General Information
Conflict of Interest
- Varvara-COI-1.pdf
- Burgwinkel-COI.pdf
- Olivier-COI.pdf
- Held-COI.pdf
- Tai-COI.pdf
- Naudet-COI.pdf
- Locher-COI.pdf
Request Clinical Trials
Associated Trial(s):- NCT03390504 - A Phase 3 Study of Erdafitinib Compared With Vinflunine or Docetaxel or Pembrolizumab in Subjects With Advanced Urothelial Cancer and Selected FGFR Gene Aberrations
- NCT03180736 - A Phase 3 Study Comparing Pomalidomide and Dexamethasone With or Without Daratumumab in Subjects With Relapsed or Refractory Multiple Myeloma Who Have Received at Least One Prior Line of Therapy With Both Lenalidomide and a Proteasome Inhibitor
Request Clinical Trials
Data Request Status
Status: OngoingResearch Proposal
Project Title: Data-sharing, reanalysis and sensitivity analysis of OS and PFS of pivotal randomized controlled trials in oncology
Scientific Abstract:
Background: The sharing of Individual Participant Data (IPD) remains sub-optimal but allows the reanalysis of clinical trials to evaluate the reproducibility of results and assess their robustness. Oncology trials' results may also be impacted by the censoring rules used in the statistical plan, which may bias the outcomes in favour of one of the study groups.
Objectives: To evaluate the extent of data sharing of Randomised Controlled Trials (RCTs) that supported European Medicines Agency (EMA) approvals of cancer medicines and the reproducibility of their primary analysis, together with the robustness of their findings to censoring rule change.
Study design: This is a cross-sectional study that aims to reanalyse the results of 60 randomly selected RCTs among the 'main studies' that led to an EMA approval between January 01, 2020, and December 31, 2024.
Participants: All the participants included in the 60 studies that have been selected.
Outcomes: Overall survival (OS) and progression/disease-free survival (PFS).
Statistical Analysis: For each study, we will ask for documents on study conduct and IPD, which will be used to reanalyse OS and PFS and compare the original analysis to ours to determine whether the study can be reproduced. The proportion of RCTs that will grant us access to the data and of studies reproducible for both endpoints will be reported.
We will also assess the risk of informative censoring in RCTs by examining early censoring patterns in treatment groups and by performing a sensitivity analysis.
Brief Project Background and Statement of Project Significance:
Data reuse is extremely useful to (i) avoid duplicate studies, (ii) perform meta-analysis, (iii) test alternative research hypotheses and (iv) explore different analytical methods using the same dataset.
While data sharing is now widely recognized as extremely important for advancing research and providing new insights, its implementation is still not optimal.
This project aims to analyze the level of data sharing and reproducibility of clinical trials, with a focus on main studies that led to EMA approval in the oncological field. Main studies are the most fundamental part in the process for a drug to be granted market authorization, as they are the studies on which the decision of the Committee for Medicinal Products for Human Use is based.
The relevance that these studies have for the authorization process indicates that it is important that their findings are robust and verifiable, and that their data should be made available also to other researchers.
This project's focus is on understanding how reproducible the analyses that have led to EMA-approved oncological medicines are, and we will do so using the same material and methods as described in the studies that we will reproduce.
A similar work has been done in an earlier paper, where the authors hypothesized that at least half of the studies would have been reproducible. This has not shown itself to be true, as they obtained the necessary material only for 10 studies out of 62.
There are many reasons why a trial could be labelled as not eligible for data sharing, and one of the most common is because follow-up data collection is still ongoing. This is extremely relevant in the context of anticancer treatments, where long-term follow-up of clinical trials is essential.
A study from 2021 focused on the level of data sharing by pharmaceutical companies for anticancer drugs developed in the previous 10 years and showed that, while 45% of the trials were eligible for data sharing, the most common reason for lack of eligibility was the presence of follow-up studies. However, the presence of long-term follow-up should not be seen as an obstacle to sharing the already existing Individual Patient Data (IPD) because "If the trial data are sufficiently mature to support widespread use of the drug, they must also be sufficiently mature to be shareable with the public for assessment".
The low eligibility rate for data sharing, especially in oncological trials, is also worrisome because not all the studies labelled as eligible for IPD request always share sufficient information. A study published in 2023 showed that out of the 203 clinical trials on the FDA-approved anticancer medicines for the treatment of solid tumors from the past decade, 91 were considered eligible for IPD request, but only 70 IPD packages with great variability of completeness of key data and supporting documents were obtained. This is 34% of the original number of trials.
We used the results from the above-mentioned studies to estimate the proportion of reproducible studies, which we used to calculate the sample size needed for this project to evaluate the level of data sharing and reproducibility in oncological trials now.
Specific Aims of the Project:
This project aims to evaluate the level of data sharing and reproducibility in pivotal oncological trials that supported Marketing Authorisation Applications (MAA) to the EMA and led to positive decisions.
The objective is to reanalyze the studies in our sample according to the original protocol in use at he time of the MAA and to evaluate whether the results are reproducible or if the discrepancies between the original results and the ones obtained in our study affect the conclusions.
We also want to evaluate the presence of differential censoring in these randomly selected studies and evaluate the robustness of the study results to the switch of the censoring rule.
Study Design: Individual trial analysis
What is the purpose of the analysis being proposed? Please select all that apply.: Confirm or validate previously conducted research on treatment effectiveness Other
Software Used: Python, R
Data Source and Inclusion/Exclusion Criteria to be used to define the patient sample for your study:
We will include every patient that has been included in the original study. No exclusion criteria.
Our sample is made up of 60 studies, of which 2 are on YODA, 33 on Vivli, and the other are on different platforms or need to be requested directly from the sponsor. Therefore, we will request data from other sources as well but the data from these different studies will not be pooled or analysed together.
The complete lists of the studies in our sample is: NCT00836654, NCT01546038, NCT01744366, NCT01777152, NCT02003222, NCT02242942, NCT02319837, NCT02409342, NCT02437318, NCT02504372, NCT02569242, NCT02614794, NCT02677896, NCT02684292, NCT02743494, NCT02811861, NCT02928224, NCT02998528, NCT03043872, NCT03053440, NCT03070392, NCT03085095, NCT03106779, NCT03110562, NCT03151811, NCT03173248, NCT03180736, NCT03191786, NCT03215706, NCT03257267, NCT03274492, NCT03336333, NCT03353753, NCT03390504, NCT03409614, NCT03425643, NCT03434379, NCT03474107, NCT03493854, NCT03504397, NCT03517449, NCT03521154, NCT03553836, NCT03575351, NCT03594747, NCT03653507, NCT03663205, NCT03731260, NCT03734016, NCT03778931, NCT03783442, NCT03789604, NCT03829969, NCT03914612, NCT04003636, NCT04008030, NCT04223856, NCT04269200, NCT04322539, ,NCT04988295.
Primary and Secondary Outcome Measure(s) and how they will be categorized/defined for your study:
The primary objective of this meta-study focuses on the reproducibility and robustness to censoring rule change of results for the two main efficacy endpoints in oncology, i.e.,
1. Overall survival [OS]
2. Progression-free survival [PFS] -- or disease-free survival [DFS] -- depending on whether the setting is metastatic or adjuvant
Main Predictor/Independent Variable and how it will be categorized/defined for your study: Disease progression and/or death.
Other Variables of Interest that will be used in your analysis and how they will be categorized/defined for your study: We will focus on the variables analysed in the original protocols.
Statistical Analysis Plan:
For each study OS and PFS will be reanalyzed following the protocol and the SAP in force at the time of analysis presented in the MAA and for each endpoint the following characteristics will be detailed: (i) any deviations from the initial protocol that could influence the effect size, (ii) estimand (iii) effect size [HR with confidence interval], (iv) p-value, and (v) conclusion (positive or negative).
Two independent researchers with expertise in the field will discuss whether quantitative changes affect the conclusions. If they agree that the conclusions remain consistent, the study will be considered reproducible. If they disagree, the underlying factors contributing to the discrepancy will be examined, and the analysis may be repeated after addressing these differences. If they find the results consistent, the study will be considered reproduced with verification. Otherwise, the results will be considered as not reproduced.
Once all the trials have been re-analysed, the percentage of reproducible studies will be calculated.
Significant differential censoring will also be assessed using the reverse Kaplan-Meier analysis.
Sensitivity analyses will also be performed for each study, using different censoring methods to assess the impact of patient attrition bias on OS and PFS.
Narrative Summary:
Our study aims to evaluate the level of data sharing and reproducibility of main studies that supported EMA authorizations of oncological medicines and the robustness of their conclusion due to different censoring rules and attrition bias.
We sampled 60 RCTs among the pivotal trials that supported a positive opinion from the CHMP between 01/01/2020 and 31/12/2024.For each of them we will use the IPD,protocols and information on how the study was conducted to reanalyze OS PFS to see if the study is reproducible.
Afterwards, we will analyze the impact of censoring for OS and PFS, performing a reverse Kaplan-Meier analysis to assess the presence of differential censoring and risk of attrition bias. Finally, we will re-analyze the IPD using different ways to handle censoring and compare the results to evaluate the robustness of the study.
Project Timeline:
-Protocol drafting: Q1 2025
-OSF registration: April 2025 (URL)
-Data access request submission: June -- July 2025
-Systematic review on censoring rules based on available protocols: Q3 2025
-Receipt of first individual participant data: Q4 2025
-Reanalyses: from Q4 2025 to Q3 2026
-Analysis completion date: dec 2026
-Results synthesis: Q1 2027
-Manuscript drafting: Q1 2027
-Results reported back to the YODA Project: end of Q1 2027
-First submitted for publication: end of Q1 2027
Dissemination Plan:
The main anticipated product of this project is a peer-reviewed manuscript presenting the results of a reanalysis of RCTs using alternative censoring rules. This manuscript will focus on the reproducibility and robustness of RCTs results and their implications for clinical decision-making. The target audience includes researchers, trialists, regulators, and healthcare professionals interested in oncology, clinical trial interpretation and reproducibility research.
We anticipate submitting the manuscript to journal in the fields of general medicine, oncology and/or evidence-based medicine. Potential suitable journals include BMJ, BMJ EBM, JCE or BMC Medicine.
Additional dissemination products may include conference presentations (e.g., at the Share-CTD network or at the World Conference on Research Integrity).
Bibliography:
[1] C. Locher, G. Le Goff, A. Le Louarn, U. Mansmann, and F. Naudet, "Making data sharing the norm in medical research," BMJ, vol. 382, p. 1434, Jul. 2023, doi: 10.1136/bmj.p1434.
[2] D. B. Taichman et al., "Data Sharing Statements for Clinical Trials – A Requirement of the International Committee of Medical Journal Editors," N Engl J Med, vol. 376, no. 23, pp. 2277--2279, Jun. 2017, doi: 10.1056/NEJMe1705439.
[3] E. Loder, H. Macdonald, T. Bloom, and K. Abbasi, "Mandatory data and code sharing for research published by The BMJ," BMJ, vol. 384, p. q324, Mar. 2024, doi: 10.1136/bmj.q324.
[4] M. Siebert et al., "Data-sharing and re-analysis for main studies assessed by the European Medicines Agency-a cross-sectional study on European Public Assessment Reports," BMC Med, vol. 20, no. 1, p. 177, May 2022, doi: 10.1186/s12916-022-02377-2.
[5] N. D. Modi et al., "Audit of Data Sharing by Pharmaceutical Companies for Anticancer Medicines Approved by the US Food and Drug Administration," JAMA Oncol, vol. 8, no. 9, pp. 1310--1316, Sep. 2022, doi: 10.1001/jamaoncol.2022.2867.
[6] N. D. Modi et al., "Clinical Study Report and Individual Participant Data Transparency for US Food and Drug Administration-Approved Anticancer Drugs: A Call for Systematic Data Availability," J Clin Oncol, vol. 42, no. 32, pp. 3773--3777, Nov. 2024, doi: 10.1200/JCO.24.00539.
[7] A. M. Hopkins et al., "Heterogeneity and Utility of Pharmaceutical Company Sharing of Individual-Participant Data Packages," JAMA Oncol, vol. 9, no. 12, pp. 1621--1626, Dec. 2023, doi: 10.1001/jamaoncol.2023.3996.
[8] S. Gilboa, Y. Pras, A. Mataraso, D. Bomze, G. Markel, and T. Meirson, "Informative censoring of surrogate end-point data in phase 3 oncology trials," Eur J Cancer, vol. 153, pp. 190--202, Aug. 2021, doi: 10.1016/j.ejca.2021.04.044.
[9] "ICH Official web site : ICH." Accessed: Jan. 06, 2025. [Online]. Available: https://www.ich.org/
[10] "Attrition bias | Catalog of Bias." Accessed: Mar. 06, 2025. [Online]. Available: https://catalogofbias.org/biases/attrition-bias/
[11] V. Lesan, T. Olivier, and V. Prasad, "Progression-free survival estimates are shaped by specific censoring rules: Implications for PFS as an endpoint in cancer randomized trials," Eur J Cancer, vol. 202, p. 114022, May 2024, doi: 10.1016/j.ejca.2024.114022.
[12] "Download medicine data | European Medicines Agency (EMA)." Accessed: Jan. 06, 2025. [Online]. Available: https://www.ema.europa.eu/en/medicines/download-medicine-data
[13] "ATCDDD – ATC/DDD Index." Accessed: Jan. 06, 2025. [Online]. Available: https://atcddd.fhi.no/atc_ddd_index/
Supplementary Material: Protocol_anonymized.docx