Oncofusion

start / graph / concept-mrd-as-an-endpoint-in-kmt2a-r-immunotherapy-and-menin-trials

Explainermoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q34-mrd-monitoring

MRD as a trial endpoint in KMT2A-rearranged leukaemia: high negativity rates, and almost none of them say which assay

also called MRD-negative complete response · MRD endpoint · MRD as surrogate endpoint · MRD negativity rate

New drugs for this leukaemia are increasingly judged by how often they clear leftover disease rather than by how long patients live, because the answer arrives years sooner. The problem is that the published reports usually do not say which leftover-disease test was used or how sensitive it was, so the headline percentages cannot be compared with each other.

The same thing, for a scientist

MRD negativity is now the proximate endpoint for every new agent in this disease. **Tisagenlecleucel** in 35 children under 3 years, 29 of 38 screened with KMT2A-rearranged ALL: MRD-negative complete response with or without haematological recovery in **24 of 28 (86%)** of those with measurable disease; 12-month overall survival 84%, event-free survival 69%, stringent event-free survival 41% [E1]. **Revumenib** phase II in relapsed/refractory KMT2A-rearranged acute leukaemia: overall response 63.2% with **15 of 22 (68.2%)** having no detectable residual disease [E2]; the phase I report states clearance of residual disease using *sensitive clinical assays* alongside hallmarks of differentiation into normal haematopoietic cells [E3]. **Blinatumomab** added to Interfant-06 chemotherapy improved 2-year disease-free survival to 82% against 49% for historical chemotherapy [E4]. Reviewers summarise CD19 CAR-T as inducing MRD-negative remissions in 80-90% of heavily pretreated patients [E5]. **None of the primary reports quoted here states the MRD assay, its analyte, its analytical sensitivity or the tissue** in the abstract. That is not pedantry: the same marrow can be negative by flow and positive by sequencing [E6], the gap between the two reclassified 38.7% of paediatric B-ALL patients [E6], and in this subtype a phenotype-based assay is defeated by the lineage switch that CD19-directed therapy itself provokes.

Go deeper
Why MRD became the endpoint, and what is given up

The argument for it is strong. MRD is the strongest prognostic factor in ALL, it allows risk-group assignment, and its use as a surrogate endpoint would significantly shorten follow-up in trials of new agents - which is the explicit rationale stated by the architects of the standards 7. In a disease affecting a few hundred infants a year worldwide, an endpoint that reads out in months rather than years is close to a precondition for testing anything.

What is given up is comparability, in three layers.

  1. Assay. Flow, ASO-PCR, fusion transcript, genomic breakpoint and sequencing are five different measurements with different failure modes. An 86% MRD-negativity rate by flow and a 68% by an unnamed assay are not on the same scale.
  2. Timepoint. In frontline infant therapy, MRD negativity rises from 19.4% at end of induction to 63.1% after MARMA on the same patients 8. A relapsed-setting response rate at an unspecified post-treatment timepoint is not comparable with either.
  3. Denominator. Revumenib's 15 of 22 is among responders, from 57 efficacy-evaluable and 94 treated patients 2; tisagenlecleucel's 24 of 28 is among those with measurable disease, from 35 infused and 38 screened 1. Both are legitimate and neither is a rate per treated patient.
The subtype-specific reason this matters more here

Two of the three drug classes above create the exact conditions under which a phenotype-based MRD assay fails.

CD19-directed therapy provokes lineage switch. Seven KMT2A-rearranged patients treated with CD19 CAR-T all achieved marrow complete remission by flow cytometry, and two developed clonally related AML within a month 9. An MRD-negativity rate reported by flow in a CD19-directed trial in this subtype is measuring the one thing that is guaranteed to look clear.

Menin inhibition produces differentiation. The phase I report pairs clearance of residual disease with hallmarks of differentiation into normal haematopoietic cells, including differentiation syndrome, in the same sentence 3, and grade 3 or higher differentiation syndrome occurred in 16.0% of phase II patients 2. A flow MRD panel discriminates blasts from normal maturing cells by maturation pattern; a drug whose mechanism includes maturation attacks that discriminant. Whether it does so in practice has not been measured, because no menin-inhibitor trial has reported paired phenotype-based and lineage-independent molecular MRD.

The fix is cheap and specific. In KMT2A-rearranged disease a lineage-independent molecular MRD target - the genomic fusion junction - exists for essentially every patient 10. Reporting it alongside whatever the trial's primary MRD assay is would cost one qPCR per timepoint and would make the numbers comparable across drug classes and across trials. No trial in this disease currently does this.

Facts and figures
proposed reporting standardreport a lineage-independent patient-specific genomic fusion junction MRD alongside the trial's primary MRD assay; feasible because a junction target is obtainable in essentially every KMT2A-rearranged patient
currently done byno trial in this disease
reported mrd negativity
[
  {
    "agent": "tisagenlecleucel",
    "setting": "relapsed/refractory, children <3 years, 76% KMT2A-r",
    "rate": "24/28 (86%) of those with measurable disease",
    "assay_stated": false,
    "source_pmid": "36084658"
  },
  {
    "agent": "revumenib",
    "setting": "relapsed/refractory KMT2A-r acute leukaemia, phase II",
    "rate": "15/22 (68.2%) with no detectable residual disease",
    "assay_stated": false,
    "source_pmid": "39121437"
  },
  {
    "agent": "revumenib",
    "setting": "relapsed/refractory acute leukaemia, phase I",
    "rate": "clearance of residual disease using 'sensitive clinical assays'",
    "assay_stated": false,
    "source_pmid": "36922593"
  },
  {
    "agent": "CD19 CAR T (class summary)",
    "setting": "heavily pretreated paediatric BCP-ALL",
    "rate": "MRD-negative remission in 80-90%",
    "assay_stated": false,
    "source_pmid": "42489798"
  }
]
frontline comparator by timepoint PCT
{
  "end_of_induction": 19.4,
  "end_of_consolidation": 40.4,
  "after_MARMA": 63.1,
  "source_pmid": "33405950"
}
why assay identity changes the number
{
  "flow_vs_sequencing_reclassified_upward_pct": 38.7,
  "source_pmid": "29284596"
}
subtype specific hazards
[
  "CD19-directed therapy provokes lineage switch, which a B-lineage flow assay scores negative",
  "menin inhibition produces differentiation, which erodes the blast-versus-normal discriminant flow relies on"
]
How this could bite us back
  • MRD negativity as an endpoint rewards drugs that clear the compartment the assay samples. A drug that clears blood faster than marrow, or that matures cells rather than killing them, scores well on the wrong mechanism, and both behaviours are documented in this drug class.
  • An 86% MRD-negative response rate in infants after CAR-T sits alongside a 41% stringent event-free survival at 12 months in the same cohort. The endpoint and the outcome disagree by a factor of two, which is the clearest available warning against treating MRD negativity as the result.
  • Quoting a responder-subgroup MRD-negativity rate (68.2%) without the complete-remission rate it is drawn from (22.8%) is the standard way this number is presented, and it inflates the apparent effect roughly threefold.
  • If MRD becomes a registrational surrogate in a disease where the commonest escape mechanism defeats the commonest MRD assay, approval will be granted on a measurement that the resistance mechanism is specifically invisible to.
  • The remedy - report a genomic-junction MRD alongside the primary assay - requires shipping DNA to one of a small number of breakpoint-characterisation laboratories, so a reporting standard that improves comparability would also make trial participation harder for centres outside that network.
Still open
  1. What assay produced the AUGMENT-101 no-detectable-residual-disease figure? It is among the most quoted numbers in this disease and the primary report's abstract does not name it.
  2. In a menin-inhibitor trial, would a genomic-junction MRD and a flow MRD on the same marrow agree? The preclinical ctDNA work found compartments diverging under this drug class; the clinical version of that experiment has not been done.
  3. Should a lineage-independent molecular MRD be a required secondary endpoint in KMT2A-rearranged trials? The target is obtainable in essentially every patient, and without it the field cannot compare MRD-negativity rates between drug classes.
  4. In the infant tisagenlecleucel cohort, was a myeloid panel applied at MRD timepoints? No lineage switch is reported and the paper does not state whether the assay could have detected one.
Every source behind this entry (10)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In an international retrospective cohort of 38 children under 3 years screened for tisagenlecleucel across 15 hospitals in ten European countries, of whom 29 (76%) had KMT2A-rearranged ALL and 35 (92%) received an infusion, measurable residual disease-negative complete response with or without haematological recovery occurred in 24 of 28 patients (86%) who had measurable disease, with 12-month overall survival 84%, event-free survival 69% and stringent event-free survival 41%.

    PMID 36084658DOI
    What this evidence cannot show

    Retrospective real-world cohort with a standardised reporting form rather than a prospective trial; MRD assay, sensitivity and tissue are not stated in the abstract and will have varied across 15 centres in ten countries. The 24-of-28 denominator is patients with measurable disease, not the 35 infused or the 38 screened. Median follow-up 14 months, and the gap between 69% event-free and 41% stringent event-free survival shows how much the outcome definition moves the number.

    Evidence that argues against this (1)
    • In this same population CD19-directed therapy is the treatment most strongly associated with lineage switch, and a lineage-switched clone would be scored MRD-negative by a B-lineage assay; the paper reports no lineage-switch events but also does not state that a myeloid panel was applied. Gardner R et al. 2016
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the phase II registrational portion of AUGMENT-101, 94 patients with relapsed or refractory KMT2A-rearranged acute leukaemia received revumenib; among 57 efficacy-evaluable patients the complete remission plus complete remission with partial haematologic recovery rate was 22.8% and overall response rate 63.2%, with 15 of 22 patients (68.2%) having no detectable residual disease; grade 3 or higher differentiation syndrome occurred in 16.0%.

    PMID 39121437DOI
    What this evidence cannot show

    The abstract does not state the MRD assay, its sensitivity, the tissue or how the 22-patient denominator relates to the 57 efficacy-evaluable and 94 treated patients. Median age 37 years with a range from 1.3 years, so this is predominantly an adult cohort and its MRD practice is adult practice.

    Evidence that argues against this (1)
    • The complete remission plus CRh rate was 22.8% while the no-detectable-residual-disease figure is 68.2% of a responder subgroup; quoting the second without the first, which is common, substantially overstates what the drug achieves per treated patient. Issa GC et al. 2025
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The first-in-human phase I trial of revumenib in relapsed or refractory acute leukaemia reported a 30% rate of complete remission or complete remission with partial haematologic recovery in the efficacy analysis population, demonstrated clearance of residual disease using sensitive clinical assays, and identified hallmarks of differentiation into normal haematopoietic cells including differentiation syndrome.

    The menin inhibitor revumenib in KMT2A-rearranged or NPM1-mutant leukaemia. · Issa GC, Aldoss I, DiPersio J, et al. · Nature 2023
    PMID 36922593DOI
    What this evidence cannot show

    The phrase sensitive clinical assays is plural and unnamed in the abstract; the assays may be specified in the full text, which this beat did not retrieve. The trial enrolled both KMT2A-rearranged and NPM1-mutant leukaemias, and differentiation is a myeloid phenomenon so its relevance to infant B-lineage disease is not established here.

    Evidence that argues against this (1)
    • Differentiation and residual-disease clearance are reported in the same sentence without separation, so the founding evidence cannot distinguish cells cleared from cells matured; equally, the co-occurrence does not show the MRD clearance was artefactual. Issa GC et al. 2023
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    A single course of sequential blinatumomab added to Interfant-06 chemotherapy improved 2-year disease-free survival to 82% compared with 49% from the historical chemotherapy-only approach in infant KMT2A-rearranged B-ALL, and inotuzumab ozogamicin shows high induction response and MRD-negativity rates in relapsed disease while carrying a risk of sinusoidal obstruction syndrome.

    PMID 40584390DOI
    What this evidence cannot show

    A narrative review reporting a single-arm result against a historical comparator at second hand. The randomised test is Interfant-21, which has not reported. Reviews are used in this graph as maps rather than as primary evidence.

    Evidence that argues against this (1)
    • The historical comparator is itself heterogeneous: Interfant-06 event-free survival was 10-12 percentage points higher in West European and North American groups than elsewhere, so a single-arm comparison against a pooled historical number is compared against a figure describing no actual population. Pieters R et al. 2019
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    CD19-directed CAR T-cell therapy induces MRD-negative remissions in 80-90% of heavily pretreated patients with paediatric B-cell precursor ALL, with durable survival in 40-50% without mandatory haematopoietic stem cell transplant consolidation, and randomised paediatric trials show blinatumomab improves survival when incorporated into frontline and first-relapse therapy.

    PMID 42489798DOI
    What this evidence cannot show

    A narrative review summarising ranges across trials in paediatric BCP-ALL generally, not KMT2A-rearranged disease specifically, and without stating the MRD assays behind the 80-90% figure. Used here to show how MRD-negativity rates are quoted in the field, not as a primary source for the rate.

    Evidence that argues against this (1)
    • The same review names antigen escape among the resistance mechanisms these therapies face, and in KMT2A-rearranged disease the dominant form of escape is lineage switch, which a B-lineage MRD assay reports as a negative. Gardner R et al. 2016
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    High-throughput sequencing of IGH and TRG identified 55 (38.7%) more patients as MRD-positive than flow cytometry at a 0.01% threshold across 619 paired paediatric B-ALL marrow samples, and those discrepant patients had worse outcomes than flow-MRD-negative patients.

    PMID 29284596DOI
    What this evidence cannot show

    Paediatric B-ALL, no KMT2A-rearranged or infant subgroup, and one of the two source trials selected patients on their flow MRD result. Cited to establish that assay identity changes the reported negativity rate, not to make a claim about infants.

    Evidence that argues against this (1)
    • Five-year survival estimates were similar by both methods at the same threshold, so assay substitution changed classification without changing average prognostic performance - which weakens the argument that any one assay is the correct one for an endpoint. Wood B et al. 2018
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    MRD diagnostics is the strongest prognostic factor in ALL, guides treatment decisions in frontline, relapsed and transplant settings, is used to evaluate the efficacy of innovative drugs including antibodies and small molecules within clinical trials, and might well be used as a surrogate endpoint thereby significantly shortening follow-up; the techniques need to be sensitive to 10^-4 or better, broadly applicable, accurate, reliable, fast and affordable.

    PMID 25999452DOI
    What this evidence cannot show

    A review by the developers of the EuroFlow and EuroMRD standards; the surrogate-endpoint statement is a prospective argument made in 2015 rather than a regulatory fact. It describes ALL generally.

    Evidence that argues against this (1)
    • The same review states that classical flow does not reach a solid 10^-4 while classical ASO-PCR is slow and labour intensive, so by the authors' own criteria neither established technique fully met the standard they set for a surrogate endpoint at the time they proposed it. van Dongen JJ et al. 2015
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In Interfant-06, the proportion of KMT2A-rearranged infants who were MRD-negative rose from 42 of 217 (19.4%) at end of induction to 72 of 178 (40.4%) at end of consolidation and 106 of 168 (63.1%) after MARMA.

    PMID 33405950DOI
    What this evidence cannot show

    Read in full text. The three denominators are different sets of patients because sampling was not mandatory, so the rise is not strictly within-patient. It is used here to show the dependence of an MRD-negativity rate on timepoint, which does not require a closed cohort.

    Evidence that argues against this (1)
    • Patients who converted from positive at end of induction to negative at end of consolidation had almost the same outcome as those negative at both timepoints (65.7% versus 72.0%), so a later negativity rate is not simply a diluted version of an earlier one and carries real prognostic content. Stutterheim J et al. 2021
  9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Seven patients with MLL-rearranged B-ALL treated with CD19-specific CAR T cells all achieved complete remission in the bone marrow by flow cytometry, and within one month of infusion two developed acute myeloid leukaemia clonally related to their B-ALL.

    PMID 26907630DOI
    What this evidence cannot show

    Seven patients and two events; establishes that the failure mode exists rather than its frequency. The complete remissions were assessed by the same flow assay whose blind spot the cases demonstrate.

    Evidence that argues against this (1)
    • In the infant tisagenlecleucel cohort of 35 infused patients, 76% with KMT2A rearrangement, no lineage switch is reported, so the phenomenon is not universal even in the highest-risk population for it. Ghorashian S et al. 2022
  10. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Across 3,401 acute leukaemia patients, at least one patient-specific KMT2A genomic fusion sequence was obtained in every case by long-distance inverse PCR, long-distance multiplex PCR or targeted NGS and used to establish patient-specific qPCR assays for MRD monitoring, with discovery rates of about 95% for LD-PCR and nearly 100% for targeted NGS.

    The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023
    PMID 37019990DOI
    What this evidence cannot show

    Read in full text. A referral-based series at one specialist centre; the near-universal availability of a target is a property of that workflow, not of routine practice everywhere, and adding this assay to a multinational trial would require the same referral logistics.

    Evidence that argues against this (1)
    • Interfant-06 named two centres, Frankfurt and Catania, as the source of breakpoint sequences, so a recommendation to report a genomic-junction MRD alongside every trial endpoint concentrates a global requirement on a small number of laboratories. Stutterheim J et al. 2021
Connected to