The International Myeloma Foundation Proudly Announces we have completed and finalised the EMA-CHMP Qualification Advice to i2TEAMM Novel Biomarker Procedure Application on the Use of MRDnegCR as an Intermediate Early Endpoint for Conditional Market Approval in Myeloma Clinical Trials
Another Remarkable Milestone for the IMF and i2TEAMM, Following the FDA-ODAC Unanimous 12-0 Vote in April 2024
STUDIO CITY, CA, July 2, 2025 — The International Myeloma Foundation (IMF) and the collaborative stakeholder group i2TEAMM (International Independent Team for Endpoint Approval of Myeloma MRD), which includes top global myeloma experts, proudly announce the completion of i2TEAMM Europe’s application on the use of minimal residual disease (MRD) as an early endpoint in myeloma clinical trials. Building on the momentum gained from the 12-0 unanimous vote given by the U.S. Food and Drug Administration’s Oncologic Drugs Advisory Committee (ODAC) in April 2024, the i2TEAMM submitted an application (Doc Ref: EMADOC-360526170-2374344) to the EMA-CHMP letter on November 15, 2024, completing the application on the use of MRD “as an early endpoint in clinical trials conducted in patients with multiple myeloma in order to support regulatory decisions.”
Below is the full scientific advice letter On 5/22/2025, the CHMP met and adopted the advice to be given to the applicant for full context.
IMF Chairperson of the Board S. Vincent Rajkumar, MD, said: "I am very pleased with the CHMP letter on MRD. We are an international organization, and we are committed to accelerating the availability of effective new treatments to myeloma patients worldwide. I applaud the efforts of the i2TEAMM investigators and our entire team at the International Myeloma Foundation."
"The completion of the IMF/ i2TEAMM application to the EMA-CHMP represents one of the most significant milestones in our collective mission to accelerate innovation and cure in multiple myeloma. The regulatory pathway of MRD as a valid early endpoint brings us one step closer to delivering more timely, effective therapies to patients. This scientific deliverable reflects the strength of global collaboration and data-driven progress in the field," said Nikhil Munshi, MD (Dana-Farber Cancer Institute — Boston, MA), a member of the i2TEAMM Executive Committee, the IMF Scientific Advisory Board, and the IMF Board of Directors.
Jesús San Miguel, MD, PhD (University of Navarra — Pamplona, Spain), who is also a member of the i2TEAMM Executive Committee and the IMF Scientific Advisory Board, praised the EMA-CHMP's decision and said: “This was the result of a long-standing effort of the International Myeloma Foundation, Academic Centers, and Industry that worked together under the umbrella of the i2TEAMM project. The completion of the application on MRD as an endpoint will represent a sensitive early readout for conditional drug approval, allowing patients timely access to newer treatment options although trials should be adequately planned to demonstrate a benefit in PFS or OS."
“The International Myeloma Foundation is elated at the groundbreaking outcomes of the i2TEAMM’s efforts, both in the FDA-ODAC and the EMA-CHMP. As these remarkable milestones bring us closer to a cure, we look forward to a future where the IMF’s vision can finally be achieved: A world where every myeloma patient can live life to the fullest, unburdened by the disease,” said IMF Interim CEO and Senior VP of Strategic Planning Diane Moran, RN, MA, EdM.
FULL SCIENTIFIC ADVICE LETTER
Initial Qualification Procedure
Minimal Residual Disease in Multiple Myeloma
On 15/11/2024, the Applicant i2TEAMM Europe requested Qualification Advice, pursuant to Article 57(1)(n) of Regulation (EC) 726/2004 of the European Parliament and of the Council, on the use of Minimal Residual Disease (MRD) as an early endpoint in clinical trials conducted in patients with Multiple Myeloma (MM) in order to support regulatory decisions. Serena Marchetti, Hilke Zander and Elina Asikanius were appointed as Coordinators. The EMA Scientific Officer for the procedure was Myriam Chapelin. Lukas Aguirre Dávila, Michal Zwiewka, Macarena Gajardo Álvarez, Joerg Zinserling, Torsten Holm Nielsen, Odoardo Olimpieri, Johanna Lähteenvuo, Karri Penttila, Olli Tenhunen, Pierre Demolis and Filip Josephson were appointed as Qualification Team members. The procedure started on 25/11/2024. On 08/05/2025, the SAWP agreed on the advice to be given to the Applicant.
On 22/05/2025, the CHMP adopted the advice to be given to the Applicant. The response given by the CHMP is based on the questions and supporting documentation submitted by the Applicant, considered in the light of the current state of the art in the relevant scientific fields
The response given by the CHMP is based on the questions and supporting documentation submitted by the Applicant, considered in the light of the current state of the art in the relevant scientific fields.
Based on the Coordinators’ reports the CHMP gave the following answers: Questions on Clinical development
Question 1
Based on the coordinators' reports, the CHMP provided the following answers regarding clinical development:Does the CHMP agree that the rate of MRDnegCR, at 10-5 or higher threshold and measured at 9±3 or 12±3 months, translates into clinically meaningful benefit in patients with NDTE, NDTinE, and RR MM?
CHMP answer
CHMP agrees that MRD negativity at 10-5 or higher threshold has a prognostic value at patient level. The data presented appear to support this positive effect in different MM settings, i.e., NDTE, NDTinE and RRMM. However, surrogacy at trial level has not been sufficiently established at the present time.
Methodological uncertainties have been identified in the extensive data analyses presented (further discussed under ‘Other comments not directly related to the questions’ below) hampering interpretation at this time.
Moreover, it is not straightforward to generalise the claimed benefit of MRDnegCR to all new treatment modalities in all settings for future studies, considering relevant differences in potential treatmentrelated toxicities as well as specificities associated with specific situations (e.g., patients with extramedullary disease) and different settings. For comments related to the proposed timing of MRD assessment (i.e., at 9 ± 3 or 12 ± 3 months) reference is made to the answer to question 2.
In the context of a qualification of MRDnegCR as a marker of clinical benefit to support regulatory decision making, there are several additional aspects that should be taken into account. Several technologies are available to assess MRD (i.e., MFC multiparametric flow cytometry, NGS nextgeneration sequencing, PCR polymerase chain reaction) and no consensus has been reached on the optimal methodology to be applied. Although current evidence suggests that depth of response (i.e.-05 or -06) might be relevant irrespectively of the detection methodology applied, it is emphasised that a high level of standardisation in MRD testing has not been achieved yet and guidance is needed to ensure that MRD assessments are consistent amongst laboratories.
Question 2
Does the CHMP agree that a new therapy or combination of therapies that increases MRDnegCR rate by approximately 10% to 20%, at 10-5 or higher threshold and measured at 9±3 or 12±3 months, provides a major therapeutic advantage over approved therapies for patients with MM?
CHMP answer
CHMP emphasizes that major therapeutic advantage (MTA) has a specific meaning in the current European regulatory framework. It is evaluated on a case-by-case basis during assessment for a conditional marketing authorisation in the context of all available therapies in the sought indication. Therefore, the discussion should not be limited to the comparator arm employed in the clinical trial intended to support registration.
The Applicant proposes a minimum 10-20% difference in MRDnegCR rate to declare MTA in comparison with available treatment options, based on the results of a meta-analysis (Paiva et al 2023). This percentage is derived from combined data obtained from the NDTE, NDTI and RRMM settings.
As initial comment, the Applicant’s claim that a difference of 10-20% in MRDnegCR rates will translate to a certain hazard ratio in PFS is not agreed. CHMP rather emphasizes that, in general, the magnitude of an increase in MRDnegCR considered clinically relevant should be determined on a case-by-case basis and may differ per treatment setting. Furthermore, treatment-related toxicity is always part of the B/R assessment and will as well contribute to MTA conclusion.
Time points for MRD assessments cannot be easily generalised. The proposed 12 ±3 months seem practical and could be considered acceptable in most cases, provided a tighter assessment window than ±3 months is implemented. In effect, the proposed ±3 months interval may lead to imbalances between treatment arms able to affect interpretation of the data. Specificities associated with the different settings should be taken into account, too, as for instance a 12-month time point would not be considered adequate for assessment of treatments employed in the induction phase before transplantation.
Question 3
Does the CHMP agree with the context of use for MRDnegCR, classified at 10-5 or higher threshold and measured at 9±3 or 12±3 months, as an early clinical endpoint in clinical trials to support regulatory approval of new indications for MM while PFS and OS results are maturing?
CHMP answer
CHMP agrees that, depending on the setting, a role for MRDnegCR as an endpoint to support (conditional) approval of a compound while the obligation to demonstrate long-term benefit remains, can be envisaged. This implies that the trials should be adequately planned to demonstrate a benefit in PFS or OS (EMA/CHMP/205/95 Rev.6).
The predictive value of an increase in MRDnegCR regarding long-term clinical benefit in MM has not been fully established at this time. Therefore, based on the presented data, generally MRDnegCR is not considered able to replace PFS or OS for regulatory approval for obtaining a full marketing authorisation at the present time. CHMP emphasises that, in particular when compounds with novel mechanisms of action are explored, a comprehensive evaluation of toxicity with sufficient follow up is needed for an adequate benefit/risk assessment. MRDnegCR is, per definition, a parameter able to capture antitumor activity but unable to reflect or capture the treatment related toxicity. In the most extreme scenario, MRD negativity could be reached with a treatment poorly tolerated leading to extreme toxicity with detriment in OS. Notably, tolerability can largely differ between subgroups (e.g. elderly). Age can also impact treatment targets: indeed, long-term disease control, excellent QoL and similar survival has been reported in elderly patients with MM with less aggressive approaches (Wildes 2014). Therefore, acceptability of MRD as primary evidence for efficacy for early approvals might depend on the context and targeted clinical setting.
Related to the context of use, it includes a statement “…while PFS and OS endpoints are maturing”. It has not been discussed how such a study would concretely be designed and whether the intention is to state that PFS and OS are expected to become available from the same trial. The MRD analysis will mean an early read out and unblinding of the trial which could compromise study integrity.
Furthermore, the MRD analysis, with e.g. 10-20% difference between arms as proposed, would require fewer patients than an adequately powered analysis for a time-to-event endpoint. Hence designing a trial which can at the same time reliably assess both MRD and time to event endpoints might be challenging.
Finally, it is emphasized that the treatment regimens in MM are complex. Often, a combination of multiple treatments with different durations and induction, consolidation and maintenance therapies are administered. For these reasons, it is not possible to agree to a universal time point for MRD assessment nor that MRD is a relevant endpoint for all the clinical questions of interest. In addition, the full spectrum of treatment may not be fully captured by defining one timepoint for MRDnegCR assessment because MRD status can be transient (D’Agostino 2024, Guerrero 2024).
In conclusion, the suitability of MRDnegCR at a prespecified timepoint as an endpoint in clinical trials in order to support regulatory decision making depends on the treatment setting and study design and will be based on a case-by-case assessment.
Question 4
The totality of data presented by the i2TEAMM Europe meta-analysis, and by the independent meta-analysis conducted by University of Miami (EVIDENCE), provides evidence for the use of MRDnegCR as an early endpoint to measure therapeutic benefit and support benefit-risk assessments for regulatory approval in MM. Does the CHMP agree?
CHMP answer
CHMP does not consider the available data sufficient to support the use of MRDnegCR rate as an early, stand alone, clinical endpoint in randomised controlled trials in multiple myeloma to establish, on its own, the benefit/risk of new treatments or combinations and support a full MAA. Support from longterm endpoints, measured at a later stage, would still be needed to demonstrate the benefit. At an individual patient level, the value of MRDnegCR for disease prognosis is acknowledged and already used in clinical care. However, trial level surrogacy has not been sufficiently established at the present time.
As pointed out in the answer to question 3, MRDnegCR as an endpoint in clinical trials in specific settings, together with the totality of the available evidence, could support regulatory decision in the context of a CMA, if confirmed by subsequent efficacy and safety data from long term clinically relevant endpoints (e.g., PFS/OS).
Other comments not directly related to the questions
The designs and statistical approaches of the studies feeding into the (patient-level and) trial-level correlational investigations also need to be evaluated to some extent although these were not the scope of the pre-defined questions. The following paragraph provides some considerations in relation to this.
The efforts to collect and analyse individual patient data from a high number of studies are appreciated and the challenges acknowledged. Both patient- and trial-level associations were investigated. As the main question is whether and how an effect (experimental treatment vs control treatment) on MRDnegCR translates into long term effects on clinical endpoints, the trial-level associations represent the most important analyses. The patient-level associations support that MRDnegCR is in the chain from treatment to outcome at an individual level.
Considerations related to the study designs and statistical approaches
Endpoint definitions
In order to be considered for MRD negativity, a patient first needs to have had a CR or sCR. It is assumed that conventional IMWG criteria were applied for determination of CR status, but this is currently unclear. The assessment schedule for response was not reported in the overview of the studies, nor was the definition of response status. It is also unclear how the use of new MM therapy prior to disease progression was handled for the assessment of response in these studies. This may have some impact on both the patient and trial-level correlations.
PFS was also based on investigator-determined progressions. While this is understood from the perspective of pooling and analysing the trials, progression based on an independent central review should usually be the basis for the primary endpoint in a pivotal open-label trial. It is not clear how this might affect the results of the trial-level associations. Regarding the censoring rules for PFS, it is stated that “Patients without documented PD who are alive will be censored on the date of the last documented disease evaluation, regardless of subsequent new MM therapies or any missing tumor assessment before the last disease evaluation.” This definition is in line with the preferred CHMP approach for PFS. Usually, an evaluation of the results would include an investigation of whether patients remained in follow up after receiving new MM therapy and other investigations related to potentially informative censoring. However, given the number of the studies where this is not possible for this procedure, it is considered likely that informative censoring influences the estimation of the patient-level and trial-level associations.
In section 3.4 of the SAP version 3.1, various scenarios are presented to clarify when a “success” (MRD negativity) would be concluded. Scenario 1 refers to the event of PD or death prior to the MRD window. In the example given, the patient had not achieved a CR/sCR before PD/Death, so this is clearly a failure. It is not fully clear how patients who did achieve a (best objective response) of CR/sCR before progressing were handled – i.e., whether MRD status was considered a failure or missing.
It is not clear from the study descriptions when MRD was actually measured. An overview of the time windows for each study, along with median time would be useful.
Methodology
A full evaluation of the methodology implemented to estimate the patient-level and trial-level associations is outside the scope of this procedure. The approaches were discussed in the previous advice and were generally considered to be acceptable.
The primary definition of PFS is the time from the start date specific to the two-arm comparison to the date of first documented disease progression or death due to any cause, whichever occurred first. Burzykowski et al. (2004) note that the comparison of survival times of patients according to tumour response (in their case) is likely to be “length biased” in the sense that patients who experience long survival times are more likely to be responders than non-responders which biases the survival of responders upwards. This is also relevant for the current case, where patients not only need to be responders, but also have an MRD measurement at 9 ± 3 months (or 12 ± 3 months) and be alive and progression free. Some results from ‘landmark analyses’ were presented but these were in the form of Kaplan-Meier curves rather than the Global ORs. Also, given the landmark was the time of the determination of MRD status, and not a specific timepoint for all patients, it is not clear how this may affect the curves and comparisons.
Presentation of information and results
It is not clear which studies have been used for the different sets of results.
No information is presented on the timing of response assessments, and study-level results for MRDnegCR and PFS. The time windows and ranges for the relevant MRD assessments are also missing.
Interpretation of results
No overview of the percentages of MRDnegCR reported in the trials is presented. For the comparison of the outcome of MTDnegCR, risk differences or risk ratios are a preferred measure rather than odds ratios.
Concerns regarding use of MRDnegCR as an intermediate endpoint for PFS/OS
Response-based outcomes in trials are estimated using the ITT analysis set, or the full analysis sample, for which all patients are included in the denominator for the calculation of the percentage of responders in the experimental and control arms. The responders (in this case patients with MRDnegCR) are included in the numerator. The non-responders are a mix of patients including patients who progressed or died, those who had all measurements but did not have a CR but were MRD negative, and patients for whom the assessments were non-evaluable or missing due to participant withdrawal. This latter category could be seen as an “experimental imperfection” rather than representing a true non-responder status of the patient.
Particularly in open-label trials, an imbalance in early withdrawal from the control arm after randomization might challenge interpretation of the results in terms of MRDneg CR as due to the lack of assessments of response they are counted as non-responders in the estimates of the percentage of MRDnegCR patients. In view of this, a difference in the responder rates may not necessarily represent a true difference in the activity of the treatment.
As discussed in the answer to Question 3, another risk associated with the use of MRDnegCR as an intermediate endpoint for conditional approval, with longer-term PFS or OS results delivered as a specific obligation is that patients in the control arm might choose to discontinue their allocated treatment in order to “crossover” and receive the newly approved therapy. In extreme cases, the PFS or OS results may no longer be interpretable. This issue should always be anticipated at the design stage of the trial and measures should be put in place to minimize the potential for crossover, to the extent possible. Analyses to account for the crossover should also be pre-specified and a data collection plan that supports these analyses should be defined as part of the design of the study (e.g. the collection of time-dependent prognostic factors).
ABOUT THE INTERNATIONAL MYELOMA FOUNDATION
Founded in 1990, the International Myeloma Foundation (IMF) is the first and largest global foundation focusing specifically on multiple myeloma. The Foundation's reach extends to more than 525,000 members in 140 countries worldwide. The IMF is dedicated to improving the quality of life of myeloma patients while working toward prevention and a cure by focusing on four key areas: research, education, support, and advocacy. The IMF has conducted more than 250 educational seminars worldwide, maintains a world-renowned InfoLine, and in 2001, established the International Myeloma Working Group (IMWG), a collaborative research initiative focused on improving myeloma treatment options for patients. In 2012, the IMF launched the Black Swan Research Initiative®, a groundbreaking research project aimed at curing myeloma. The IMF can be reached at (800) 452-CURE (2873). The global website is www.myeloma.org.
Follow the IMF on:
X/Twitter: @IMFmyeloma
Instagram: @imfmyeloma
Facebook: @myeloma
LinkedIn: International Myeloma Foundation
Bluesky: @imfmyeloma.bsky.social
Media Contacts:
Peter Anton
[email protected]
Jason London
[email protected]