Measurable residual disease (MRD) as a surrogate end point for clinical drug approval in acute myeloid leukemia (AML): Perspectives from the MRD Partnership and Alliance in AML Clinical Treatment Consortium.

INTRODUCTION

Our understanding of acute myeloid leukemia (AML) and the drivers of leukemogenesis has dramatically evolved with advancements in diagnostic technologies, including flow cytometry and DNA sequencing, enabling better characterization of heterogeneous patient subgroups. Consequently, the treatment landscape has shifted toward using more targeted therapies, allowing more patients to achieve clinical remissions with varying duration in both frontline and relapsed/refractory settings. 1 , 2 Despite this progress, considerable unmet medical needs remain to improve the survival of patients with AML. 3 , 4 Historically, new AML drugs were approved based on large randomized controlled trials (RCTs) with overall survival (OS) as the primary end point. More recently, event‐free survival (EFS) has gained acceptance, allowing faster approval of drugs targeting AML molecular subsets and facilitating pivotal studies for rare disease subgroups. 5 , 6 The development of more sensitive surrogate markers of treatment failure that can be measured much earlier than current morphologic methods could further accelerate AML drug development.

A surrogate end point is a biomarker that can be measured before the true clinical end point and is in the causal pathway of the disease process; hence, interventions are expected to affect both the surrogate end point and the true clinical end point (Figure 1 ). 7 Measurable residual disease (MRD) may have the potential to serve as a surrogate end point for OS in AML. 8 , 9 MRD is the detection of phenotypic or genotypic characteristics at levels below the threshold to define complete remission (CR) by conventional morphology (5%) and is considered the achievement of a deep response . 1 Numerous studies indicate that the absence of MRD is associated with superior disease‐free survival (DFS) and OS in patients with AML who receive different backbones of treatment across AML subtypes, age groups, specimen sources, times of MRD assessment, and detection methods. 10 , 11 , 12 , 13 Current MRD guidelines offer recommendations to support clinical decisions at the patient level. 10 , 11 , 14 , 15 Specifically, these guidelines provide directions on MRD reporting, validating methodologies, timing of MRD measurements, and clinically relevant prognostic thresholds to define relapse risk both in patients who achieve CR and in those undergoing allogeneic hematopoietic stem cell transplantation (allo‐HSCT). 10 , 11 , 14 , 15 Despite the widespread clinical use of MRD to guide treatment decisions at the patient level, MRD has not been adopted as a primary end point in AML clinical trials for drug approvals. Such adoption requires robust validation of MRD‐based surrogate end points as correlates of OS and as more sensitive measures of the quality of disease control among patients achieving morphologic CR. Parallel MRD surrogacy initiatives are ongoing at different stages in other hematologic malignancies, including multiple myeloma, 16 , 17 , 18 acute lymphoblastic leukemia, 19 , 20 and lymphoma. 21

For an end point to be considered a surrogate, it must be more than a correlation between two end point measurements (e.g., MRD and OS) at the patient level. 22 Rather, a surrogate end point must fully capture the treatment effect on the clinical benefit end point, demonstrating trial‐level correlation across multiple clinical trials in a meta‐analysis. 23 However, recent recommendations by the US Food and Drug Administration (FDA) Oncologic Drugs Advisory Committee (ODAC) in April 2024 suggest that data demonstrating individual patient–level associations may be used to support MRD as an early end point for accelerated approvals. The ODAC unanimously supported the use of MRD‐negative CR as an early end point for accelerated approvals in multiple myeloma clinical trials based on strong individual patient–level associations with OS and progression‐free survival despite moderate (yet encouraging) trial‐level associations. 18 This initiative in multiple myeloma may serve as a model to follow for other hematologic malignancies.

The MRD Partnership and Alliance in AML Leukemia Clinical Treatment (MPAACT) is a research alliance that was founded in 2018 among several pharmaceutical companies and academic leaders and was later broadened to include additional industry partners, forming a consortium. MPAACT was created to develop a pathway to establish MRD as a surrogate end point in AML clinical trials, with the goal of accelerating new treatment options for patients. MPAACT has been working collaboratively with regulatory agencies, health technology assessment (HTA) bodies, academic and industry partners, technology vendors, and patient groups. In addition, MPAACT is developing and validating MRD testing methodologies to establish consistent and high‐quality MRD testing for future AML clinical trials.

In this report, we summarize current MRD assessment methods, their use in recent clinical trials, regulatory guidance, and statistical approaches to validate MRD as a surrogate end point as well as patient access considerations.

METHODS OF MRD ASSESSMENT

Various technologies have been developed to assess MRD in AML, including multiparameter flow cytometry (MFC), quantitative polymerase chain reaction (qPCR), real‐time qPCR (RT‐qPCR), digital‐droplet PCR (ddPCR), and next‐generation sequencing (NGS). Table 1 summarizes the advantages and disadvantages of each technology. 1 , 11 , 24

RT‐qPCR‐based MRD assessment is most applicable to distinct molecular AML subgroups, including gene fusions arising from translocations/inversions or hotspot mutations (e.g., NPM1, RUNX1/RUNXT1, CBFB/MYH11, PML/RARA, KMT2A/MLLT3, t[11q23], BCR/ABL ). 11 PCR‐based MRD in molecularly defined AML subgroups is strongly correlated with clinical outcomes, and its assessment is recommended by the European LeukemiaNet (ELN) guidelines. 1 , 10 , 11 However, because of the molecular heterogeneity of AML, only 40%–60% of patients with AML have an abnormality that can be tracked using PCR. 1 , 11 Because missing MRD data can be interpreted as a lack of response or failure by regulatory agencies, 25 , 26 RT‐qPCR is primarily suitable for use in clinical trials involving well defined populations of specific AML genomic subgroups.

MFC‐based MRD evaluation detects leukemic blasts in bone marrow or peripheral blood samples through phenotypic characterization. MFC‐based MRD assessment applies the leukemia‐associated immunophenotype and/or the different from normal strategy to identify and quantify blasts with aberrant expression of cell surface protein markers as the MRD target. 1 , 10 MFC‐based MRD methods are applicable to approximately 85%–90% of patients with AML, making them broadly applicable. 11 However, 10%–15% of patients cannot be reliably monitored using MFC‐based methods because of leukemic cells that lack sufficiently specific immunophenotypic aberrancies. 27 Pure monocytic AMLs are particularly challenging in this regard. A further consideration is that immunophenotypic aberrancies may represent preleukemic or low‐risk myelodysplastic syndrome clones, although usually these are associated with more minor immunophenotypic alterations than AML cells. 28 , 29 , 30 , 31 The ELN Working Group Flow Cytometry Expert Panel has provided recommendations for harmonizing and integrating leukemia‐associated immunophenotype/different from normal aberrant immunophenotypes to detect MRD with a minimum sensitivity of 0.1%. 10 Both the FDA and the ELN recommend reporting the lower limit of detection (LOD) with MFC‐MRD results to substantiate assay performance. 10 , 25 Nonetheless, a lack of interlaboratory standardization may limit the application of MFC‐MRD. 11 , 32 , 33

High‐sensitivity NGS‐based MRD assessment has been shown to inform AML prognosis and therapy in patients with mutations in FLT3‐ITD , 34 , 35 , 36 , 37 NPM1 , 38 , 39 and IDH . 40 Commercially available targeted NGS assays, with a reported LOD as low as 5 × 10 −5 , are commonly used to monitor MRD in patients with known NPM1 or FLT3‐ITD mutations. 36 , 37 , 41 For example, a study of gilteritinib for posttransplantation maintenance in patients with FLT3‐ITD AML used these targeted NGS assays. 36 , 42 Advantages of NGS include the ability to batch test banked samples and to simultaneously obtain information on fusions or mutations. The markers included in a targeted NGS panel should capture all potential disease‐driving aberrations (i.e., somatic gene exon coverages) without sacrificing sensitivity or specificity. 1 A 30‐gene to 50‐gene NGS panel detecting a minimum variant allele frequency (VAF) of 0.1% can identify at least one somatic driver mutation in approximately 80% of patients who have AML. 43 , 44 , 45

Although the NGS multigene MRD assays carry the promise of broader applicability to a highly heterogeneous disease, key limitations include: (1) lower sensitivity compared with MFC and RT‐qPCR outside of molecularly defined subgroups, (2) longer turn‐around time, (3) increased cost, (4) difficulty distinguishing leukemic from preleukemic variants, (5) challenges in quantifying subclones, and (6) insufficient sensitivity to detect low‐VAF variants. 11 , 24 , 45 , 46 Within a 30‐gene to 50‐gene NGS panel, the LOD of each variant in the panel may vary. For example, some variants may be detected with an LOD <0.1%, whereas others have an LOD of 1%. Locally developed NGS panels may have a lower overall sensitivity, with an LOD approximating 1%, making them less relevant for MRD detection.

In its 2022 guidance for industry, the FDA referenced previous acceptance of bone marrow MRD <0.01% as evidence of efficacy. 26 Newer variations on standard NGS workflow, including target error‐corrected sequencing, ultradeep sequencing, or duplex sequencing, offer potential gains in sensitivity. 47 , 48 , 49 A recent study of a 29‐gene panel using duplex sequencing detected VAFs ranging from 2.1 × 10 −2 to 9.1 × 10 −6 . 48 Efforts are also underway to validate panel‐based RNA fusion sequencing to evaluate MRD. 50 However, these newer variations on NGS workflow are currently for research use only and have yet to be validated in a clinical setting. Based on pilot testing conducted by MPAACT, existing NGS platforms may not consistently reach 0.01% sensitivity across the large range of genes needed to encompass the mutational heterogeneity of patients with AML. In addition, IDH1/IDH2 mutations can be identified in preleukemic clones, potentially complicating their interpretation as indicators of relapse in IDH1/IDH2 ‐mutated AML. 51 , 52

Dynamic approaches—such as assessing log reduction from baseline or between two treatment timepoints—may be alternatives to a binary definition of MRD positivity at a single timepoint, particularly in patients with high initial tumor burden. However, obtaining good‐quality samples from multiple timepoints is a challenge, as demonstrated by clinical trials using sequential monitoring. 53 , 54 , 55 The quality of samples by itself is a critical requirement from a health authority standpoint, and nonoptimal samples as well as missing samples are considered MRD‐positive by default. 25 , 26

Gaps in analytical and test validation, as well as inconsistent assay performance and MRD data generation, often prevent the inclusion of MRD data in prescribing information. However, MRD data in prescribing information can help clinicians assess a patient's likelihood of relapse and guide treatment decisions. In a study evaluating the acceptance of MRD data into US prescribing information between January 2014 and February 2021, 55 (28%) of 196 drug applications included such data. 56 Among the applications with MRD data, 13% were for AML, and none had MRD data included in the final label. The predominant reasons for excluding MRD data from US prescribing information were issues with analytical and/or test validation (65%) and assay performance (24%). 56 A review of data from an internal FDA database revealed similar findings. 57 Among 13 new drug and biologics licensing applications for hematologic malignancies submitted between 2014 and 2016, only six products had MRD data added to their prescribing information. Reasons for excluding MRD data from the label included missing data, inadequate test performance data (e.g., LOD), inconsistent sample sources (bone marrow or peripheral blood), high test failure rates (e.g., failure to identify a clonal rearrangement), insufficient test validation, and improper statistical analyses. 57 Hence, rigorous and complete MRD data collection and assay performance reporting are essential.

MPAACT aims to address previous analytical and methodological issues that have prevented the integration of MRD data into product information. To achieve this goal, MPAACT collaborates with cross‐industry, academic, and technology partners to develop and validate well established and emerging MFC‐MRD and NGS‐MRD assays. MPAACT has also engaged with US and EU regulatory agencies to discuss methods used to harmonize and validate MFC‐MRD and NGS‐MRD assays. The goals of these ongoing, collaborative efforts are to establish uniformity and rigor in MRD measurements needed for inclusion in prescribing information, support MRD as a surrogate end point in clinical trials, and inform MRD detection methods in routine practice.

MRD assessment in clinical trials of adult and pediatric patients with AML

A 2020 meta‐analysis of 81 studies involving 11,151 adult and pediatric patients with AML indicated that achieving MRD‐negative status was associated with superior 5‐year DFS (64% vs. 25%, respectively) and OS (68% vs. 34%, respectively) compared with MRD‐positive status. 13 The effect of MRD on DFS and OS was observed in all evaluated subgroups: adult and pediatric patients, MRD assessed at induction and during and after consolidation, detection method (MFC, PCR, and NGS), AML subtype (core‐binding factor or noncore‐binding factor), and specimen source (bone marrow or peripheral blood). 13 A subsequent meta‐analysis reported a similar effect of MRD status on OS when comparing studies that included only patients who achieved CR versus studies that included all patients. 12 Although these meta‐analyses provided broad insights into the effect of MRD in various subgroups, adult and pediatric patients differed in molecular drivers of disease and standard‐of‐care treatment regimens (including transplantation); trials also differed in methodologies and clinical context. Herein, we summarize the current trial evidence in adult and pediatric patients with AML separately.

In adults, the prognostic value of MRD for relapse and OS has been extensively studied in treatment‐naive patients who received intensive chemotherapy as their initial AML therapy, including older patients. 45 , 58 , 59 , 60 , 61 Data supporting MRD assessments are emerging for newer therapies, including add‐ons to traditional intensive chemotherapy, lower intensity regimens for patients deemed unfit for traditional regimens, as well as maintenance therapies for patients who achieved remission with or without HSCT. 6 , 34 , 35 , 36 , 38 , 39 , 40 , 62 , 63 , 64 , 65 , 66 , 67 Table 2 summarizes the studies on MRD assessments and newer therapies in adult patients with AML. 6 , 34 , 35 , 36 , 38 , 39 , 40 , 62 , 64 , 65 , 66 , 68 , 69 Overall, despite various settings and MRD assessment methods, MRD positivity consistently predicted relapse and shorter OS in multivariate analyses. 35 , 38 , 39 , 64 , 65 The timepoints for measuring MRD included day 21 of the first or second induction, the end of the first or second induction courses (EOI1 and EOI2, respectively), first or second CR, as well as pre‐HSCT and post‐HSCT. A few trials also longitudinally measured MRD. 6 , 64 Uniformly, more patients receiving experimental treatments converted from MRD‐positive to MRD‐negative status, 6 , 34 , 38 , 63 , 65 which translated to improved survival and lower rates of relapse. Furthermore, the potential role of MRD‐guided maintenance therapy posttransplantation was demonstrated in patients who had FLT3 ‐mutated relapsed/refractory AML. 36 , 70 FLT3 inhibitor maintenance reduced the risk of relapse only in patients who were MRD‐positive before and/or after HSCT but provided no advantage for MRD‐negative patients. 36 , 70

Similar to the findings in adult patients, positive MRD status is an independent adverse prognostic factor for EFS, relapse‐free survival (RFS), cumulative incidence of relapse, and OS in pediatric patients with newly diagnosed AML. 71 , 72 , 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 Table 3 summarizes the pediatric population, MRD assessment timepoints, and clinical outcomes in recent representative trials. 71 , 73 , 75 , 85 , 86 , 87 Overall, these studies assessed MRD using different methods and cutoffs at various timepoints (from EOI1 to posttransplantation) and reported various clinical outcomes (RFS, EFS, OS, and cumulative incidence of relapse). Negative MRD status consistently demonstrated prognostic significance and improved clinical outcomes. Similar conclusions were reported in a systematic review of 13 prospective and retrospective studies 73 , 75 , 76 , 77 , 78 , 79 , 82 , 83 , 84 , 85 , 86 , 87 , 88 in patients aged birth to 21 years with de‐novo AML who received either chemotherapy alone or chemotherapy followed by HSCT. 89 Included studies in the systematic review assessed MRD with either MFC or RT‐qPCR. MRD was assessed 15–28 days after treatment initiation, with many studies measuring MRD after each treatment cycle, whereas one study assessed MRD 14 days before HSCT. 89 Although some studies suggested that MRD negativity at EOI2 or later may be superior to MRD negativity at EOI1 in its association with improved outcomes, 71 , 77 , 80 , 81 , 82 , 90 one study indicated that both the EOI1 and EOI2 timepoints were predictive of outcomes. 72 The systematic review concluded that the optimal timepoint and MRD assessment technique could not be defined. 89

MFC with a single threshold (predominantly ≥0.1%) to define MRD positivity has been the most common method to measure MRD in both pediatric and adult trials. However, more recent investigations in adults have used other methods, such as RT‐qPCR, 38 , 39 digital PCR, 6 , 40 NGS, 34 , 35 and messenger RNA quantification with peripheral blood and/or bone marrow samples. 70 A study using NGS and MFC to detect MRD in 482 adult patients with newly diagnosed AML who achieved CR after induction therapy revealed that NGS may provide additive prognostic value. 45 In contrast to adult studies, methods other than MFC to assess MRD were less common in the pediatric population.

Another major difference between adult and pediatric trials assessing MRD is the inclusion of a comparison arm. A comparison arm is more common in adult clinical trials. Therefore, treatment effects on MRD were demonstrated in many adult clinical trials that compared experimental treatments versus controls (Table 2 ) 6 , 34 , 38 , 39 , 62 , 65 but were not as well documented in pediatric AML trials. 91

Despite considerable data supporting the prognostic value of MRD, limitations in these analyses prevented regulatory agencies from using MRD data for regulatory approval in AML. 92 These limitations included insufficient data to determine clinically relevant MRD levels, different MRD assessment methods, lack of assay sensitivity information, and dissimilar MRD thresholds and assessment timepoints. In addition, not all studies evaluated MRD only in patients in CR, and the favorable association between MRD, DFS, and OS could be caused in part by differences in remission status. 92 To overcome these limitations, MPAACT has engaged with US and EU regulatory agencies, whose guidance informed the proposed meta‐analysis to evaluate MRD as a surrogate end point in AML clinical trials, with adults and pediatric patients analyzed separately, unless future feasibility assessments support the pooling of pediatric and adult data sets.

Current regulatory guidance

Developing MRD as a clinical trial surrogate end point requires support from regulatory agencies. Since 2019, MPAACT has had health authority interactions with the FDA, the European Medicines Agency (EMA), and the Medicines and Healthcare Products Regulatory Agency and received their guidance on developing a meta‐analysis to evaluate MRD as a surrogate end point. The recommendations encompassed the inclusion of large numbers of RCTs, qualification criteria and strategies, data set homogeneity, statistical methods used to evaluate surrogacy, and sensitivity analysis.

Guidance from the US Food and Drug Administration

The FDA has defined a surrogate end point as a biomarker intended to substitute for a clinical end point . 93 A surrogate end point should be reasonably likely to predict clinical benefit to support accelerated approval for a new treatment. 93 In relapsed or refractory acute leukemia, the FDA has accepted bone marrow MRD <0.01% as supporting evidence of efficacy for new drugs if durable CR is achieved. 26 This guidance currently extends to AML. However, the agency acknowledges that the MRD level needed to substantiate an efficacy claim may change as technologies advance and new clinical findings are published. 26 Furthermore, the FDA considers it impractical to use MRD to identify relapse and determine duration of response (DOR), EFS, or RFS in clinical trials because of the need for frequent bone marrow sampling. Hence, until a peripheral blood MRD assay is validated, monitoring patients for hematological relapse to define DOR, EFS, and RFS may be more practical. 26 CR and DOR have been acceptable end points for accelerated approvals in AML. 25 , 26 , 94 Therefore, the relative efficacy end point is not MRD negativity alone, but MRD‐x CR, where x is a prespecified level of MRD. 25 Determination of the threshold level to establish MRD‐x and its effect on outcomes should be data‐driven based on meta‐analyses of many RCTs, both positive and negative.

Guidance from the European Medicines Agency

The EMA advises that MRD should be established as a predictive factor in addition to being a prognostic factor to convincingly demonstrate clinical benefits based on MRD. Other efficacy end points must support the MRD analysis, including the absence of negative trends in EFS or OS at the time of the primary review. A substantial effect by the drug candidate on MRD itself would also be required to suggest a potential benefit on survival end points. MRD at the time of CR is the most clinically relevant timepoint for assessment and is also recommended by the ELN. 10

The EMA recently published its Letter of Support for MRD as a Surrogate Efficacy Endpoint in Clinical Studies with Acute Myeloid Leukemia based on qualification guidance provided to MPAACT. 95 The letter provided recommendations on the general approach, analysis population, methodology for the meta‐analysis, and MRD assessment methods. The EMA supported the two‐stage approach proposed by MPAACT (see Statistical Considerations, below). The EMA also recommended distinguishing between transplanted and nontransplanted patients and analyzing surrogacy in homogeneous groups of patients first. Furthermore, it suggested exploratory sensitivity analyses (e.g., intensive vs. nonintensive regimens) and prespecified analyses on heterogeneous subgroups (e.g., pediatric patients, fit vs. unfit patients). 95

Beyond statistical considerations, the EMA also provided guidance on MRD assessment methods. 95 Although various MRD methods can be grouped for the meta‐analyses, subgroup analyses by methods (NGS vs. PCR vs. MFC) should be performed. 95 Furthermore, different MRD negativity thresholds should be evaluated. The EMA acknowledged the challenges in developing an NGS‐MRD methodology with acceptable sensitivity and specificity across both known variants (e.g., for disease monitoring) and de novo or unknown variants (e.g., in baseline samples). In principle, the EMA supported MPAACT's proposal to use a tiering approach with de novo mutations called at a higher VAF cutoff than known mutations . 95

Guidance from the Medicines and Healthcare Products Regulatory Agency

The Medicines and Healthcare Products Regulatory Agency advised that, first, MPAACT should demonstrate age as a noncritical factor before combining adult and pediatric data sets for further analyses. This approach will strengthen the support for MRD as a surrogate end point and generalizability of the findings. In the sensitivity analysis on age, surrogacy ideally should be demonstrated across a range of end points instead of one specific end point. In addition, data from various timepoints will increase applicability for future trials. Addressing the kinetics of MRD is important but is currently limited by data availability.

Early engagements with health authorities have guided MPAACT's proposed meta‐analyses and efforts to validate and standardize MRD assessment methodologies. MPAACT plans to seek ongoing input from health authorities to collaboratively overcome challenges that prevent the deployment of MRD surrogacy in global registration trials.

Statistical considerations

For an end point to be considered a surrogate, it must be more than a correlation between two end point measurements (e.g., MRD and OS) at the patient level (i.e., individual patient–level association). 22 Rather, a surrogate end point must reliably and accurately predict the treatment effect on the clinical benefit end point, demonstrating trial‐level correlation across multiple clinical trials in a meta‐analysis (i.e., trial‐level association). 23 Strong trial‐level correlation is the primary criterion for validating a surrogate end point that can replace the true clinical end point in randomized trials. 23 , 96 A meta‐analytic, multitrial approach is the preferred and crucial method for surrogate end point validation in oncology across different tumor types. 25 Meta‐analyses conducted to validate surrogacy differ from standard meta‐analyses commonly published in the literature, such as the analyses performed by Short et al. 13 and Segerink et al. 89 Table 4 summarizes the differences among individual patient–level analysis (stage 1), trial‐level analysis (stage 2), and a standard meta‐analysis.

MPAACT will conduct a meta‐analysis to determine whether an MRD status–based end point can serve as a valid surrogate for OS in both adult and pediatric patients with newly diagnosed AML. The meta‐analytic approach will closely align with regulatory guidance and other successful prior surrogacy initiatives, such as 30‐month CR as a surrogate end point for traditional approvals in follicular lymphoma 97 and MRD as an early clinical end point for accelerated approvals in multiple myeloma. 18 Treatment effects measured on both MRD end point(s) and OS from RCTs are required for trial‐level correlation and individual patient–level evaluation.

MPAACT has established strong partnerships with AML trial sponsors, academic investigators, and AML study groups to share individual patient trial data. Based on a simulation study examining the estimation performance of trial‐level surrogacy using either conventional or model‐based approaches, the trial‐level surrogacy can be estimated accurately and precisely when including a sufficient number of trials (at least six) and if the range of treatment effects across trials is sufficiently large (hazard ratio, 0.5–2.0). 98 As of April 2025, data set transfers were completed for 10 RCTs and initiated for two single‐arm studies. In addition, data transfer or data use agreements were in process for an additional four RCTs. Hence, we anticipate that the proposed meta‐analyses will be feasible.

The Mayo Clinic SHARE (Sargent's Hypothesis Analysis in the Research of Endpoints) team serves as the independent statistical data‐monitoring center and will perform the proposed analyses. After preparatory analysis, data analysis will be conducted in two stages. as below (Figure 2 ).

Preparatory analyses

Before stage 1 analysis, single‐arm trials and control arms from RCTs will be used to inform the selection of the primary surrogate candidate. Key aspects to be evaluated include: (1) a threshold to define MRD‐negative status, (2) the optimal timepoint for MRD end point assessments; (3) the optimal method for incorporating conventional response measures, and (4) whether MRD negativity based on the same threshold, but determined by different methods (MFC, qPCR, or NGS), carries the same or similar prediction power for OS. To maintain data independence for the trial‐level correlation analyses, experimental arms in RCTs will not be used in the preparatory analyses.

Stage 1: Individual patient–level surrogacy

The evaluation of individual patient–level surrogacy based on RCT data is needed in surrogate end point validation alongside trial‐level analysis. Individual patient–level surrogacy is estimated by the global odds ratio using a Plackett copula model, which simultaneously estimates both individual patient–level and trial‐level correlations. 99 The global odds ratio measures the ratio of the odds of being alive for patients who achieved MRD negativity compared with those who did not. The higher the value above 1.0, the stronger the individual patient–/patient‐level correlation. A 95% confidence interval excluding 1.0 indicates statistical significance. A landmark survival analysis beginning at the time of MRD assessment, pooling patients across trials, will be used as a supportive analysis to assess the prognostic value of the MRD end point.

Stage 2: Trial‐level surrogacy

In stage 2, formal trial‐level surrogacy will be the primary measure for decision making on the use of MRD as a surrogate end point. Trial‐level surrogacy measures how precisely the treatment effect measured by OS (e.g., hazard ratio) can be predicted according to the treatment effect measured by the MRD end point (e.g., odds ratio). The trial‐level surrogacy is measured by the coefficient of determination estimated by two methods: the weighted least‐squares regression method 100 and the joint modeling method using a bivariate Plackett copula model. 99 Both of these coefficient of determination R 2 values are bound within the interval from 0 to 1, with values near one indicating strong trial‐level surrogacy and values near zero suggesting no surrogacy. Point estimates and 95% confidence intervals will be reported along with graphic representations of the relation between treatment effects on MRD and OS.

Data analysis and allo‐HSCT

Allo‐HSCT is the most potent consolidation therapy for patients with AML and is the standard treatment for patients with high‐risk or relapsed AML. 101 , 102 Because the option to receive HSCT depends on a patient's response to initial treatment, HSCT can be considered an outcome of the treatment rather than an independent new course of treatment. Hence, survival outcomes will be assessed regardless of whether patients received HSCT. Similarly, the option to receive consolidation and postconsolidation treatment is also an outcome of treatment because only fit patients who achieve remission qualify to continue treatments. Therefore, OS and EFS analyses will not be censored at the time of allo‐HSCT or treatment modification. MRD at the time of CR before and after allo‐HSCT, including CR with incomplete recovery, CR with partial hematologic recovery, and CR with incomplete platelet recovery, will be considered.

Subgroup and supplemental analyses

Depending on the number of trials and the amount of patient‐level data available, patient subgroup analyses will be performed based on cytogenetic risks according to ELN guidelines, 101 age, sex, region, and Eastern Cooperative Oncology Group performance status. Trial subgroups will include treatment type (e.g., targeted therapy vs. chemotherapy), setting (intensive vs. nonintensive chemotherapy), treatment sequence (protocol‐defined salvage or consolidation therapy), prior transplantation, and MRD assessment methods (molecular [NGS or PCR] vs. cellular [MFC] methods).

We will also estimate the surrogate threshold effect (STE). STE is the minimal treatment effect size on the surrogate end point that needs to be observed to conclude a nonzero treatment effect on the true end point. 103 STE estimation offers insights into the usefulness of the surrogate in particular applications, but its use may be limited by the absence of clear guidance on STE values indicating adequate surrogacy. We will report the STE estimation results if the convergence of the prediction limit functions is satisfied.

Potential updates to the proposed statistical analysis plan

Recently, the meta‐analytic approach was used to validate MRD as an early clinical end point reasonably likely to predict clinical benefit, supporting accelerated approvals in multiple myeloma clinical trials. 18 The conclusion was based on strong individual patient–level associations and encouraging trial‐level associations. Consequently, MPAACT is exploring the option of establishing individual patient–level associations to support MRD as an early end point for accelerated AML approvals while formal surrogacy with trial‐level correlation, which requires more RCTs, is being established.

Patient access

In addition to support from regulatory agencies, acceptance by HTA bodies and payers is needed for an MRD surrogate end point to be considered for reimbursement purposes and patient access. Whereas regulatory agencies are primarily concerned with evidence supporting efficacy and safety, HTA bodies focus on evaluating the long‐term comparative effectiveness and economic effect of health technologies while also taking into account unmet needs, equity, and severity of disease. 104 Hence, relying on surrogate end points may create challenges for health care decision makers. For example, surrogate end points may fail to capture the overall risk‐benefit profile of a health technology. 105 In addition, from the perspective of health care systems, superiority demonstrated with a surrogate end point may not translate into value for money spent. 105 Finally, larger treatment effects observed on surrogate end points than on the final outcomes may lead to an overestimation of clinical benefits. 104

For MRD to be accepted as a surrogate end point by HTA bodies, the validation process extends beyond regulatory requirements. In addition to the requirements summarized in Figure 1 , HTA bodies are also concerned with the magnitude of the surrogates treatment effect that would translate into a meaningful benefit on the final outcome of interest, 105 such that the final relevant outcomes may include not only clinical events (e.g., death) but also health‐related quality of life. 104 Methodological guidance from HTA bodies is critical to establish MRD as a surrogate end point for reimbursement purposes. However, in a recent survey, only seven of 73 international HTA bodies offered detailed guidance on surrogate end point methodological considerations. 106 Furthermore, HTA bodies vary in their focus within their methodological guidance documents. These findings underscore the opportunities and need for HTA bodies to clarify and harmonize methodological approaches for validating surrogate end points.

MPAACT is working to understand HTA and payer needs and to address potential concerns. In addition, whereas reimbursement discussions traditionally center on specific products, the broader implications of MRD as a surrogate end point require a nonproduct, disease‐level dialog. We will seek input from the European Network for Health Technology Assessment, country‐level HTA bodies, as well as academic and clinician experts in our process.

DISCUSSION

Here, we describe a comprehensive framework to establish MRD as a surrogate for OS in AML clinical trials to support regulatory approvals. Despite substantial evidence supporting the prognostic value of MRD, regulatory acceptance of MRD end points for AML approvals is hindered by inconsistencies in assessment methods, lack of standardized assay sensitivity, various MRD thresholds and assessment timepoints, and inclusion of non‐CR patients across studies. MPAACT will address these limitations by developing a uniformly defined MRD end point across all RCTs in the meta‐analysis to rigorously assess MRD surrogacy.

Although the number of RCTs obtained by MPAACT should be sufficient to support the proposed individual patient–level analyses, potential challenges for trial‐level surrogacy evaluation remain, including incomplete data (e.g., missing MRD or OS end point data). Nevertheless, a robust individual patient–level association with a moderate trial‐level association may support accelerated approvals within the current regulatory framework. Recently, the FDA ODAC unanimously supported MRD as an early end point that is reasonably likely to predict treatment effects on long‐term end points and can support accelerated approvals in multiple myeloma clinical trials. 18 The ODAC decision provided a concrete pathway for validating an early end point for accelerated approvals in AML if formal trial‐level surrogacy could not be established in the short term. Specifically, to qualify an early end point for accelerated approval, a strong individual patient–level correlation can be considered sufficient, and a promising trial‐level correlation will be supportive. The primary and supportive analyses of this approach differ from our initially proposed meta‐analysis. MPAACT is consulting with regulatory and HTA bodies on using individual patient–level correlation to qualify for early end point in accelerated AML approvals, while formal trial‐level surrogacy is being established.

It is important to acknowledge that MPAACT's approach may leave some questions partially addressed or unanswered. To minimize heterogeneity of the analysis cohort, trials in relapsed/refractory disease and pediatric patients were excluded from the current data set. MRD surrogacy in these populations will require a separate validation. Likewise, once established, the applicability of this MRD framework among specific clinically and molecularly defined subgroups may be further validated in other academic and real‐world data sets. Innovation is critically important for improving the outcome of patients with AML. Although the purpose of MPAACT is to establish a standard framework for the use of MRD as a surrogate end point, we recognize that ongoing efforts and investment will likely refine it further. These efforts may include evaluating MRD measured by emergent technologies, such as more sensitive or blood‐based assays, evolving treatment strategies and objectives (e.g., OS vs. EFS), and updated MRD assessment workflows and pipelines, including advancements in artificial intelligence or machine learning in assessing MRD. 107 , 108 , 109 In addition, improving the affordability of these techniques will be vital to enable their global deployment.

Upon successful demonstration and acceptance of MRD as a surrogate end point for clinical trials, the next step will be to standardize MRD testing across trials. Treatment centers often use MRD protocols developed locally or through collaborative groups. These tests are validated using individual laboratories' internal standards. The interlaboratory work led by the ELN is an important step toward standardizing MRD measurement on a broader scale. 10 Nevertheless, local guidelines and the ELN standardization efforts have not been developed using standards expected by health authorities. These tests may inform decisions at the patient level, such as guiding consolidation or transplantation decisions, but are not suited to support regulatory approvals to make new treatments available to patients.

Beyond conducting a robust surrogacy analysis, parallel health policy initiatives are essential to ensure that therapies with demonstrated clinical benefits using MRD end points are reimbursed and accessible to patients. To achieve this reimbursement status, HTA bodies must first accept MRD as a valid end point to define clinical benefit. The use of co‐primary or key secondary end points in AML trials may bridge this gap by demonstrating early MRD benefits that support conditional approval/reimbursement, followed by conventional time‐dependent end points at the final read‐out of the trial. A successful example is the approval of blinatumomab in MRD‐positive precursor B‐cell acute lymphoblastic leukemia based on the results of the BLAST (BiTE Antibody Blinatumomab in Adult Patients with Minimal Residual Disease of B‐precursor Acute Lymphoblastic Leukemia) trial (ClinicalTrials.gov identifier NCT01207388 ), which used MRD as a primary end point and RFS/OS as secondary end points. 20 , 110 Another example is the approval of arsenic trioxide in relapsed acute promyelocytic leukemia based on MRD negativity as a secondary end point. 111

Our work to validate MRD as a surrogate end point for OS requires coordinated efforts among diverse stakeholders. Academic scientists and AML study groups play pivotal roles in pioneering novel AML therapeutics. The proposed meta‐analysis relies heavily on data from their RCTs. Furthermore, their expertise will facilitate the design of the meta‐analysis and data interpretation. Other key contributors include diagnostic companies, regulators, payers, and, importantly, patients and their advocates. The extensive and ongoing collaboration between MPAACT and global partners from industry and academia includes the sharing of data, resources, and intellectual capacity. This robust collaboration underscores our shared commitment and the strength of this alliance. MPAACT is uniquely positioned to lead the cross‐industry initiative to advance MRD as a clinical trial surrogate end point, to accelerate AML drug development and deliver innovative treatments to patients.

AUTHOR CONTRIBUTIONS

Michael Boyiadzis: Conceptualization, data curation, investigation, supervision, visualization, project administration, writing—original draft, and writing–revising and editing. Andrew H. Wei: Data curation, investigation, and writing–revising and editing. Bruno Paiva: Data curation, investigation, and writing–revising and editing. Sylvie D. Freeman: Data curation, investigation, and writing–revising and editing. Gertjan Kaspers: Data curation, investigation, and writing–revising and editing. Brenda Chyla: Data curation, supervision, visualization, project administration, and writing–revising and editing. Sarah Hersey: Data curation, visualization, project administration, writing–original draft, and writing–revising and editing. Reshma Patel: Data curation, supervision, writing–original draft, and writing–revising and editing. Brian Maloney: Data curation, writing–original draft, and writing–revising and editing. Marie‐Helene Blanchet Zumofen: Data curation and writing–revising and editing. Marlies Van Hoef: Data curation, writing–original draft, and writing–revising and editing. Beate Wulff: Data curation, supervision, and writing–revising and editing. Vanessa Obourn: Data curation, visualization, writing–original draft, and writing–revising and editing. Shalaka Patel: Data curation, visualization, writing–original draft, and writing–revising and editing. Daniel Lopes de Menezes: Data curation, supervision, and writing–revising and editing. Qian Shi: Conceptualization, methodology, writing–original draft, and writing–revising and editing. Bourras‐Rezki Bengoudifa: Methodology, writing–original draft, and writing–revising and editing. Natalie Dimier: Methodology and writing–revising and editing. Thomas J. Prior: Methodology and writing–revising and editing. Gail J. Roboz: Conceptualization, methodology, data curation, investigation, and writing–revising and editing. Thomas Prebet: Conceptualization, data curation, investigation, supervision, visualization, project administration, writing–original draft, and writing–revising and editing. All authors approved the final version of the article.

CONFLICT OF INTEREST STATEMENT

Michael Boyiadzis is an employee of OncoVerity and participates in the employee stock program. Andrew H. Wei reports Institutional research funding from Novartis, AbbVie, Amgen, Astex, AstraZeneca, BMS, Janssen, Novartis, Servier, and Syndax; advisory board fees from AbbVie, Amgen, Astellas, AstraZeneca, Beigene, Bristol Myers Squibb Company (BMS), Gilead, GlaxoSmithKline, Janssen, Jazz Pharma, Novartis, Pfizer, Roche, and Servier, personal/consulting fees from AbbVie, Aculeus, Novartis, Servier, and Shoreline; participation in speakers' bureaus for AbbVie, Astellas, BMS, Novartis, and Servier; and financial benefits as an employee of the Walter and Eliza Hall Institute related to milestone and royalty payments made to the institute in connection with the development of Venetoclax outside the submitted work. Bruno Paiva reports unrestricted grants from Celgene, EngMab, Roche, Sanofi, and Takeda Pharmaceutical Company; personal/consulting fees from BMS/Celgene, Janssen Biotech, and Sanofi US Services Inc.; and honoraria for lectures/membership on advisory boards with Adaptive Biotechnologies, Amgen, BMS, Celgene, Janssen Biotech, Kite Pharma, Roche, Sanofi, Takeda Pharmaceutical Company, and The Binding Site outside the submitted work. Sylvie D. Freeman reports institutional research funding from AstraZeneca, BMS/Celgene, Cytek, and Jazz Pharmaceuticals; personal/consulting fees from the Measurable Residual Disease Partnership and Alliance in Acute Myeloid Leukemia Clinical Treatment Consortium (MPAACT) and Novartis; honoraria from BMS/Celgene, Janssen, and Novartis; and jointly holds a patent (US Patent Application 15/842,240, 2018, 2024) with Vyas (Oxford University), all outside the submitted work. Brenda Chyla is an employee of AbbVie and owns stock and stock options in the company. Sarah Hersey reports personal/consulting fees and support for other professional activities from the Personalized Medicine Coalition; board membership at the Personalized Medicine Coalition; independent board membership at Sampled; has a board observer role at Precede Biosciences; reports ownership of stock in Johnson and Johnson and Novartis; and is an employee of BMS and owns stock in the company, all outside the submitted work. Reshma Patel is an employee of Johnson and Johnson. Brian Maloney reports previous employment with Johnson and Johnson. Marie‐Helene Blanchet Zumofen is an employee of F. Hoffmann‐La Roche Ltd and reports stock ownership in Roche Holding AG and Moderna. Marlies Van Hoef reports ownership of Van Hoef Consulting and is a former employee of Novartis Pharma outside the submitted work. Beate Wulff reports employment, equity ownership, and patent royalties from F. Hoffmann‐La Roche; and is a member of the Board of Directors for the aPODD Foundation, UK. Vanessa Obourn is an employee of Novartis and reports stock ownership in Novartis and BMS. Shalaka Patel is an employee of BMS. Daniel Lopes de Menezes reports employment, ownership of equity stocks and options, and patent royalties from BMS. Qian Shi reports institutional grants/contracts from BMS, Janssen Biotech, the MPAACT Consortium, Novartis, Regeneron Pharmaceuticals, and Roche/Genentech USA Inc.; personal/consulting fees from AbbVie, Chugai Pharmaceutical Company Ltd., HopeAI, Kronos Bio, and Regeneron Pharmaceuticals; data and safety monitoring at BeOne Medicines USA Inc., BMS College of Engineering, Exelixis Inc., Genmab, the Hoosier Cancer Research Network, Mirati Therapeutics Inc., and Yiviva Inc.; and support for other professional activities from Chugai Pharmaceuticals Company Ltd. outside the submitted work. Bourras‐Rezki Bengoudifa is an employee of BMS. Natalie Dimier is an employee of Hoffmann‐La Roche Ltd. Thomas J. Prior reports stock and stock option ownership in Johnson & Johnson; and he is an employee of Janssen Research & Development LLC and participates in their employee stock program outside the submitted work. Gail J. Roboz reports grants/contracts from Cellectis and Janssen; and personal/consulting fees from AbbVie, Actinium Pharmaceuticals, Agios Pharmaceuticals Inc., Amgen, Amphivena, Argenyx, Array BioPharma, Astellas Pharma, Astex, AstraZeneca, Bayer, Bluebird Bio, BMS, Caribou Biosciences Inc., Celgene Corporation, Celltrion Healthcare, Daiichi Sankyo Company Ltd., Eisai Inc., Epizyme, F. Hoffmann‐La Roche, GlaxoSmithKline, Geron, Glycomimetics, Helsinn Therapeutics Inc., Janssen Global Services LLC, Jasper Pharmaceuticals, Jazz Pharmaceuticals, Molecular Partners, Morphosys, Neogenomics, Novartis, Oncoprecision, OncoVerity, Orsenix, Otsuka Pharmaceutical, Pfizer, Rigel Pharmaceuticals Inc., Sandoz, Syndax, Takeda Oncology, Telix, and Trovagene outside the submitted work. Thomas Prebet is an employee of BMS. Gertjan Kaspers disclosed no conflicts of interest.

ACKNOWLEDGMENTS

Medical writing and editorial support for this article was provided by Annie Cheang, PharmD, MS, from Med‐Ed Link, under the direction of the authors, and was funded by the Measurable Residual Disease Partnership and Alliance in Acute Myeloid Leukemia Clinical Treatment Consortium.

Boyiadzis M, Wei AH, Paiva B, et al. Measurable residual disease (MRD) as a surrogate end point for clinical drug approval in acute myeloid leukemia (AML): perspectives from the MRD Partnership and Alliance in AML Clinical Treatment Consortium. Cancer. 2025;e35960. doi: 10.1002/cncr.35960

The first two authors contributed equally to this article.

The last two authors jointly supervised this work.

REFERENCES

1. Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345‐1377. doi: 10.1182/blood.2022016867

2. DiNardo CD, Wei AH. How I treat acute myeloid leukemia in the era of new drugs. Blood. 2020;135(2):85‐96. doi: 10.1182/blood.2019001239

3. National Cancer Institute (NCI) . Surveillance, Epidemiology, and End Results (SEER) Program. Surveillance Research Program, Division of Cancer Control and Prevention. NCI. 2024. Accessed October 30, 2024. https://seer.cancer.gov/statistics‐network/explorer/overview.html

4. Ganzel C, Sun Z, Cripe LD, et al. Very poor long‐term survival in past and more recent studies for relapsed AML patients: The ECOG‐ACRIN experience. Am J Hematol. 2018;93(8):1074‐1081. doi: 10.1002/ajh.25162

5. Norsworthy KJ, Gao X, Ko CW, et al. Response rate, event‐free survival, and overall survival in newly diagnosed acute myeloid leukemia: US Food and Drug Administration trial‐level and patient‐level analyses. J Clin Oncol. 2022;40(8):847‐854. doi: 10.1200/jco.21.01548

6. Montesinos P, Recher C, Vives S, et al. Ivosidenib and azacitidine in IDH1‐mutated acute myeloid leukemia. N Engl J Med. 2022;386(16):1519‐1531. doi: 10.1056/nejmoa2117344

7. Fleming TR, Powers JH. Biomarkers and surrogate endpoints in clinical trials. Stat Med. 2012;31(25):2973‐2984. doi: 10.1002/sim.5403

8. Aitken MJL, Ravandi F, Patel KP, Short NJ. Prognostic and therapeutic implications of measurable residual disease in acute myeloid leukemia. J Hematol Oncol. 2021;14(1):137. doi: 10.1186/s13045-021-01148-5

9. Dillon R, Potter N, Freeman S, Russell N. How we use molecular minimal residual disease (MRD) testing in acute myeloid leukaemia (AML). Br J Haematol. 2021;193(2):231‐244. doi: 10.1111/bjh.17185

10. Heuser M, Freeman SD, Ossenkoppele GJ, et al. Update on MRD in acute myeloid leukemia: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2021;138(26):2753‐2767. doi: 10.1182/blood.2021013626

11. Schuurhuis GJ, Heuser M, Freeman S, et al. Minimal/measurable residual disease in AML: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2018;131(12):1275‐1291. doi: 10.1182/blood-2017-09-801498

12. Short NJ, Fu C, Berry DA, et al. Association of hematologic response and assay sensitivity on the prognostic impact of measurable residual disease in acute myeloid leukemia: a systematic review and meta‐analysis. Leukemia. 2022;36(12):2817‐2826. doi: 10.1038/s41375-022-01692-0

13. Short NJ, Zhou S, Fu C, et al. Association of measurable residual disease with survival outcomes in patients with acute myeloid leukemia: a systematic review and meta‐analysis. JAMA Oncol. 2020;6(12):1890‐1899. doi: 10.1001/jamaoncol.2020.4600

14. National Comprehensive Cancer Network (NCCN) . NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®). Acute Myeloid Leukemia, Version 2.2025. NCCN. 2025. Accessed May 13, 2025. https://www.nccn.org/professionals/physician_gls/pdf/aml.pdf

15. Dholaria B, Savani BN, Hamilton BK, et al. Hematopoietic cell transplantation in the treatment of newly diagnosed adult acute myeloid leukemia: an evidence‐based review from the American Society of Transplantation and Cellular Therapy. Transplant Cell Ther. 2021;27(1):6‐20. doi: 10.1016/j.bbmt.2020.09.020

16. Munshi NC, Avet‐Loiseau H, Anderson KC, et al. A large meta‐analysis establishes the role of MRD negativity in long‐term survival outcomes in patients with multiple myeloma. Blood Adv. 2020;4(23):5988‐5999. doi: 10.1182/bloodadvances.2020002827

17. Anderson KC, Auclair D, Adam SJ, et al. Minimal residual disease in myeloma: application for clinical care and new drug registration. Clin Cancer Res. 2021;27(19):5195‐5212. doi: 10.1158/1078-0432.ccr-21-1059

18. U.S. Food and Drug Administration (FDA), Center for Drug Evaluation and Research . Final Summary Minutes Of the Oncologic Drugs Advisory Committee Meeting, April 12, 2024. FDA; 2024. Accessed October 30, 2024. https://www.fda.gov/media/180108/download

19. Hein K, Short N, Jabbour E, Yilmaz M. Clinical value of measurable residual disease in acute lymphoblastic leukemia. Blood Lymphat Cancer. 2022;12:7‐16. doi: 10.2147/blctt.s270134

20. Jen EY, Xu Q, Schetter A, et al. FDA approval: blinatumomab for patients with B‐cell precursor acute lymphoblastic leukemia in morphologic remission with minimal residual disease. Clin Cancer Res. 2019;25(2):473‐477. doi: 10.1158/1078-0432.ccr-18-2337

21. Giudice ID, Starza ID, Foà R. Does MRD have a role in the management of iNHL? Hematol Am Soc Hematol Educ Program. 2021;2021(1):320‐330. doi: 10.1182/hematology.2021000312

22. Duke–Margolis Center for Health Policy . Minimal residual disease as a surrogate endpoint in hematologic cancer trials. 2016. Accessed October 30, 2024. https://healthpolicy.duke.edu/sites/default/files/2020‐03/9.7.16_discussion%20guide.pdf

23. Buyse M, Molenberghs G, Burzykowski T, Renard D, Geys H. The validation of surrogate endpoints in meta‐analyses of randomized experiments. Biostatistics. 2000;1(1):49‐67. doi: 10.1093/biostatistics/1.1.49

24. Chen X, Wood BL. Monitoring minimal residual disease in acute leukemia: technical challenges and interpretive complexities. Blood Rev. 2017;31(2):63‐75. doi: 10.1016/j.blre.2016.09.006

25. U.S. Food and Drug Administration (FDA) . Hematologic Malignancies: Regulatory Considerations for Use of Minimal Residual Disease in Development of Drug and Biological Products for Treatment. Guidance for Industry. FDA; 2020. Accessed May 13, 2025. https://www.fda.gov/regulatory‐information/search‐fda‐guidance‐documents/hematologic‐malignancies‐regulatory‐considerations‐use‐minimal‐residual‐disease‐development‐drug‐and

26. U.S. Food and Drug Administration (FDA) . Acute Myeloid Leukemia: Developing Drugs and Biological Products for Treatment, Guidance for Industry. FDA; 2022. Accessed May 24, 2025. https://www.fda.gov/regulatory‐information/search‐fda‐guidance‐documents/acute‐myeloid‐leukemia‐developing‐drugs‐and‐biological‐products‐treatment‐0

27. Grimwade D, Freeman SD. Defining minimal residual disease in acute myeloid leukemia: which platforms are ready for “prime time”. Blood. 2014;124(23):3345‐3355. doi: 10.1182/blood-2014-05-577593

28. Loghavi S, DiNardo CD, Takahashi K, et al. Clonal hematopoiesis and Its implications for flow cytometric assessment of measurable residual disease in patients with NPM1‐mutated acute myeloid leukemia [abstract]. Blood. 2020;136(suppl 1):38‐39. doi: 10.1182/blood-2020-139871

29. Wang SA, Jorgensen JL, Hu S, et al. Validation of a 12‐color flow cytometry assay for acute myeloid leukemia minimal/measurable residual disease detection. Cytometry B Clin Cytom. 2023;104(5):356‐366. doi: 10.1002/cyto.b.22140

30. Simoes C, Gonzalez C, Vergez F, et al. Data‐driven flow cytometry classification of blast differentiation in older patients with acute myeloid leukemia. Blood Neoplasia. 2024;1(3):100027. doi: 10.1016/j.bneo.2024.100027

31. Kern W, Westers TM, Bellos F, et al. Multicenter prospective evaluation of diagnostic potential of flow cytometric aberrancies in myelodysplastic syndromes by the ELN iMDS flow working group. Cytometry B Clin Cytom. 2023;104(1):51‐65. doi: 10.1002/cyto.b.22105

32. Tettero JM, Freeman S, Buecklein V, et al. Technical aspects of flow cytometry‐based measurable residual disease quantification in acute myeloid leukemia: experience of the European LeukemiaNet MRD Working Party. Hemasphere. 2022;6(1):e676. doi: 10.1097/hs9.0000000000000676

33. Paiva B, Vidriales MB, Sempere A, et al. Impact of measurable residual disease by decentralized flow cytometry: a PETHEMA real‐world study in 1076 patients with acute myeloid leukemia. Leukemia. 2021;35(8):2358‐2370. doi: 10.1038/s41375-021-01126-3

34. Levis M, Shi W, Chang K, et al. FLT3 inhibitors added to induction therapy induce deeper remissions. Blood. 2020;135(1):75‐78. doi: 10.1182/blood.2019002180

35. Grob T, Sanders MA, Vonk CM, et al. Prognostic value of FLT3‐internal tandem duplication residual disease in acute myeloid leukemia. J Clin Oncol. 2023;41(4):756‐765. doi: 10.1200/jco.22.00715

36. Levis MJ, Hamadani M, Logan B, et al. Gilteritinib as post‐transplant maintenance for acute myeloid leukemia with internal tandem duplication mutation of FLT3. J Clin Oncol. 2024:42(15):1766‐1775. doi: 10.1200/jco.23.02474

37. Dillon LW, Gui G, Ravindra N, et al. Measurable residual FLT3 internal tandem duplication before allogeneic transplant for acute myeloid leukemia. JAMA Oncol. 2024;10(8):1104‐1110. doi: 10.1001/jamaoncol.2024.0985

38. Lambert J, Lambert J, Nibourel O, et al. MRD assessed by WT1 and NPM1 transcript levels identifies distinct outcomes in AML patients and is influenced by gemtuzumab ozogamicin. Oncotarget. 2014;5(15):6280‐6288. doi: 10.18632/oncotarget.2196

39. Kapp‐Schwoerer S, Weber D, Corbacioglu A, et al. Impact of gemtuzumab ozogamicin on MRD and relapse risk in patients with NPM1‐mutated AML: results from the AMLSG 09‐09 trial. Blood. 2020;136(26):3041‐3050. doi: 10.1182/blood.2020005998

40. Stein EM, DiNardo CD, Fathi AT, et al. Ivosidenib or enasidenib combined with intensive chemotherapy in patients with newly diagnosed AML: a phase 1 study. Blood. 2021;137(13):1792‐1803. doi: 10.1182/blood.2020007233

41. Invivoscribe . AML‐FLT2 ITD MRD Assay. Catalog no. FLT3 ITD MRD. Invivoscribe, Inc. 2025. Accessed April 23, 2025. https://catalog.invivoscribe.com/product/flt3‐itd‐mrd‐testing‐by‐ngs /

42. Levis MJ, Perl AE, Altman JK, et al. A next‐generation sequencing‐based assay for minimal residual disease assessment in AML patients with FLT3‐ITD mutations. Blood Adv. 2018;2(8):825‐831. doi: 10.1182/bloodadvances.2018015925

43. Thol F, Gabdoulline R, Liebich A, et al. Measurable residual disease monitoring by NGS before allogeneic hematopoietic cell transplantation in AML. Blood. 2018;132(16):1703‐1713. doi: 10.1182/blood-2018-02-829911

44. Patkar N, Kakirde C, Shaikh AF, et al. Clinical impact of panel‐based error‐corrected next generation sequencing versus flow cytometry to detect measurable residual disease (MRD) in acute myeloid leukemia (AML). Leukemia. 2021;35(5):1392‐1404. doi: 10.1038/s41375-021-01131-6

45. Jongen‐Lavrencic M, Grob T, Hanekamp D, et al. Molecular minimal residual disease in acute myeloid leukemia. N Engl J Med. 2018;378(13):1189‐1199. doi: 10.1056/nejmoa1716863

46. Klco JM, Miller CA, Griffith M, et al. Association between mutation clearance after induction therapy and outcomes in acute myeloid leukemia. JAMA. 2015;314(8):811‐822. doi: 10.1001/jama.2015.9643

47. Dillon LW, Gui G, Page KM, et al. DNA sequencing to detect residual disease in adults with acute myeloid leukemia prior to hematopoietic cell transplant. JAMA. 2023;329(9):745‐755. doi: 10.1001/jama.2023.1363

48. Dillon LW, Higgins J, Nasif H, et al. Quantification of measurable residual disease using duplex sequencing in adults with acute myeloid leukemia. Haematologica. 2024;109(2):401‐410. doi: 10.3324/haematol.2023.283520

49. Kennedy SR, Schmitt MW, Fox EJ, et al. Detecting ultralow‐frequency mutations by duplex sequencing. Nat Protoc. 2014;9(11):2586‐2606. doi: 10.1038/nprot.2014.170

50. Hoffmeister LM, Suttorp J, Walter C, et al. Panel‐based RNA fusion sequencing improves diagnostics of pediatric acute myeloid leukemia. Leukemia. 2024;38(3):538‐544. doi: 10.1038/s41375-023-02102-9

51. Ok CY, Loghavi S, Sui D, et al. Persistent IDH1/2 mutations in remission can predict relapse in patients with acute myeloid leukemia. Haematologica. 2019;104(2):305‐311. doi: 10.3324/haematol.2018.191148

52. Jendro J, Bischof L, Ussmann J, et al. Measurable residual disease detection of IDH1 or IDH2 mutations post but not prior to allogeneic hematopoietic stem cell transplantation show high prognostic relevance in patients with acute myeloid leukemia [abstract]. Blood. 2024;144(suppl 1):4334. doi: 10.1182/blood-2024-207490

53. Willekens C, Blanchet O, Renneville A, et al. Prospective long‐term minimal residual disease monitoring using RQ‐PCR in RUNX1‐RUNX1T1‐positive acute myeloid leukemia: results of the French CBF‐2006 trial. Haematologica. 2016;101(3):328‐335. doi: 10.3324/haematol.2015.131946

54. Ivey A, Hills RK, Simpson MA, et al. Assessment of minimal residual disease in standard‐risk AML. N Engl J Med. 2016;374(5):422‐433. doi: 10.1056/nejmoa1507471

55. Jourdan E, Boissel N, Chevret S, et al. Prospective evaluation of gene mutations and minimal residual disease in patients with core binding factor acute myeloid leukemia. Blood. 2013;121(12):2213‐2223. doi: 10.1182/blood-2012-10-462879

56. Baines AC, Yazdy MS, Kasamon YL, et al. Minimal residual disease data in hematologic malignancy drug applications and labeling: an FDA perspective. Clin Cancer Res. 2023;29(15):2748‐2752. doi: 10.1158/1078-0432.ccr-22-3579

57. Gormley N, Bhatnagar V, Ehrlich LA, et al. FDA analysis of MRD data in hematologic malignancy applications [abstract]. J Clin Oncol. 2017;35(15 suppl):2541. doi: 10.1200/jco.2017.35.15_suppl.2541

58. Terwijn M, van Putten WL, Kelder A, et al. High prognostic impact of flow cytometric minimal residual disease detection in acute myeloid leukemia: data from the HOVON/SAKK AML 42A study. J Clin Oncol. 2013;31(31):3889‐3897. doi: 10.1200/jco.2012.45.9628

59. Freeman SD, Hills RK, Virgo P, et al. Measurable residual disease at induction redefines partial response in acute myeloid leukemia and stratifies outcomes in patients at standard risk without NPM1 mutations. J Clin Oncol. 2018;36(15):1486‐1497. doi: 10.1200/jco.2017.76.3425

60. Venditti A, Piciocchi A, Candoni A, et al. GIMEMA AML1310 trial of risk‐adapted, MRD‐directed therapy for young adults with newly diagnosed acute myeloid leukemia. Blood. 2019;134(12):935‐945. doi: 10.1182/blood.2018886960

61. Freeman SD, Virgo P, Couzens S, et al. Prognostic relevance of treatment response measured by flow cytometric residual disease detection in older patients with acute myeloid leukemia. J Clin Oncol. 2013;31(32):4123‐4131. doi: 10.1200/jco.2013.49.1753

62. Wei AH, Montesinos P, Ivanov V, et al. Venetoclax plus LDAC for newly diagnosed AML ineligible for intensive chemotherapy: a phase 3 randomized placebo‐controlled trial. Blood. 2020;135(24):2137‐2145. doi: 10.1182/blood.2020004856

63. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and venetoclax in previously untreated acute myeloid leukemia. N Engl J Med. 2020;383(7):617‐629. doi: 10.1056/nejmoa2012971

64. Pratz KW, Jonas BA, Pullarkat V, et al. Measurable residual disease response and prognosis in treatment‐naïve acute myeloid leukemia with venetoclax and azacitidine. J Clin Oncol. 2022;40(8):855‐865. doi: 10.1200/jco.21.01546

65. Roboz GJ, Ravandi F, Wei AH, et al. Oral azacitidine prolongs survival of patients with AML in remission independently of measurable residual disease status. Blood. 2022;139(14):2145‐2155. doi: 10.1182/blood.2021013404

66. Freeman SD, Thomas A, Thomas I, et al. Fractionated vs single‐dose gemtuzumab ozogamicin with determinants of benefit in older patients with AML: the UK NCRI AML18 trial. Blood. 2023;142(20):1697‐1707. doi: 10.1182/blood.2023020630

67. Simoes C, Paiva B, Martínez‐Cuadrón D, et al. Measurable residual disease in elderly acute myeloid leukemia: results from the PETHEMA‐FLUGAZA phase 3 clinical trial. Blood Adv. 2021;5(3):760‐770. doi: 10.1182/bloodadvances.2020003195

68. Walter RB, Gooley TA, Wood BL, et al. Impact of pretransplantation minimal residual disease, as detected by multiparametric flow cytometry, on outcome of myeloablative hematopoietic cell transplantation for acute myeloid leukemia. J Clin Oncol. 2011;29(9):1190‐1197. doi: 10.1200/jco.2010.31.8121

69. Getta BM, Devlin SM, Levine RL, et al. Multicolor flow cytometry and multigene next‐generation sequencing are complementary and highly predictive for relapse in acute myeloid leukemia after allogeneic transplantation. Biol Blood Marrow Transplant. 2017;23(7):1064‐1071. doi: 10.1016/j.bbmt.2017.03.017

70. Terao T, Matsuoka K, Ueda H, et al. Gilteritinib maintenance therapy post‐allogenic stem‐cell transplantation improves the prognosis of patients with FLT3‐mutated AML [abstract]. Blood. 2022;140(suppl 1):3290‐3291. doi: 10.1182/blood-2022-163876

71. van Weelderen RE, Klein K, Harrison CJ, et al. Measurable residual disease and fusion partner independently predict survival and relapse risk in childhood KMT2A‐rearranged acute myeloid leukemia: a study by the International Berlin–Frankfurt–Münster Study Group. J Clin Oncol. 2023;41(16):2963‐2974. doi: 10.1200/jco.22.02120

72. Benetton M, Merli P, Walter C, et al. Molecular measurable residual disease assessment before hematopoietic stem cell transplantation in pediatric acute myeloid leukemia patients: a retrospective study by the I‐BFM Study Group. Biomedicines. 2022;10(7):1530. doi: 10.3390/biomedicines10071530

73. Rubnitz JE, Lacayo NJ, Inaba H, et al. Clofarabine can replace anthracyclines and etoposide in remission induction therapy for childhood acute myeloid leukemia: the AML08 multicenter, randomized phase III trial. J Clin Oncol. 2019;37(23):2072‐2081. doi: 10.1200/jco.19.00327

74. Buldini B, Maurer‐Granofszky M, Varotto E, Dworzak MN. Flow‐cytometric monitoring of minimal residual disease in pediatric patients with acute myeloid leukemia: recent advances and future strategies. Front Pediatr. 2019;7:412. doi: 10.3389/fped.2019.00412

75. Tierens A, Bjorklund E, Siitonen S, et al. Residual disease detected by flow cytometry is an independent predictor of survival in childhood acute myeloid leukaemia: results of the NOPHO‐AML 2004 study. Br J Haematol. 2016;174(4):600‐609. doi: 10.1111/bjh.14093

76. Sievers EL, Lange BJ, Alonzo TA, et al. Immunophenotypic evidence of leukemia after induction therapy predicts relapse: results from a prospective Children's Cancer Group study of 252 patients with acute myeloid leukemia. Blood. 2003;101(9):3398‐3406. doi: 10.1182/blood-2002-10-3064

77. Rubnitz JE, Inaba H, Dahl G, et al. Minimal residual disease‐directed therapy for childhood acute myeloid leukaemia: results of the AML02 multicentre trial. Lancet Oncol. 2010;11(6):543‐552. doi: 10.1016/s1470-2045(10)70090-5

78. Keino D, Kinoshita A, Tomizawa D, et al. Residual disease detected by multidimensional flow cytometry shows prognostic significance in childhood acute myeloid leukemia with intermediate cytogenetics and negative FLT3‐ITD: a report from the Tokyo Children's Cancer Study Group. Int J Hematol. 2016;103(4):416‐422. doi: 10.1007/s12185-016-1937-y

79. Jacobsohn DA, Tse WT, Chaleff S, et al. High WT1 gene expression before haematopoietic stem cell transplant in children with acute myeloid leukaemia predicts poor event‐free survival. Br J Haematol. 2009;146(6):669‐674. doi: 10.1111/j.1365-2141.2009.07770.x

80. Zeng HM, Hu GH, Lu AD, Jia YP, Zuo YX, Zhang LP. Predictive impact of residual disease detected using multiparametric flow cytometry on risk stratification of paediatric acute myeloid leukaemia with normal karyotype. Int J Lab Hematol. 2021;43(4):752‐759. doi: 10.1111/ijlh.13570

81. Antoniou E, Hahn S, Sendker S, von Neuhoff N, Reinhardt D, Schneider M. P503: MRD monitoring during intensive chemotherapy in pediatric AML—data of the AML‐BFM group [poster]. Hemasphere. 2022;6:402‐403. doi: 10.1097/01.hs9.0000844900.31768.1d

82. Zhang L, Cao Z, Ruan M, et al. Monitoring the AML1/ETO fusion transcript to predict outcome in childhood acute myeloid leukemia. Pediatr Blood Cancer. 2014;61(10):1761‐1766. doi: 10.1002/pbc.25109

83. van der Velden VH, van der Sluijs‐Geling A, Gibson BE, et al. Clinical significance of flowcytometric minimal residual disease detection in pediatric acute myeloid leukemia patients treated according to the DCOG ANLL97/MRC AML12 protocol. Leukemia. 2010;24(9):1599‐1606. doi: 10.1038/leu.2010.153

84. Coustan‐Smith E, Ribeiro RC, Rubnitz JE, et al. Clinical significance of residual disease during treatment in childhood acute myeloid leukaemia. Br J Haematol. 2003;123(2):243‐252. doi: 10.1046/j.1365-2141.2003.04610.x

85. Buldini B, Rizzati F, Masetti R, et al. Prognostic significance of flow‐cytometry evaluation of minimal residual disease in children with acute myeloid leukaemia treated according to the AIEOP‐AML 2002/01 study protocol. Br J Haematol. 2017;177(1):116‐126. doi: 10.1111/bjh.14523

86. Steinbach D, Bader P, Willasch A, et al. Prospective validation of a new method of monitoring minimal residual disease in childhood acute myelogenous leukemia. Clin Cancer Res. 2015;21(6):1353‐1359. doi: 10.1158/1078-0432.ccr-14-1999

87. Loken MR, Alonzo TA, Pardo L, et al. Residual disease detected by multidimensional flow cytometry signifies high relapse risk in patients with de novo acute myeloid leukemia: a report from Children's Oncology Group. Blood. 2012;120(8):1581‐1588. doi: 10.1182/blood-2012-02-408336

88. Langebrake C, Creutzig U, Dworzak M, et al. Residual disease monitoring in childhood acute myeloid leukemia by multiparameter flow cytometry: the MRD‐AML‐BFM Study Group. J Clin Oncol. 2006;24(22):3686‐3692. doi: 10.1200/jco.2005.05.4312

89. Segerink WH, de Haas V, Kaspers GJL. Measurable residual disease in pediatric acute myeloid leukemia: a systematic review. Expert Rev Anticancer Ther. 2021;21(4):451‐459. doi: 10.1080/14737140.2021.1860763

90. Jacobsohn DA, Loken MR, Fei M, et al. Outcomes of measurable residual disease in pediatric acute myeloid leukemia before and after hematopoietic stem cell transplant: validation of difference from normal flow cytometry with chimerism studies and Wilms tumor 1 gene expression. Biol Blood Marrow Transplant. 2018;24(10):2040‐2046. doi: 10.1016/j.bbmt.2018.06.010

91. Tierens A, Arad‐Cohen N, Cheuk D, et al. Mitoxantrone versus liposomal daunorubicin in induction of pediatric AML with risk stratification based on flow cytometry measurement of residual disease. J Clin Oncol. 2024;42(18):2174‐2185. doi: 10.1200/jco.23.01841

92. Przepiorka D, de Claro RA, Pazdur R. The role of acute myeloid leukemia minimal residual disease in regulatory decision‐making. JAMA Oncol. 2021;7(5):784. doi: 10.1001/jamaoncol.2021.0122

93. FDA‐NIH Biomarker Working Group . BEST (Biomarkers, EndpointS, and other Tools) Resource [Internet]. US Food and Drug Administration; 2021. Accessed October 30. 2024. https://www.ncbi.nlm.nih.gov/books/NBK402286/on

94. U.S. Food and Drug Administration (FDA) . Clinical Trial Considerations to Support Accelerated Approval of Oncology Therapeutics: Guidance for Industry (Draft Guidance). FDA; 2023. Accessed October 30, 2024. https://www.fda.gov/media/166431/download

95. Cooke E; European Medicines Agency . Letter of Support for Measurable Residual Disease (MRD) as a Surrogate Efficacy Endpoint in Clinical Studies with Acute Myeloid Leukemia. EMADOC‐1700519818‐1285629. European Medicines Agency. February 15, 2024. Accessed May 13, 2025. https://www.ema.europa.eu/en/documents/other/letter‐support‐measurable‐residual‐disease‐mrd‐surrogate‐efficacy‐endpoint‐clinical‐studies‐acute‐myeloid‐leukemia_en.pdf

96. Buyse M, Saad ED, Burzykowski T, Regan MM, Sweeney CS. Surrogacy beyond prognosis: the importance of “trial‐level” surrogacy. Oncologist. 2022;27(4):266‐271. doi: 10.1093/oncolo/oyac006

97. Shi Q, Flowers CR, Hiddemann W, et al. Thirty‐month complete response as a surrogate end point in first‐line follicular lymphoma therapy: an individual patient‐level analysis of multiple randomized trials. J Clin Oncol. 2017;35(5):552‐560. doi: 10.1200/jco.2016.70.8651

98. Shi Q, Renfro LA, Bot BM, Burzykowski T, Buyse M, Sargent DJ. Comparative assessment of trial‐level surrogacy measures for candidate time‐to‐event surrogate endpoints in clinical trials. Comput Stat Data Anal. 2011;55(9):2748‐2757. doi: 10.1016/j.csda.2011.03.014

99. Burzykowski T, Molenberghs G, Buyse M. The validation of surrogate end points by using data from randomized clinical trials: a case‐study in advanced colorectal cancer. J R Stat Soc Ser A Stat Soc. 2004;167(1):103‐124. doi: 10.1111/j.1467-985x.2004.00293.x

100. Sargent DJ, Conley BA, Allegra C, Collette L. Clinical trial designs for predictive marker validation in cancer treatment trials. J Clin Oncol. 2005;23(9):2020‐2027. doi: 10.1200/jco.2005.01.112

101. Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424‐447. doi: 10.1182/blood-2016-08-733196

102. Heuser M, Ofran Y, Boissel N, et al. Acute myeloid leukaemia in adult patients: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow‐up. Ann Oncol. 2020;31(6):697‐712. doi: 10.1016/j.annonc.2020.02.018

103. Burzykowski T, Buyse M. Surrogate threshold effect: an alternative measure for meta‐analytic surrogate endpoint validation. Pharm Stat. 2006;5(3):173‐186. doi: 10.1002/pst.207

104. Ciani O, Grigore B, Blommestein H, et al. Validity of surrogate endpoints and their impact on coverage recommendations: a retrospective analysis across international health technology assessment agencies. Med Decis Making. 2021;41(4):439‐452. doi: 10.1177/0272989x21994553

105. Ciani O, Buyse M, Drummond M, Rasi G, Saad ED, Taylor RS. Use of surrogate end points in healthcare policy: a proposal for adoption of a validation framework. Nat Rev Drug Discov. 2016;15(7):516. doi: 10.1038/nrd.2016.81

106. Grigore B, Ciani O, Dams F, et al. Surrogate endpoints in health technology assessment: an international review of methodological guidelines. Pharmacoeconomics. 2020;38(10):1055‐1070. doi: 10.1007/s40273-020-00935-1

107. Ko BS, Wang YF, Li JL, et al. Clinically validated machine learning algorithm for detecting residual diseases with multicolor flow cytometry analysis in acute myeloid leukemia and myelodysplastic syndrome. EBioMedicine. 2018;37:91‐100. doi: 10.1016/j.ebiom.2018.10.042

108. Mocking TR, Kelder A, Reuvekamp T, et al. Computational assessment of measurable residual disease in acute myeloid leukemia using mixture models. Commun Med (Lond). 2024;4(1):271. doi: 10.1038/s43856-024-00700-x

109. McCarthy N, Gui G, Dumezy F, et al. Pre‐emptive detection and evolution of relapse in acute myeloid leukemia by flow cytometric measurable residual disease surveillance. Leukemia. 2024;38(8):1667‐1673. doi: 10.1038/s41375-024-02300-z

110. Gökbuget N, Dombret H, Bonifacio M, et al. Blinatumomab for minimal residual disease in adults with B‐cell precursor acute lymphoblastic leukemia. Blood. 2018;131(14):1522‐1531.

111. Russell N, Burnett A, Hills R, et al. Attenuated arsenic trioxide plus ATRA therapy for newly diagnosed and relapsed APL: long‐term follow‐up of the AML17 trial. Blood. 2018;132(13):1452‐1454. doi: 10.1182/blood-2018-05-851824