-analysis. Combination of multiple trials in meta-analysis also allows researchers to investigate potential effect modifiers – things that influence how effective the treatment is, which are particular to individual patients (such as genetic biomarkers) that alter treatment effectiveness.

In this project, we focus on different ways to estimate treatment effects in biomarker subgroups. We consider several different meta-analysis methods to combine data from RCTs to investigate effect modification. Some methods only use summary data such as the average treatment effect or the proportion of patients with a genetic mutation in the study. Other methods use individual participant data (IPD), which provide information on the clinical outcome or the characteristics for each individual in the study; for example, the exact combination of genetic mutations for each patient.

To compare the performance of these different statistical methods, we will apply them to studies investigating treatments for NSCLC. The differences in genetic biomarkers between patients could potentially bias the treatment effect estimates. We hope that by illustrating how different statistical methods can be applied to a pre-existing dataset we will demonstrate how these methods can be successfully applied in future meta-analyses.

Requested Studies:

A Phase III, Randomised, Double-blind, Placebo-controlled, Multi-centre, International Study of MEDI4736 as Sequential Therapy in Patients With Locally Advanced, Unresectable Non-Small Cell Lung Cancer (Stage III) Who Have Not Progressed Following Definitive, Platinum-based, Concurrent Chemoradiation Therapy (PACIFIC)
Data Contributor: AstraZeneca
Study ID: NCT02125461
Sponsor ID: D4191C00001

A Phase II, Open-label, Multicenter, Randomized Study to Investigate the Efficacy and Safety of MPDL3280A (Anti−PD-L1 Antibody) Compared With Docetaxel in Patients With Non−Small Cell Lung Cancer After Platinum Failure
Data Contributor: Roche
Study ID: NCT01903993
Sponsor ID: GO28753

A Phase III, Open Label, Randomized Study of Atezolizumab (Anti-PD-L1 Antibody) Compared With a Platinum Agent (Cisplatin or Carboplatin) in Combination With Either Pemetrexed or Gemcitabine for PD-L1-Selected, Chemotherapy-Naive Patients With Stage IV Non-Squamous Or Squamous Non-Small Cell Lung Cancer
Data Contributor: Roche
Study ID: NCT02409342
Sponsor ID: GO29431

A Phase III, Open-Label, Multicenter, Randomized Study Evaluating the Efficacy and Safety of Atezolizumab (MPDL3280A, Anti-PD-L1 Antibody) in Combination With Carboplatin+Paclitaxel or Atezolizumab in Combination With Carboplatin+Nab-Paclitaxel Versus Carboplatin+Nab-Paclitaxel in Chemotherapy-Naive Patients With Stage IV Squamous Non-Small Cell Lung Cancer
Data Contributor: Roche
Study ID: NCT02367794
Sponsor ID: GO29437

A Phase III, Open-Label, Multicenter, Randomized Study to Investigate the Efficacy and Safety of Atezolizumab (Anti-PD-L1 Antibody) Compared With Docetaxel in Patients With Non-Small Cell Lung Cancer After Failure With Platinum Containing Chemotherapy
Data Contributor: Roche
Study ID: NCT02008227
Sponsor ID: GO28915

A Phase III, Open-Label, Randomized Study of Atezolizumab (MPDL3280A, Anti-PD-L1 Antibody) in Combination With Carboplatin+Paclitaxel With or Without Bevacizumab Compared With Carboplatin + Paclitaxel + Bevacizumab in Chemotherapy-Naïve Patients With Stage IV Non-Squamous Non-Small Cell Lung Cancer
Data Contributor: Roche
Study ID: NCT02366143
Sponsor ID: GO29436

A Phase III Multicenter, Randomized, Open-Label Study Evaluating the Efficacy and Safety of Atezolizumab (MPDL3280A, Anti-PD-L1 Antibody) in Combination With Carboplatin+Nab-Paclitaxel for Chemotherapy-Naive Patients With Stage IV Non-Squamous Non-Small Cell Lung Cancer
Data Contributor: Roche
Study ID: NCT02367781
Sponsor ID: GO29537

The additional data that will be analysed alongside the individual participant data (IPD) obtained from Vivli is publicly available data that can be used for secondary analysis such as in a meta-analysis. We will only be using aggregated data taken from published study reports of the trials listed below and no additional IPD will be uploaded to the Vivli platform. The additional data will be extracted from 30 research papers representing 31 RCTs which were identified from the Horita (2022) review (DOI: 10.3390/cancers15010185) for which IPD is not available on the Vivli platform. It is important that aggregate data from these 31 trials can be utilised alongside IPD from 7 trials provided via Vivli to ensure that the maximum amount of information is used for evaluation of treatment effectiveness and surrogacy. Please find the NCT numbers for each of the 31 trials for which we will be uploading aggregate data obtained from published trial reports for use alongside IPD from 7 studies obtained via the Vivli platform: NCT02395172 NCT01673867 NCT03302234 NCT01642004 NCT02041533 NCT02578680 NCT02785952 NCT01285609 NCT02477826 NCT02591615 NCT01905657 NCT03656094 NCT02039674 NCT03057106 NCT00527735 NCT02220894 NCT01633970 NCT02775435 NCT03215706 NCT02352948 (comprised of sub-studies A and B) NCT02142738 NCT02453282 NCT03088540 NCT03117049 NCT02613507 NCT03607539 NCT03134872 NCT03629925 NCT03728556 NCT03789604
Data Contributor: I WILL BRING MY OWN