Landscape
Menin inhibitors at ASH 2026: which trial cohorts can you actually compare?
Two drugs approved on CR + CRh, a combination programme that counts CR alone, and two first-line Phase 3 pairs that look identical until you read the endpoint text. Every claim links to the registry record or FDA summary behind it.
Two menin inhibitors are on the US market, a third has two Phase 3 trials running, and the 68th ASH Annual Meeting (New Orleans and virtual, 12 to 15 December 2026, abstracts online since 4 November) will put a new wave of cohort data in front of the field. The natural move is to line the response rates up. This page sets out which cohorts share enough of a definition to be read side by side, and which do not, from the registry record and the FDA's own approval summaries.
On 17 November 2026 we read every ClinicalTrials.gov record that lists revumenib, ziftomenib, bleximenib, enzomenib, icovamenib or BN104 in a leukaemia or myelodysplastic indication: 50 interventional studies (any status except withdrawn) and 116 treatment arms that give a menin inhibitor. Each arm was read as a cohort, by disease, genotype, line of therapy and backbone, next to the trial's primary endpoint as the record words it.
The field, at a glance
| Asset | Industry sponsor | US regulatory and clinical status | Registry records read |
|---|---|---|---|
| Revumenib (Revuforj) | Syndax Pharmaceuticals | Approved: relapsed or refractory acute leukaemia with a KMT2A translocation (15 November 2024); relapsed or refractory AML with a susceptible NPM1 mutation (24 October 2025); adults and children from 1 year | 24 (6 with Syndax as lead sponsor) |
| Ziftomenib (Komzifti) | Kura Oncology | Approved: adults with relapsed or refractory AML with a susceptible NPM1 mutation (13 November 2025) | 15 (5 Kura, 1 Kyowa Kirin) |
| Bleximenib (JNJ-75276617) | Janssen Research & Development | Phase 3: cAMeLot-2 (NCT06852222) and HOVON 181 (NCT07223814) | 6 (4 Janssen) |
| Enzomenib (DSP-5336) | Sumitomo Pharma America | Phase 1/2: Horizen-1 (NCT04988555) | 2 |
| BN104 | Servier (I.R.I.S.) for the Phase 1/2; two investigator studies in China | Phase 1/2 (NCT06052813) | 3 |
| Icovamenib (BMF-219) | Biomea Fusion | The one leukaemia study on record (NCT05153330) is terminated | 1 |
Both approvals rest on CR + CRh
For ziftomenib, the FDA states that efficacy was established on the rate of complete remission plus CR with partial haematological recovery in KO-MEN-001 (NCT04067336), an open-label single-arm trial in 112 adults with relapsed or refractory NPM1-mutant AML: CR + CRh 21.4% (95% CI 14.2–30.2), CR 17.0%, CRh 4.5%, median duration of CR + CRh 5 months, median follow-up 4.2 months.
For revumenib in NPM1-mutant AML, the FDA lists CR, CRh and CR + CRh duration as the main efficacy outcome measures in a single-arm cohort of AUGMENT-101 (NCT04065399): CR + CRh 23.1% (95% CI 13.5–35.2), median duration 4.5 months.
Two points apart, with confidence intervals that overlap almost entirely, from two different single-arm trials with different follow-up, under an indication that for revumenib extends to children from one year of age. These are the reference numbers every new NPM1 cohort at ASH will be held against, informally, in hallway conversations and slide decks. That is the reason the definition behind each new number matters more than the number: a cohort reported on CR alone will look worse against 21–23% than it is, and a cohort reported on a broader composite will look better.
One class, different response definitions
The registry records of the main programmes do not count the same thing as their primary efficacy endpoint:
| Trial | Asset and setting | Primary efficacy endpoint (registry wording) | Response criteria named |
|---|---|---|---|
| KO-MEN-001 | Ziftomenib, single agent | “Evidence of anti-leukemia activity”, assessed by the CR + CRh rate (Phase 1a, 1b and 2) | None named |
| AUGMENT-101 | Revumenib, single agent | “CR+CRh rate (Phase 2 [Cohorts 2A-2C])” | None named |
| Horizen-1 | Enzomenib, alone and in combination | “complete response (CR) and complete response with partial hematologic recovery (CRh) in Phase 2” | “FDA guidance and ELN2017” |
| cAMeLot-1 | Bleximenib, single agent | “Rate of Complete Remission or Complete Remission with Partial Hematologic Recovery (CR/CRh)” (Phase 2) | None named |
| NCT05735184 | Ziftomenib with venetoclax/azacitidine, 7+3 or quizartinib | “Complete remission (CR) rate” | ELN 2022 |
The combination study NCT05735184, the one that carries ziftomenib with venetoclax and azacitidine, with 7+3 and with quizartinib across 13 arms, counts CR alone by ELN 2022. Its secondary outcomes list composite complete remission and morphologic leukaemia-free state, also by ELN 2022; CRh is not named as an outcome in the record. The single-agent programmes count CR + CRh.
A CR rate from the ziftomenib combination study and a CR + CRh rate from a single-agent cohort are different quantities, before any question of population or line. Anyone building a slide that puts "menin inhibitor + ven/aza" next to "menin inhibitor alone" needs the CR component of the single-agent number, which the FDA summary gives for ziftomenib (17.0%) and which the registry does not give for anyone. The cleanest comparisons at ASH will be the ones where a sponsor reports CR and CRh separately.
Cohorts the registry lets you read side by side
Across the 116 arms, only a handful of groupings share disease, genotype, line, backbone and the primary endpoint definition across more than one trial. These are the two that hold up on a reading of the records:
| Population and backbone | Arms | Primary endpoint | Still to check before comparing |
|---|---|---|---|
| Relapsed or refractory NPM1-mutant AML, single agent | KO-MEN-001 Phase 2 cohort (ziftomenib); Horizen-1 Arm H (enzomenib); NCT06052813 Phase II cohort A (BN104); NCT07623616 (ziftomenib, Kyowa Kirin, 6 planned) | CR + CRh in all four records | Only Horizen-1 names a response-criteria edition. Prior venetoclax or menin-inhibitor exposure is not stated at arm level in any of them. |
| Untreated KMT2A-rearranged AML, with venetoclax and azacitidine | NCT05735184 cohort B-3 (ziftomenib, Phase 1 expansion); RAVEN NCT07605949 (revumenib, Phase 2, 88 planned, not yet recruiting) | CR in both | ELN 2022 in the first; “the protocol Response Criteria” in RAVEN. Phase 1 expansion against Phase 2. |
AUGMENT-101 is not in the first row for a structural reason: its record registers a single arm for every cohort, so the registry on its own cannot isolate the NPM1 cohort. The FDA summary above is where that cohort's numbers are stated.
A third grouping looks comparable and is not. KO-MEN-001 Sub-study 3 and Horizen-1 Arm I both give a single agent in relapsed or refractory KMT2A-rearranged ALL. In the records, Sub-study 3's primary outcome is a dose finding (minimum biologically effective dose or recommended Phase 2 dose) assessed by CR, while Horizen-1's Phase 2 primary is CR + CRh.
The first-line Phase 3s: alike in design, different in the fine print
Intensive chemotherapy: KO-MEN-017 against HOVON 181
| KO-MEN-017, intensive study | HOVON 181 | |
|---|---|---|
| Sponsor | Kura Oncology | HOVON, with the German-Austrian AML Study Group |
| Population | Untreated NPM1-mutated or KMT2A-rearranged AML | Newly diagnosed KMT2A-rearranged or NPM1-mutant AML eligible for intensive chemotherapy |
| Arms | Ziftomenib + 7+3, then ziftomenib maintenance; ziftomenib + 7+3, then placebo maintenance; placebo throughout | Bleximenib + induction and consolidation, then bleximenib maintenance; bleximenib, then placebo maintenance; placebo throughout. Daunorubicin or idarubicin. |
| Primary endpoint, all countries | EFS: randomisation to treatment failure, haematologic relapse following CR, or death | EFS: randomisation to failure to achieve CR after remission induction, haematologic relapse after CR, or death |
| Additional primary endpoint | US and US-reference countries only: CR with central bone-marrow MRD negativity in NPM1-mutant patients, ELN 2022, investigator-assessed | None |
| Masking | Quadruple | Double |
| Planned enrolment | 1,300 across both KO-MEN-017 studies | 875 |
| Start / primary completion in the record | 26 September 2025 / November 2031 | 1 March 2026 / June 2030 |
Strip the drug names and these are the same experiment: a menin inhibitor added to intensive induction, a maintenance randomisation, a placebo arm, and event-free survival as the endpoint. That makes the definitions the whole story for anyone reading one against the other when they report, around 2030–2031 on the records' dates. HOVON 181 writes its failure event as no CR after induction, which ties its EFS to the protocol's CR definition and induction window; KO-MEN-017's record says "treatment failure" without defining it. And only KO-MEN-017 carries a second, US-only primary endpoint on MRD-negative CR in NPM1-mutant patients, a structure consistent with seeking a US read earlier than an EFS event count would allow.
Non-intensive: KO-MEN-017 against cAMeLot-2
| KO-MEN-017, non-intensive study | cAMeLot-2 | |
|---|---|---|
| Sponsor | Kura Oncology | Janssen Research & Development |
| Comparison | Ziftomenib or placebo, with venetoclax + azacitidine | Bleximenib or placebo, with venetoclax + azacitidine |
| Population | Untreated NPM1-mutated or KMT2A-rearranged AML | Newly diagnosed KMT2A-rearranged or NPM1-mutant AML, ineligible for intensive chemotherapy |
| Primary endpoints | OS in all countries; CR per ELN 2022, investigator-assessed, in US and US-reference countries only | CR (defined in the record: marrow blasts < 5%, no circulating blasts, no extramedullary disease, ANC ≥ 1.0 × 10⁹/L, platelets ≥ 100 × 10⁹/L) and OS |
| Masking | Quadruple | Triple |
| Start / primary completion in the record | 26 September 2025 / November 2031 | 4 June 2025 / 22 June 2029 |
Both test a menin inhibitor against placebo on a venetoclax and azacitidine backbone, and both pool NPM1-mutant and KMT2A-rearranged patients. cAMeLot-2 makes CR a primary endpoint everywhere and writes out a full-recovery CR definition in the record; KO-MEN-017 makes CR primary only in the US and US-reference countries, per ELN 2022, investigator-assessed.
What the registry cannot tell you
Stated plainly rather than filled in:
- Prior therapy is not read here. Number of prior lines, prior venetoclax, prior menin inhibitor, prior transplant and age group sit in eligibility criteria and protocols, and they can separate two cohorts that share every field above.
- Pooled genotypes. Several first-line arms enrol NPM1-mutant and KMT2A-rearranged patients together; a pooled response rate is not a genotype-level one.
- Trial-level endpoints. A record's primary outcomes often apply to one part of the trial only (a Phase 2 cohort, a sub-study, US sites); each cohort above was checked against the outcome text, not assumed.
- One record, Horizen-1 (NCT04988555), cites "ELN2017" for some response outcomes and "ELN2027" for two Phase 2 outcomes, and writes "MFLS" for morphologic leukaemia-free state. No 2027 edition of the ELN recommendations exists; this page reports the record as written and does not assume which edition is meant.
- ASH 2026 abstracts are not yet read into this page. They will be added after reading at source, and the page will be updated during the meeting.
The cohort table behind this page
116 menin-inhibitor treatment arms (CSV), read from ClinicalTrials.gov on 2026-11-17: one row per arm, with disease, genotype, line, backbone, the primary and secondary endpoints that apply to that arm's part of the trial (with the trial-level set alongside), the response criteria named, and a link to the record. Values the reading rules could not establish are marked "not stated", and arms whose outcomes name trial parts the arm label does not are marked "unclear", rather than inferred.
Landscape reviewed 17 November 2026. This page is a reference on menin inhibitors in acute leukaemia, maintained as registries, regulatory actions and congress disclosures change; it is not a news briefing and does not claim to be exhaustive. Every fact links to the primary record behind it, and trial details are read from the registry itself on the review date rather than from a release describing it. Enrolment figures are planned targets unless the record states otherwise. No claim of superiority or inferiority is made about any programme, and no cross-trial efficacy comparison is made: the differences described are structural. Prognyx has no affiliation with any company named here and holds no position in their securities. Names and trademarks belong to their owners. Nothing here is medical or investment advice. Found an error? Write to hello@prognyx.com and it will be corrected.
Frequently asked
Which menin inhibitors are approved in the United States?
Two. Revumenib (Revuforj, Syndax) was approved on 15 November 2024 for relapsed or refractory acute leukaemia with a KMT2A translocation and on 24 October 2025 for relapsed or refractory AML with a susceptible NPM1 mutation, in adults and children from one year of age. Ziftomenib (Komzifti, Kura Oncology) was approved on 13 November 2025 for adults with relapsed or refractory AML with a susceptible NPM1 mutation.
Can response rates from different menin-inhibitor trials be compared?
Only with care, and often not at all. The single-agent programmes register CR + CRh as their primary efficacy endpoint, while the ziftomenib combination study NCT05735184 registers CR alone by ELN 2022. Populations also differ in genotype, line of therapy, prior treatment and age, and several first-line trials pool NPM1-mutant and KMT2A-rearranged patients. Cross-trial comparisons are hypothesis-generating at best.
What is the difference between CR and CR + CRh?
Complete remission (CR) requires the marrow and blood criteria of remission together with full recovery of neutrophil and platelet counts. CR with partial haematological recovery (CRh) meets the marrow criteria but with only partial recovery of those counts. CR + CRh is the sum of both, so it is always at least as high as CR alone: in the FDA summary for ziftomenib, CR was 17.0% and CR + CRh 21.4%.
Which Phase 3 trials test menin inhibitors in first-line AML?
On the registry at the review date: KO-MEN-017 (NCT07007312, ziftomenib, Kura), with an intensive and a non-intensive study under one record; cAMeLot-2 (NCT06852222, bleximenib with venetoclax and azacitidine, Janssen); and HOVON 181 (NCT07223814, bleximenib with intensive chemotherapy, HOVON with the German-Austrian AML Study Group). All are randomised, placebo-controlled and enrol NPM1-mutant or KMT2A-rearranged AML.
How is this page maintained?
Each registry record is re-read from ClinicalTrials.gov rather than carried forward, and a dated copy of every record is kept so that field-level changes between reads can be reported. Regulatory statements are traced to the FDA approval summaries. Where something cannot be confirmed at a primary record, the page says so rather than estimating.
Your cohort, read the same way.
This page reads a whole class from the public record. A client brief applies the same discipline to one asset and one decision: which competitor cohorts your data will be held against, on which definition, and what that does to the comparison. Name the asset.
Scope a pilot → The KRAS G12C landscape