AUGMENT-101: the partner composition of the trial that approved a menin inhibitor, with KMT2A::AFF1 at n=2 evaluable
also called AUGMENT-101 translocation distribution · revumenib registration cohort partner mix
The trial that got the first menin inhibitor approved published which partner gene each patient had. Only two of the fifty-seven patients whose response counted had the partner this project's whole mechanistic story is built around, and a quarter had no partner identified at all.
The same thing, for a scientist
The registration-enabling phase 2 KMT2A-rearranged portion of AUGMENT-101 (NCT04065399) treated 94 patients and assessed efficacy in 57. Table 1 of the primary report gives the translocation distribution for both populations. Efficacy population (n=57): t(9;11)(p22;q23) 11 (19.3%), t(11;19)(q23;p13) 13 (22.8%), t(10;11)(p12;q23) 7 (12.3%), t(6;11)(q27;q23) 7 (12.3%), **t(4;11)(q21;q23) 2 (3.5%)**, t(1;11) 2 (3.5%), other translocations 2 (3.5%), fusion partner unknown 13 (22.8%). Safety population (n=94): t(9;11) 21 (22.3%), t(11;19) 17 (18.1%), t(10;11) 12 (12.8%), t(6;11) 8 (8.5%), t(4;11) 7 (7.4%), t(1;11) 2 (2.1%), other 3 (3.2%), unknown 24 (25.5%). The primary endpoint was CR+CRh 22.8% (95% CI 12.7-35.8) in the efficacy population; the text states that an exploratory analysis with limited sample sizes demonstrated efficacy across various translocation partners and refers the numbers to a data supplement table, giving none in the body. The cohort was 83.0% AML and 14.9% ALL.
The number that matters for this graph
Two. Two of the 57 efficacy-evaluable patients carried t(4;11)/KMT2A::AFF1 1. Seven of 94 in the safety population. This graph's central bet is the menin-KMT2A interface in KMT2A::AFF1 disease; the trial that established clinical proof of that interface contains two evaluable AFF1 patients. Whatever the approval establishes, it does not establish an AFF1-specific response rate, and no arithmetic on 2 patients can.
A quarter of patients have no partner at all
13 of 57 (22.8%) and 24 of 94 (25.5%) are recorded as fusion partner unknown 1. Eligibility was based on local testing with central confirmation by FISH split signal, which establishes that KMT2A is rearranged and identifies nothing about what it is rearranged with. So the largest single category in the efficacy population, after t(11;19), is the category with no partner information. Any partner-stratified reanalysis of this trial starts by discarding a quarter of the data.
t(11;19) is the largest arm and is band-ambiguous
t(11;19)(q23;p13) at 13/57 is the biggest identified arm, and the sub-band is not given - p13.1 would be KMT2A::ELL and p13.3 KMT2A::MLLT1, partners that sit on opposite sides of the paediatric AML risk classification 2. The largest arm in the registration trial is therefore a pooling of two partners that this literature does not regard as equivalent. See concept-cytogenetic-band-versus-molecular-partner-misassignment.
What the trial actually says about partners
The body text reads: an exploratory analysis with limited sample sizes demonstrated efficacy for KMT2A-rearranged disease across various translocation partners, with the numbers in Data Supplement Table S3 1. That is an assertion of no-difference with the evidence placed outside the article. The PMC record for this paper carries pmc-prop-has-supplement: no, so the supplement is not retrievable through that route and Table S3 was not read for this entry. The class claim for menin inhibition across partners currently rests, in the primary literature, on a sentence.
| metric | KMT2A translocation distribution in the AUGMENT-101 registration cohort |
|---|---|
| value | 2 |
| unit | patients with t(4;11)/KMT2A::AFF1 in the efficacy population (3.5% of 57) |
| population | relapsed/refractory KMT2A-rearranged acute leukaemia, aged 30 days and older |
| nct | NCT04065399 |
| partner confirmation method | local testing with central FISH split signal (identifies rearrangement, not partner) |
| partner response data location | Data Supplement Table S3, not carried in the PMC record, not read |
| region | 22 sites in five countries |
| year | 2025 |
| prespecified | partner subgroup analysis described by the authors as exploratory with limited sample sizes |
| source | PMID 39121437 |
efficacy population — the numbers behind the chart
{
"n": 57,
"t(9;11)": 11,
"t(11;19)": 13,
"t(10;11)": 7,
"t(6;11)": 7,
"t(4;11)": 2,
"t(1;11)": 2,
"other": 2,
"partner_unknown": 13
}safety population — the numbers behind the chart
{
"n": 94,
"t(9;11)": 21,
"t(11;19)": 17,
"t(10;11)": 12,
"t(6;11)": 8,
"t(4;11)": 7,
"t(1;11)": 2,
"other": 3,
"partner_unknown": 24
}lineage mix — the numbers behind the chart
{
"AML_pct": 83,
"ALL_pct": 14.9,
"ambiguous_pct": 2.1
}primary endpoint
{
"cr_crh_pct": 22.8,
"ci": "12.7-35.8",
"orr_pct": 63.2
}- A drug approved for KMT2A-rearranged acute leukaemia will be prescribed to KMT2A::AFF1 infants on the strength of a trial containing two evaluable AFF1 patients; the class label is doing work the data cannot support for that specific fusion.
- Central confirmation by FISH split signal makes partner-blind eligibility structurally unavoidable, so every menin trial run this way will generate a quarter of its data with no partner attached - the reporting gap is designed in, not accidental.
- If response does differ by partner, the arms in this trial are 2, 7, 11 and 13 patients, so the trial that defines the class claim is permanently unable to test it.
- What does Data Supplement Table S3 actually show? Retrieving it is the single highest-value open action in this beat and requires only publisher access.
- Were the 13 t(11;19) patients sub-banded to 19p13.1 versus 19p13.3, and were the 7 t(10;11) patients resolved to MLLT10 versus ABI1?
- Can the 13 unknown-partner patients be resolved retrospectively by RNA sequencing of banked material, and did they respond differently?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In AUGMENT-101's registration-enabling phase 2 KMT2A-rearranged portion, the efficacy population (n=57) comprised t(9;11) 11 (19.3%), t(11;19) 13 (22.8%), t(10;11) 7 (12.3%), t(6;11) 7 (12.3%), t(4;11) 2 (3.5%), t(1;11) 2 (3.5%), other 2 (3.5%) and fusion partner unknown 13 (22.8%); the safety population (n=94) comprised t(9;11) 21, t(11;19) 17, t(10;11) 12, t(6;11) 8, t(4;11) 7, t(1;11) 2, other 3 and unknown 24; CR+CRh was 22.8% (95% CI 12.7-35.8) and efficacy by translocation partner was reported only as an exploratory analysis referred to a data supplement.
Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101). · Issa GC, Aldoss I, et al. · J Clin Oncol 2025What this evidence cannot show
Read from the PMC full text of the primary report (PMC11687943, 63,271 characters), Table 1 and Results/Discussion body. The per-partner response numbers are in Data Supplement Table S3, which is not carried in the PMC record (pmc-prop-has-supplement: no) and was not read; no per-partner response rate is asserted in this entry. Partner assignment rests on local testing with central FISH split-signal confirmation, which does not identify a partner, hence the 22.8-25.5% unknown category. t(11;19) is given without sub-band. Heavily pretreated relapsed/refractory population (median 2 prior therapies, 50% prior HSCT, ~65% prior venetoclax), so the partner mix reflects who survives to a salvage trial, not the disease at diagnosis.
Evidence that argues against this (2)
- The partner mix here is not the partner mix of the disease: in 629 children with KMT2A-r ALL, t(4;11)/KMT2A::AFF1 accounted for 51.5% of cases, against 3.5% of AUGMENT-101's efficacy population, because the trial population is 83% AML and relapsed/refractory. Attarbaschi A et al. 2023
- In 465 KMT2A-rearranged B-ALL cases across the age range, AFF1 was the commonest partner at 69%, again far above its representation in this trial, confirming that the registration cohort under-represents the fusion this graph's mechanism is built on. Hartmann AM et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The two genes producing a t(11;19) karyotype are prognostically distinct in this literature - KMT2A::MLLT1 (19p13.3) is in the I-BFM high-risk group in childhood AML while KMT2A::ELL (19p13.1) is not - so an unresolved t(11;19) arm pools partners the field treats differently.
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. · van Weelderen RE, Klein K, et al. · J Clin Oncol 2023What this evidence cannot show
The risk classification is from childhood AML while AUGMENT-101 enrolled mostly adults; the relevance is to arm composition rather than to a claim that the prognostic difference carries into the relapsed/refractory salvage setting.
Evidence that argues against this (1)
- Two large adult KMT2A-r AML cohorts found no partner effect on survival at all, so in an adult-majority trial the pooling of 19p13.1 with 19p13.3 may make no practical difference to observed response. Issa GC et al. 2021
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
AUGMENT-101 is registered as NCT04065399, an open-label phase 1/2 study of revumenib in relapsed or refractory acute leukaemias including KMT2A-rearranged disease.
Study of SNDX-5613 in R/R Leukemias Including Those With an MLL/KMT2A Gene Rearrangement or NPM1 Mutation (AUGMENT-101) · Syndax Pharmaceuticals · ClinicalTrials.gov 2024What this evidence cannot show
Registry records state planned design and are updated by the sponsor; enrolment figures and arm definitions in the registry may differ from those in the published report. The registry does not report outcomes by fusion partner.
Evidence that argues against this (1)
- The registry entry, like the publication, reports eligibility at the level of KMT2A rearrangement and does not define partner-based arms, so the absence of partner-stratified reporting is built into the protocol rather than being an omission at publication. Issa GC et al. 2025
- ← related toOne karyotype band, two partner genes: why t(11;19) and t(10;11) can cancel a real partner effect
the registration trial's largest identified arm is an unresolved t(11;19)
the approved menin inhibitor's registration cohort contained two evaluable KMT2A::AFF1 patients
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
KMT2A::AFF1 is 3.5% of the efficacy population and 7.4% of the safety population
central confirmation by FISH split signal leaves 22.8-25.5% with fusion partner unknown
the nearest available partner-response data sit in a supplement not carried in PMC