In January, FDA released a draft guidance entitled “Minimal Residual Disease and Complete Response in Multiple Myeloma: Use as Endpoints to Support Accelerated Approval.” This release came roughly 20 months after the Oncologic Drugs Advisory Committee (ODAC) voted unanimously that minimal residual disease (MRD) negativity, in combination with complete response (CR), is an acceptable primary endpoint to support accelerated approval for multiple myeloma (MM) therapies.
While the ODAC vote was the headline, the draft guidance is the broader story, as its consolidation of existing regulatory expectations was not a part of the ODAC conversation but has meaningful implications for both developers and the broader use of MRD in regulatory decision-making.
The April 2024 ODAC discussion centered on the core scientific question of whether MRD negativity is reasonably likely to predict clinical benefit. Independent meta-analyses of two public-private partnership studies—the EVIDENCE study and the I2TEAMM study—conducted in collaboration with FDA presented compelling individual-level associations between MRD-negative CR and progression-free survival (PFS) across newly diagnosed and relapsed or refractory settings.
These analyses drew from thousands of randomized patients across multiple phase II and phase III trials, using assays validated to a sensitivity threshold of 1 tumor cell in 100,000 normal cells (10-5) or better. The committee’s affirmative vote was a watershed moment, the product of more than a decade of cooperative work that began with FDA co-sponsored public workshops on MRD in hematologic malignancies as early as 2012. But the vote itself was binary, answering the “whether.” The new FDA draft guidance addresses the “how”—and the how is where the regulatory complexity lies.
What MRD is—and isn’t—as an endpoint
Before turning to the specifics of the guidance, it is important to call out a distinction that is foundational to the entire document but that is easily overlooked: the difference between a validated surrogate endpoint and an intermediate clinical endpoint that is reasonably likely to predict clinical benefit. This distinction has direct consequences for how MRD can—and cannot—be used in regulatory decision-making.
A validated surrogate endpoint requires robust biological plausibility linking it to the clinical outcome, demonstrated prognostic value, and evidence from clinical trials that treatment effects on the surrogate reliably correspond to treatment effects on long-term outcomes such as PFS or overall survival (OS). This last requirement, often described as trial-level correlation, is the most demanding. It asks not just whether patients who achieve MRD negativity tend to do better at an individual level, but whether a drug that produces higher MRD negativity rates reliably produces better survival outcomes at the trial level.
The overarching message that developers must internalize is that a new endpoint does not relax old principles. The guidance makes clear that MRD does not alter the underlying standards for trial design or evidence.
The meta-analyses presented at ODAC demonstrated strong individual-level associations. Patients who achieved MRD-negative CR had significantly improved PFS across disease settings.
However, trial-level correlations were more variable. In the EVIDENCE analysis, for example, trial-level correlation coefficients between 12-month MRD-negative CR and PFS ranged from 0.83 in transplant-ineligible newly diagnosed patients to 0.35 in transplant-eligible patients and essentially zero in the relapsed or refractory setting. For OS, the correlations were even weaker. FDA’s pooled analysis, combining the data from the EVIDENCE and I2TEAMM studies yielded similar results at the trial level.
This is why FDA has classified MRD not as a validated surrogate but as an intermediate clinical endpoint reasonably likely to predict benefit, which is the standard required for accelerated, but not traditional, approval. MRD-based approvals will reflect the inherent uncertainties of the accelerated pathway, including the requirement for confirmatory trials, the possibility of withdrawal if clinical benefit is not verified, and the expectation that sponsors will have a plan in place to demonstrate longer-term outcomes. The entire architecture of the guidance stems from the recognition that MRD has not cleared the bar for full surrogacy and the regulatory framework must account for that residual uncertainty.
This framing was implicit in the ODAC discussion, but it was never stated with the clarity or operational specificity that the guidance seeks to provide. Developers who treat MRD as a shortened path to approval, rather than as an intermediate endpoint embedded in a carefully constructed body of evidence, will find themselves misaligned with the agency’s expectations.
New endpoints do not relax old principles
What makes the draft guidance significant is not any single recommendation in isolation. Instead, it is the consolidation of regulatory expectations that, until now, existed in silos across pre-investigational new drug meeting feedback, individual review decision practices, the 2020 general guidance on MRD in hematologic malignancies, and the institutional knowledge of FDA reviewers who have spent years evaluating MRD data in submissions.,
For the first time, the agency has assembled these expectations into a coherent framework tailored specifically to the accelerated approval pathway for MM. This consolidation was not part of the ODAC discussion, and it warrants attention from any organization designing or funding a MM trial.
The overarching message that developers must internalize is that a new endpoint does not relax old principles. The guidance makes clear that MRD does not alter the underlying standards for trial design or evidence. Randomization, survival assessment, contribution of effect, and thoughtful trial conduct still anchor regulatory decision-making. MRD is a more sensitive tool for measuring treatment efficacy, but the rigor expected in wielding that tool is, if anything, higher than what developers have grown accustomed to with overall response rate (ORR).
The collaborative nature of the draft guidance is also worth noting. The document reflects input from across the agency, including the Oncology Center of Excellence, the Center for Drug Evaluation and Research, the Center for Biologics Evaluation and Research, and the Center for Devices and Radiological Health. This cross-center coordination is significant since MRD-based regulatory submissions sit at the intersection of drug development and diagnostic device evaluation. The draft guidance makes the integrated nature of this regulatory challenge explicit.
Randomized trials, even for accelerated approval
The guidance clearly outlines a preference for randomized trials, even when MRD is being used to support accelerated approval. This may seem counterintuitive, as accelerated approval has historically been a pathway associated with single-arm trials evaluating response rates in settings of serious unmet need. The agency leaves open the possibility for single-arm trials to support MRD-based approvals, but expresses a clear preference for randomized designs.
A well-designed randomized trial can serve the dual purpose of evaluating MRD negativity rate at an early timepoint to support accelerated approval while simultaneously collecting the long-term PFS and OS data needed for confirmatory evidence or traditional approval. The practical benefit is that sponsors can run one trial rather than two sequential studies, reducing the total development timeline even while employing a more rigorous design.
Randomized designs also provide the comparative safety data that single-arm trials inherently lack and allow for a control arm that reflects U.S. standard of care, which is specifically called for in the draft guidance. This approach may afford conduct in an earlier treatment setting, consistent with Oncology Center of Excellence’s Project Frontrunner paradigm. For sponsors, the strategic implication is that investing in a randomized trial upfront is likely the most efficient path to both accelerated and traditional approval.
Overall survival is still required
Because MRD remains an intermediate rather than a validated surrogate endpoint, the guidance requires that OS still be evaluated whether as a secondary or safety endpoint. This aligns closely with FDA’s broader guidance on survival endpoints released in August 2025, and the consistency is deliberate. MRD does not replace OS. Instead, it sits alongside it.
In an era where patients with MM may survive a decade or more on successive lines of therapy, demonstrating an OS benefit for any single regimen is increasingly challenging. Moreover, recent oncology trials have highlighted the potential for lack of correlation between early efficacy endpoints and OS. The phase III BELLINI trial in relapsed or refractory MM is a cautionary example that illustrates this dynamic. Patients randomized to venetoclax plus bortezomib and dexamethasone achieved higher ORRs, higher MRD negative rates, and longer PFS compared to those receiving placebo. And yet, interim analysis revealed significantly higher mortality in the venetoclax arm, prompting the sponsor and FDA to halt enrollment.
The guidance does not dismiss this challenge. Rather, it acknowledges the evolving treatment landscape while maintaining that long-term outcomes cannot be abandoned as the field shifts toward earlier and more sensitive measures of efficacy. For developers, the takeaway is that accelerated approval via MRD is not an exemption from the obligation to understand the implications of an investigative product for patients over the lifecycle of their disease. Thus, planning for OS evaluation from the outset, even when it is a secondary endpoint, is required.
Full enrollment before MRD readout
Another expectation that is codified in the guidance, but absent from the ODAC proceedings, is that, in randomized trials, the study should be fully enrolled before the primary MRD analysis is conducted. Historically, there was more flexibility allowing sponsors to sometimes conduct primary analyses when enrollment was nearing completion, but this guidance is more prescriptive.
The notable tightening of expectations addresses concerns about operational bias. MRD is an endpoint that can be assessed months or even a year or more before PFS data mature, especially with newer, more potent therapies that achieve deep responses earlier. If interim MRD data were analyzed while enrollment was ongoing, knowledge of early results could influence investigator behavior, referral patterns, or patient selection, potentially compromising the integrity of the final analysis.
By requiring full enrollment prior to the MRD readout, FDA is applying a standard of rigor more commonly associated with time-to-event endpoints to the evaluation of a rate-based endpoint. This signals that the agency is setting a high evidence bar for MRD-based approvals, and developers should account for this when drafting their trial timelines and data monitoring plans.
Contribution of effect: a familiar concept applied in a new context
Perhaps the most operationally consequential section of the guidance, which again received no attention at ODAC, involves the concept of contribution of effect for combination regimens. In MM, multi-drug combinations are the standard of care. Triplet and quadruplet regimens dominate both the newly diagnosed and relapsed or refractory settings.
Sponsors seeking accelerated approval based on MRD must outline a strategy for demonstrating that each component of the regimen contributes meaningfully to the observed MRD negativity rate.
This is not a novel concept in FDA regulatory thinking. The agency has long required evidence of the contribution of each drug to a combination regimen when approval is sought based on response rate. But applying this framework to MRD is new, and the practical implications are substantial. When backbone therapies already drive response rates of 80-90%, isolating the marginal contribution of an added agent at the deeper level of MRD negativity will require careful planning, whether through factorial trial designs, leveraging data from earlier-phase studies, or building a compelling totality-of-evidence argument.
The guidance also extends this logic to multi-phase treatment paradigms, requiring sponsors to elucidate the contribution of induction, consolidation, and maintenance phases to changes in MRD negativity. In a field that increasingly relies on sequential and layered treatment strategies, this suggests that regulatory hurdles rise with regimen complexity. Sponsors may benefit from discussing their contribution-of-effect strategies with the appropriate FDA review division early in development.
Evidence needed for biomarker-selected populations
The guidance also introduces expectations around biomarker-selected populations that were not part of the ODAC conversation. If a sponsor proposes to enroll a MM population defined by a specific biomarker, the guidance states that additional information will be needed to demonstrate whether MRD is an appropriate endpoint for that population. This is a thoughtful hedge as the meta-analyses presented at ODAC evaluated MRD across broad populations in the newly diagnosed and relapsed or refractory settings.
Whether MRD negativity carries the same prognostic weight in genomically defined subgroups, particularly patients with high-risk features such as del(17p), t(4;14), or gain (1q), remains an open question. Some high-risk patients achieve MRD negativity only to relapse quickly, raising concerns about whether the biomarker means the same thing in that context.
FDA is right to flag this and to require sponsors utilizing enrichment strategies to provide supporting evidence. For developers pursuing precision medicine approaches, this provision warrants early and substantive engagement with the agency.
Assay validation and the problem of missing data
Finally, the guidance devotes considerable attention to MRD assay validation and data quality, the complexity of which is easy to underestimate. The document calls for assays that have been validated for the intended clinical use, plans to mitigate risks of calibration failures with sequencing-based methods, and strategies to minimize missing MRD data.
Perhaps the most operationally consequential section of the guidance, which again received no attention at ODAC, involves the concept of contribution of effect for combination regimens.
Each of these provisions reflects lessons learned from data FDA has reviewed in this space, including the agency’s own 2017 analysis, which found that although 40 percent of hematologic malignancy submissions included MRD data, the data were inadequate in roughly one-third of those studies. A subsequent FDA review covering submissions through 2021 confirmed that, despite a growing number of sponsors proposing MRD data for inclusion in prescribing information, there was a lag in acceptance rates attributable to analytical validation deficiencies, assay performance issues, and inadequate data collection.6
Per the guidance, MRD should be assessed in the bone marrow using appropriately validated assays such as next-generation sequencing (NGS) or next-generation flow cytometry (NGF). The recommended sensitivity threshold is at least 10-5.1
The emphasis on missing data warrants attention as the guidance specifies that missing MRD assessments should generally be counted as failures. In bone marrow-based assays, the quality of the aspirate sample directly affects the reliability of the MRD result. A hemodilute or otherwise inadequate aspirate can render an individual assessment unevaluable.
Failed calibrations in sequencing-based assays can have the same effect. When missing MRD data accumulate, the integrity of the entire trial can come into question. Consequently, operational rigor at the site level, including clear instructions for bone marrow collection, standardized criteria for aspirate quality, and contingency plans for assay failures, must be embedded into the clinical trial protocol.
What the guidance does not cover
It is worth noting what the draft guidance explicitly excludes. FDA states that available data do not currently support the use of MRD as an endpoint for accelerated approval in the maintenance setting, in smoldering MM, in monoclonal gammopathy of undetermined significance (MGUS), or in extramedullary disease. These boundaries are important as they reflect the limits of the evidence base underpinning the ODAC vote.
Expansion of MRD-based approvals into these settings will require additional data and regulatory engagement. As such, developers with programs in these spaces should engage with FDA to clarify how MRD applies to their specific context.
The ODAC vote was a landmark development, the formal acceptance of a more sensitive tool for measuring treatment efficacy in a disease where existing endpoints were reaching a ceiling. But a vote endorsing an endpoint is not the same as a regulatory roadmap for using it, particularly when the endpoint in question is an intermediate clinical endpoint, rather than a validated surrogate, and the pathway is accelerated approval.
The exclusions also extend beyond disease setting. While the guidance supports MRD as a measure of treatment response and efficacy, it does not support the use of MRD status to select patients for enrollment, to direct maintenance therapy strategies, or to guide treatment de-escalation decisions.
The field has increasingly contemplated MRD-guided treatment strategies such as intensifying therapy for patients who remain MRD-positive or shortening maintenance in those who are MRD-negative, but the guidance makes clear that the regulatory science has not yet caught up to the clinical ambition. Thus, developers should take care not to conflate the agency’s endorsement of MRD as a trial endpoint with a broader embrace of MRD-direct clinical decision-making.
Notably, the guidance stays silent regarding treatment modality. The meta-analyses underpinning the ODAC discussion drew primarily from conventional drug regimens, including proteasome inhibitors, immunomodulatory agents, and monoclonal antibodies. Chimeric antigen receptor (CAR)-T cell therapies and bispecific T-cell engagers have fundamentally different response kinetics and can achieve MRD-negative responses rapidly, but with distinct durability profiles and safety considerations.
The guidance does not address whether MRD negativity carries the same prognostic weight when achieved through these modalities, or whether the timing and interpretation of MRD assessment should differ. As cell therapies and bispecifics move earlier in the treatment paradigm, incorporation of emerging MRD data is necessary, or this gap may grow more consequential.
What MRD means to patients
A final but critical dimension of the emergence of MRD as a regulatory endpoint that the guidance does not address directly is the patient perspective. Patients with MM are increasingly aware of and engaged with their MRD status. For many, achieving MRD negativity represents a major milestone and signifies a tangible marker that treatment is working at the deepest measurable level. In a disease that remains incurable for most, that signal carries significant psychological and emotional weight.
If MRD becomes an established regulatory endpoint, it has the potential to further align drug development with outcomes that are both clinically meaningful and patient-centered. Patients already understand intuitively that a deeper remission is a better remission.
A regulatory framework that recognizes that depth while maintaining the safeguards necessary for an endpoint that has not achieved full surrogacy may accelerate access to novel therapies in a way that resonates not only with the scientific community, but with the people whose lives depend on it. Careful and rigorous implementation of this guidance paves the way to alignment between regulatory science and patient priorities, enabling a truly patient centric approach.
The roadmap to the milestone
As therapies for MM have become increasingly effective, traditional endpoints such as ORR and CR are no longer sufficient to distinguish meaningful clinical benefit. The ODAC vote was a landmark development, the formal acceptance of a more sensitive tool for measuring treatment efficacy in a disease where existing endpoints were reaching a ceiling. But a vote endorsing an endpoint is not the same as a regulatory roadmap for using it, particularly when the endpoint in question is an intermediate clinical endpoint, rather than a validated surrogate, and the pathway is accelerated approval.
FDA guidance provides that roadmap for MRD, consolidating expectations around surrogate endpoint science, trial design, statistical rigor, combination development, population selection, and assay validation that have real consequences for regulatory strategy and resource allocation.
For sponsors, this guidance rewards rigor. For the field, it is a reminder that depth of response is powerful, but only when measured, interpreted, and contextualized appropriately.
Endnotes
- U.S. Food and Drug Administration. Minimal residual disease and complete response in multiple myeloma: use as endpoints to support accelerated approval. Draft guidance for industry. January 2026. Available at https://www.fda.gov/media/190647/download. Accessed March 9, 2026.
- U.S. Food and Drug Administration. April 12, 2024: Meeting of the Oncologic Drugs Advisory Committee (ODAC). Available at https://www.fda.gov/media/177736/download. Accessed March 9, 2026.
- U.S. Food and Drug Administration. Surrogate endpoint resources for drug and biologic development. Available at https://www.fda.gov/drugs/development-resources/surrogate-endpoint-resources-drug-and-biologic-development.
- Landgren O, Devlin SM. Minimal residual disease as an early endpoint for accelerated drug approval in myeloma: a roadmap. Blood Cancer Discov. 2025 Jan 8;6(1):13-22.
- U.S. Food and Drug Administration. Hematologic malignancies: regulatory considerations for use of minimal residual disease in development of drug and biological products. Guidance for industry. January 2020 (revised March 2023). Available at https://www.fda.gov/media/134605/download. Accessed March 9, 2026.
- Baines AC, et al. Minimal residual disease data in hematologic malignancy drug applications and labeling: an FDA perspective. Clin Cancer Res. 2023;29(15):2748-2752.
- U.S. Food and Drug Administration. Approaches to assessment of overall survival in oncology clinical trials. Draft guidance for industry. August 2025. Available at https://www.fda.gov/regulatory-information/search-fda-guidance-documents/approaches-assessment-overall-survival-oncology-clinical-trials. Accessed March 9, 2026.
- Merino M, et al. Irreconcilable differences: the divorce between response rates, progression-free survival, and overall survival. J Clin Oncol. 2023;41(15):2706-2712.
- Kumar SK, et al. Venetroclax or placebo in combination with bortezomib and dexamethasone in patients with relapsed or refractory multiple myeloma (BELLINI): a randomised, double-blind, mulitcentre, phase 3 trial. Lancet Oncol. 2020;21(12):1630-1642.
- Gormley N, et al. FDA analysis of MRD data in hematologic malignancy applications. J Clin Oncol. 2017;35(15):254.
