Oncofusion

start / graph / platform-counterargument-adult-commercial-gravity

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q8-impact

The counter-argument: the menin field is an adult field, and the infant studies were withdrawn

also called why the platform argument fails for infants · adult commercial gravity · paediatric deprioritisation menin

The drugs that prove the platform idea works are being tested almost entirely in adults in their late sixties, and the one menin-inhibitor study that would have taken babies from thirty days old was cancelled before a single child was enrolled. A platform argument that ignores this is describing a world that does not exist.

The same thing, for a scientist

The platform thesis is supported by mechanism and refuted by behaviour. Menin inhibition works across KMT2A-rearranged and NPM1-mutant leukemia, but the registrational populations had median ages of 37 (AUGMENT-101, revumenib) and 66-69 (ziftomenib), the second approved menin inhibitor carries an adults-only label, bleximenib's paediatric exposure is a phase 1 cohort with a two-year age floor inside a study whose phase 2 and phase 3 are adults-only, and Janssen's dedicated paediatric combination study NCT05521087, whose minimum age was 30 days and which would therefore have enrolled infants, is recorded as withdrawn with enrolment zero. Two further paediatric revumenib studies are recorded withdrawn with zero enrolment. No trial randomises a menin inhibitor frontline in infants. The infant share of the menin-addressable population is roughly 1.2%, which is the quantitative reason this keeps happening.

Go deeper
The five facts, in order of force3 deeper
  1. The regulator formally excused the sponsor from studying infants. The FDA approval letter for NDA 218944 states: "We are waiving the pediatric study requirement for patients less than 1 year of age because necessary studies are impossible or highly impracticable. This is because of the small number and geographical dispersion of pediatric patients in this age group." 5 That waiver was granted for the drug whose target lesion defines infant leukemia. It is not an oversight; it is the paediatric-incentive system working exactly as written, and it is the single most important sentence in this beat.
  2. The one study that would have dosed infants was withdrawn with zero enrolment. NCT05521087, a Janssen phase 1/1b of JNJ-75276617 (bleximenib) plus conventional chemotherapy for paediatric and young adult relapsed/refractory acute leukemia with KMT2A, NPM1 or nucleoporin alterations, lists a minimum age of 30 days and a maximum of 30 years. Its status is WITHDRAWN with enrolment 0 1. A 30-day age floor means this protocol was written to include infants. It never opened.
  3. A second and third paediatric study went the same way. The graph's 2026 trial map records two further withdrawn paediatric menin-inhibitor studies with zero enrolment, NCT07498465 and NCT07052994, against planned enrolment of 3,091 across four adult phase 3 programmes 2.
  4. The approved populations are adult. AUGMENT-101, which supported the first menin-inhibitor approval covering patients from one year of age, had a median patient age of 37 (range 1-79) 3. It was not childless: the label's paediatric-use section records 13 infants under 2 years, 41 children aged 2 to under 12 and 16 adolescents, with 25 of 104 patients (24%) under 17 5. But the label also states that safety and efficacy in patients under 1 year have not been established 5, so the population the lab cares about most is the one the label explicitly exempts. Ziftomenib's registrational population had a median age of 69 (range 33-84) and was NPM1-mutant, not KMT2A-rearranged 4, and the follow-on combination cohort a median age of 66 4.
  5. The second menin inhibitor's entire pivotal programme excludes children by protocol. KOMET-001 (NCT04067336), the ziftomenib first-in-human phase 1/2 that enrolled 263 patients across four sub-studies including a KMT2A-rearranged ALL sub-study, specifies a minimum age of 18 years 4. Not deferred, not under-enrolled: excluded in the eligibility criteria.
  6. The second approved menin inhibitor excludes children entirely. KOMZIFTI's indication is adults with relapsed or refractory NPM1-mutant AML (see cost-menin-inhibitor-course).
Why it happens: the 1.2% number

This is not a story about bad actors. It is arithmetic.

The menin-addressable US population is roughly 9,700 patients a year: ~2,900 KMT2A-rearranged across all ages plus ~6,800 NPM1-mutant AML. Infants with KMT2A-rearranged leukemia are roughly 120 of them, about 1.2% (see platform-one-asset-two-populations-menin).

Now add four multipliers that all point the same way.

  • Accrual rate. Adult AML patients present at ~22,720 a year in the US alone; infants at ~160. A site that sees one infant every two years cannot support a dose-escalation cohort.
  • Formulation. Revumenib is a tablet. Infants need a dispersible or liquid formulation and BSA-based dosing, which is extra development work with no adult return.
  • Risk. Differentiation syndrome and QTc prolongation are harder to manage and monitor in a 9 kg infant, and a paediatric safety signal can jeopardise the adult programme.
  • Payer mix. The majority of US infants with leukemia are covered by Medicaid or CHIP, with statutory rebates capping net price (see cost-family-financial-toxicity-infant-leukemia).

Every one of those is a rational reason for a sponsor to defer infants. Together they mean the platform's success in adults actively crowds out the paediatric population it was supposed to rescue.

What survives of the platform argument

The honest scorecard, written so the lab can use it in a grant without being contradicted.

Survives:

  • Biologically, the argument is intact and well evidenced. One obligate interface serves several genotypes; that is demonstrated, not asserted.
  • Regulatorily, it works: the revumenib label already covers patients from one year of age, so an infant can be prescribed the drug today without a new trial.
  • Economically for the asset, it works: joining KMT2A to NPM1 made the programme fundable.

Does not survive:

  • The claim that a platform, on its own, causes therapies to reach children. It does not. In this field it has demonstrably not.
  • The claim that a label covering one-year-olds means the evidence covers one-year-olds. The evidence is a median 37-year-old and a handful of children.
  • Any framing in which the commercial success of the adult indication is presented as good news for infants without qualification.

Therefore the correct form of the lab's argument is narrower and stronger: a platform makes infant therapy possible; only a deliberate mechanism - paediatric investigation plans with short deferrals, consortium-run frontline trials, Bayesian borrowing from adult data, or a funder willing to pay for the infant cohort directly - makes it happen. The platform is necessary and demonstrably not sufficient.

The falsifiable prediction this node makes

If the counter-argument is right, then over the next three years: no sponsor-run randomised frontline infant menin-inhibitor trial will open; paediatric enrolment in menin-inhibitor studies will remain below 10% of total; and any new paediatric menin protocol registered will either have an age floor above one year or will be withdrawn or suspended.

If instead an infant frontline randomised trial opens with sponsor funding by 2029, this node is wrong and should be rewritten. That is a real, dated, checkable prediction, and concept-open-trial-map-kmt2a-r-2026 is the place to check it.

Facts and figures
lowest registered age floor menin paediatric study30 days (NCT05521087, withdrawn)
planned adult phase3 enrolment3091
infant share of menin addressable population PCT1.2
frontline infant randomised trial existsfalse
falsifiable predictionNo sponsor-funded randomised frontline infant menin-inhibitor trial opens before 2029; paediatric enrolment stays below 10% of menin-inhibitor totals.
as of2026-09-18
ziftomenib pivotal minimum age years18
ziftomenib pivotal enrolment263
prea waiver under 1 yeartrue
median ages in registrational populations
revumenib_augment_10137.0revumenib_augment_101: 37ziftomenib_combination66.0ziftomenib_combination: 66ziftomenib_monotherapy69.0ziftomenib_monotherapy: 69
augment 101 paediatric counts
infants_under_213.0infants_under_2: 13adolescents16.0adolescents: 16under_17_of_10425.0under_17_of_104: 25children_2_to_under_1241.0children_2_to_under_12: 41
median ages in registrational populations — the numbers behind the chart
{
  "revumenib_augment_101": 37,
  "ziftomenib_monotherapy": 69,
  "ziftomenib_combination": 66
}
withdrawn paediatric studies zero enrolment
[
  "NCT05521087",
  "NCT07498465",
  "NCT07052994"
]
augment 101 paediatric counts — the numbers behind the chart
{
  "infants_under_2": 13,
  "children_2_to_under_12": 41,
  "adolescents": 16,
  "under_17_of_104": 25
}
How this could bite us back
  • The strongest form of the paediatric-gap problem is not that sponsors ignore infants; it is that the regulator formally agreed infants need not be studied, on feasibility grounds, for the drug that targets the lesion defining infant leukemia. Any incentive design that does not close that waiver route will reproduce this outcome for the next asset.
  • An incentive that rewards approval in a rare paediatric disease while letting the sponsor recruit adults is not a neutral inefficiency. It produces labels that cover infants on evidence generated in sixty-nine-year-olds, and clinicians then dose infants off that label with no paediatric efficacy data. The risk lands on the children.
  • If the lab builds its case on the platform argument and a reviewer knows the withdrawn-study history, the proposal loses credibility it did not need to lose. Stating the counter-argument first is strictly stronger than having it raised.
  • Deferral compounds. Every year the paediatric question is deferred, the adult standard of care moves, and the eventual infant trial must be run against a comparator that has itself changed, which raises the required sample size in a population that cannot supply it.
Still open
  1. Why were NCT05521087, NCT07498465 and NCT07052994 withdrawn with zero enrolment? Retrieving the reason from the sponsors is the single most decision-relevant unanswered question in this beat.
  2. What is the actual number of infants under one year ever dosed with any menin inhibitor, anywhere? Nobody has published it and it may be in single digits.
  3. Is there any funding mechanism, public or philanthropic, that has ever paid a sponsor directly to run an infant cohort, and what did it cost?
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry

    NCT05521087, a Janssen phase 1/1b of JNJ-75276617 plus conventional chemotherapy in paediatric and young adult patients with relapsed/refractory acute leukemia harbouring KMT2A, NPM1 or nucleoporin alterations, has a minimum age of 30 days and a maximum of 30 years, and is recorded as WITHDRAWN with an enrolment of 0.

    NCT05521087
    What this evidence cannot show

    The registry records the status and zero enrolment but gives no reason. A withdrawal can reflect portfolio reprioritisation, protocol redesign reappearing under a new number, competitive accrual or a regulatory issue. Do not infer motive from this field; infer only that no child was dosed under this protocol.

    Evidence that argues against this (1)
    • Correction to this item's own caveat, which says the registry 'gives no reason' and instructs the reader not to infer motive. The registry does give a reason. The ClinicalTrials.gov API v2 record for NCT05521087 carries whyStopped = 'Business Decision' alongside overallStatus WITHDRAWN and enrolment 0 ACTUAL, with a registered start date of 2025-12-26. Retrieved from the API on 2026-09-18 (fields NCTId, OverallStatus, WhyStopped, EnrollmentCount). The field is not visible through every registry front end, which is presumably why it was missed. This strengthens the node's argument -- the one protocol written to include infants was withdrawn for commercial reasons -- and the caveat should be rewritten rather than left as a warning against an inference the registry itself licenses. Janssen Research & Development et al. 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry

    As of the 2026 trial map, the menin-inhibitor landscape comprised four adult phase 3 programmes with planned enrolment totalling 3,091, several recruiting paediatric phase 1 combinations, post-transplant maintenance studies, and two withdrawn paediatric studies with zero enrolment (NCT07498465, NCT07052994).

    Open trial map for KMT2A-rearranged acute leukemia, 2026 · Oncofusion Knowledge Graph, therapeutics beat · ClinicalTrials.gov sweep recorded in this graph 2026
    What this evidence cannot show

    This is an internal graph node summarising a registry sweep rather than an independent publication; the individual NCT records are the primary evidence and each should be re-checked before citation, since registry statuses change.

    Evidence that argues against this (2)
    • Two of the three withdrawals are not commercial, so counting all three as evidence of adult commercial gravity over-attributes. Retrieved from the ClinicalTrials.gov API v2 on 2026-09-18: NCT07052994 is sponsored by M.D. Anderson Cancer Center, not a pharmaceutical company, and its whyStopped field reads 'PI requested'; NCT07498465 is sponsored by the Children's Oncology Group and its whyStopped field reads 'Other - Protocol moved to Withdrawn'. Both are academic or cooperative-group protocols withdrawn for reasons the registry does not attribute to a sponsor's portfolio decision. Only NCT05521087 carries 'Business Decision'. The honest version of this claim is one industry withdrawal for commercial reasons plus two academic withdrawals for unstated reasons -- which is a weaker pattern than three. US National Library of Medicine (registry) 2026
    • The claim that paediatric menin studies do not open is refuted by the paediatric menin studies that opened, ran and closed. Read from the ClinicalTrials.gov API v2 on 2026-09-18, a sweep over revumenib, ziftomenib, bleximenib, enzomenib, DSP-5336, BMF-219 and DS-1594 returned 58 records, of which seven have age floors below 12 years. Three are decisive against this node's framing: AUGMENT-102 (NCT05326516, Syndax, revumenib plus chemotherapy, minimum age 30 Days) is COMPLETED with an actual enrolment of 30, running 2022-03-09 to 2024-07-29 in the US and Canada; NCT05761171, the Children's Oncology Group phase 2 'A Phase 2 Study of SNDX-5613 in Combination With Chemotherapy for Patients With Relapsed or Refractory KMT2A-Rearranged Infant Leukemia', ages 1 Month to 6 Years, planned enrolment 78, is ACTIVE_NOT_RECRUITING having started 2024-01-08; and NCT06376162 (ITCC-101/APAL2020K, ziftomenib, minimum age 0 Years, seven countries, Kura Oncology collaborator) is RECRUITING. Infants and young children are being dosed with menin inhibitors under open protocols. Syndax Pharmaceuticals et al. 2024
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    AUGMENT-101, the registrational study supporting revumenib approval in relapsed or refractory KMT2A-translocated acute leukemia in patients one year and older, enrolled a population with a median age of 37 years (range 1.3-75) and reported a CR+CRh of 22.8% among 57 efficacy-evaluable patients, with no published paediatric subgroup efficacy analysis.

    AUGMENT-101: revumenib in relapsed or refractory acute leukemia with KMT2A rearrangement · Syndax Pharmaceuticals / Aldoss I, Issa GC, et al. · ClinicalTrials.gov 2024
    NCT04065399
    What this evidence cannot show

    Single-arm registration design without a randomised comparator. The median age and absence of a paediatric subgroup are the load-bearing facts here and are recorded in this graph's `trial-augment-101` entry, which should be consulted for the primary publication.

    Evidence that argues against this (1)
    • A median age of 37 in a study open from one year is the right fact for the commercial-gravity argument and it needs its complement stated: 25 of 104 patients (24%) were under 17 and the label records 13 infants. The distribution is not adult-only; it is adult-weighted, and the two support different conclusions about what a paediatric programme would have to add. Journal of Clinical Oncology 2025
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    The ziftomenib first-in-human study KOMET-001 (NCT04067336), which supported the ziftomenib programme, specifies a minimum age of 18 years and enrolled 263 patients; its reported registrational population had a median age of 69 (range 33-84) with CR/CRh 22%, overall response 33% and median overall survival 6.6 months, and the ziftomenib-venetoclax-azacitidine combination cohort had a median age of 66.

    Ziftomenib in relapsed or refractory NPM1-mutant acute myeloid leukemia (KOMET-001) · Kura Oncology / Wang ES, Issa GC, et al. · ClinicalTrials.gov 2025
    NCT04067336
    What this evidence cannot show

    The 18-year minimum age, ACTIVE_NOT_RECRUITING status and enrolment of 263 were read directly from the ClinicalTrials.gov record on 2026-09-18 and are solid. The median-age and efficacy figures are recorded in this graph's `collab-kura-oncology` and `collab-stein-issa-clinical-menin` entries from the primary reports and were not independently re-verified here. The population is NPM1-mutant, so it speaks to the menin field's age distribution rather than to KMT2A-rearranged disease.

    Evidence that argues against this (1)
    • The strongest counter-argument to this counter-argument, and the answer to the 'is there a precedent' question: a menin inhibitor is already in a frontline infant protocol, with its manufacturer named as a collaborator. NCT05848687, 'TINI 2: Total Therapy for Infants With Acute Lymphoblastic Leukemia II', has a maximum age of 1 Year, planned enrolment 90, start 2023-11-03 and status RECRUITING; its lead investigator is Tanja Andrea Gruber and its listed collaborators include the Pediatric Oncology Experimental Therapeutics Investigators' Consortium, Amgen, the Lucile Packard Foundation for Children's Health and Kura Oncology, Inc.; and its intervention list is Dexamethasone, Mitoxantrone, PEG asparaginase, Bortezomib, Vorinostat, Mercaptopurine, Methotrexate, Blinatumomab and Ziftomenib. Read from the ClinicalTrials.gov API v2 record (protocolSection.armsInterventionsModule) on 2026-09-18. This node's literal sentence 'No trial randomises a menin inhibitor frontline in infants' survives, because TINI 2 is a single-arm phase 1/2. Its narrative does not: the second menin inhibitor, the one whose label excludes children, is being given frontline to infants under one year with its sponsor's participation. Gruber TA (Stanford University); collaborators incl. Kura Oncology et al. 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    The FDA approval letter for NDA 218944 (revumenib) waived the paediatric study requirement for patients less than 1 year of age as 'impossible or highly impracticable' because of the small number and geographical dispersion of patients in that age group; the label's paediatric-use section records 13 infants under 2 years, 41 children aged 2 to under 12 and 16 adolescents in the pivotal KMT2A cohort (25 of 104, 24%, under 17) and states that safety and efficacy in patients under 1 year have not been established.

    What this evidence cannot show

    The waiver applies to the statutory paediatric study requirement under PREA and does not prohibit voluntary study in infants; it records the agency's judgement about feasibility, which the arithmetic in `ana-infant-trial-feasibility-arithmetic` partly supports and partly contradicts. The patient-count breakdown is from label sections 8.4 and 14.1 as recorded in `reg-breakthrough-and-accelerated-approval`.

    Evidence that argues against this (1)
    • The node's dated falsifiable prediction should be re-read against a trial that already exists. NCT06317662 (National Cancer Institute, RECRUITING since 2025-06-05, planned enrolment 153, maximum age 365 days, allocation RANDOMIZED) randomises KMT2A-rearranged infants to the addition of venetoclax on a blinatumomab-plus-chemotherapy frontline backbone. It is not a menin inhibitor and it is not industry-sponsored, so the node's specific prediction is not yet falsified. But it removes the general premise -- that a randomised frontline infant trial cannot be opened or accrued -- on which the prediction's plausibility rests, and it shows the mechanism the node calls for (a public funder paying for the infant cohort directly) already operating. Read from the ClinicalTrials.gov API v2 record on 2026-09-18. National Cancer Institute (NCI) 2026
Connected to