Use of an Independent Data Review Committee to Promote Best Practices for External Control Arms: A Case Study in Relapsed/Refractory Multiple Myeloma
Author(s)
Uwe Siebert, MD, MPH, MSc, ScD1, M. Alan Brookhart, PhD2, Xavier Leleu, MD, PhD3, Rakesh Popat, BSc, MBBS, MRCP, FRCPath, PhD4, Soko Setoguchi, MD, DrPH5, Nicolle Bonar, PhD6, Michael West, MSc6, Di Wang, MSc6, Mostafa Shokoohi, PhD6, Paul Spin, PhD6, Christian Hampp, PhD7, James Harnett, PharmD, MS7, Jeannette Green, MSc7, Olivier Humblet, Sc.D.7, Alexander Breskin, PhD, MPH7, Qiufei Ma, PhD7;
1UMIT TIROL-University for Health Sciences and Technology, Hall in Tirol, Austria, 2Duke University, Durham, NC, USA, 3Hôpital La Mileterie, Poitiers, France, 4University College Hospital, London, United Kingdom, 5Rutgers University, New Brunswick, NJ, USA, 6Eversana, Burlington, ON, Canada, 7Regeneron Pharmaceuticals, Inc., Tarrytown, NY, USA
1UMIT TIROL-University for Health Sciences and Technology, Hall in Tirol, Austria, 2Duke University, Durham, NC, USA, 3Hôpital La Mileterie, Poitiers, France, 4University College Hospital, London, United Kingdom, 5Rutgers University, New Brunswick, NJ, USA, 6Eversana, Burlington, ON, Canada, 7Regeneron Pharmaceuticals, Inc., Tarrytown, NY, USA
OBJECTIVES: External control arm (ECA) studies using patient-level real-world data (RWD) often act as comparators for single-arm clinical trials in the absence of a randomized study. The FDA is open to using RWD to support assessment of treatment benefit, provided data are both relevant and reliable. We describe experience in leveraging an Independent Data Review Committee (IDRC) to assess data quality, relevance, and comparability of cohorts for a RWD-derived ECA study in multiple myeloma (MM).
METHODS: R5458-ONC-21101 (NCT05673967) is a global, non-interventional study examining RWD from patients with triple-class exposed or refractory MM initiating standard-of-care (SOC) treatment to contextualize results from the linvoseltamab (anti-BCMA×CD3 antibody) 200 mg cohort of the LINKER-MM1 phase 1/2 clinical trial (NCT03761108). RWD for SOC cohort were derived from chart reviews at participating International Myeloma Working Group sites and two US-based oncology electronic health record databases (COTA; Guardian Research Network). Per study protocol, conduct of comparative analyses was contingent on IDRC approval. The IDRC included hematologists, epidemiologists, and health decision scientists who convened over four meetings and followed a predefined charter.
RESULTS: The IDRC assessed data quality (completeness) and data relevancy (generalizability of study populations, endpoint definitions, and SOC comparators), while blinded to endpoint data. Comparability of LINKER-MM1 and RW SOC cohorts was evaluated using baseline characteristics and cohort balance after applying inverse-probability-of-treatment weighting, with the IDRC endorsing formal statistical comparisons of study effectiveness endpoints. The IDRC recommended sensitivity analyses (e.g., quantitative bias analyses for unmeasured confounding and impact of missing data for prognostic factors) and causal tools (e.g., directed acyclic graphs) for robust reporting.
CONCLUSIONS: An IDRC provides impartial evaluation of data suitability for comparing clinical trial and RW SOC cohorts, minimizing biases and strengthening regulatory/health technology submissions. Selection of unbiased experts, and a predefined assessment process and charter, are essential to the IDRC’s successful implementation in RW studies.
METHODS: R5458-ONC-21101 (NCT05673967) is a global, non-interventional study examining RWD from patients with triple-class exposed or refractory MM initiating standard-of-care (SOC) treatment to contextualize results from the linvoseltamab (anti-BCMA×CD3 antibody) 200 mg cohort of the LINKER-MM1 phase 1/2 clinical trial (NCT03761108). RWD for SOC cohort were derived from chart reviews at participating International Myeloma Working Group sites and two US-based oncology electronic health record databases (COTA; Guardian Research Network). Per study protocol, conduct of comparative analyses was contingent on IDRC approval. The IDRC included hematologists, epidemiologists, and health decision scientists who convened over four meetings and followed a predefined charter.
RESULTS: The IDRC assessed data quality (completeness) and data relevancy (generalizability of study populations, endpoint definitions, and SOC comparators), while blinded to endpoint data. Comparability of LINKER-MM1 and RW SOC cohorts was evaluated using baseline characteristics and cohort balance after applying inverse-probability-of-treatment weighting, with the IDRC endorsing formal statistical comparisons of study effectiveness endpoints. The IDRC recommended sensitivity analyses (e.g., quantitative bias analyses for unmeasured confounding and impact of missing data for prognostic factors) and causal tools (e.g., directed acyclic graphs) for robust reporting.
CONCLUSIONS: An IDRC provides impartial evaluation of data suitability for comparing clinical trial and RW SOC cohorts, minimizing biases and strengthening regulatory/health technology submissions. Selection of unbiased experts, and a predefined assessment process and charter, are essential to the IDRC’s successful implementation in RW studies.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
SA47
Topic
Study Approaches
Topic Subcategory
Surveys & Expert Panels
Disease
SDC: Oncology