Oncofusion

start / graph / concept-hsct-in-kmt2a-r-leukemia

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q6b-therapeutics2

Haematopoietic stem cell transplant in KMT2A-rearranged infant leukemia — what it is for, and why most children who need it never get it

also called HSCT · allogeneic stem cell transplant · bone marrow transplant · SCT in first remission

A transplant replaces the child's blood-forming system with a donor's, and the donor immune cells then attack any remaining leukemia. For high-risk babies it is the intended plan, but in the largest trial fewer than half of the babies it was meant for actually reached it, because they relapsed or died first.

The same thing, for a scientist

Allogeneic HSCT in KMT2A-rearranged infant ALL is reserved for the high-risk stratum, where it is intended to deliver a graft-versus-leukemia effect after chemotherapy has achieved remission. In Interfant-06, high-risk infants were assigned to transplant in first remission and had a 6-year EFS of 20.9% (SE 3.4) — but only 46% actually received a transplant, because many had early events. One binding constraint is therefore the inability to hold a high-risk infant in remission long enough to reach transplant. **But transplant efficacy is not settled and may point the other way**: in the largest dataset in this disease, 256 patients with t(4;11), any transplantation was associated with WORSE disease-free and overall survival than chemotherapy alone (PMID 12057554), and Interfant-06's own protocol states there are no data supporting extensive use of SCT in infants and no randomised study has been done. That reframes every non-chemotherapy agent in this beat as a potential bridge, and it is the clearest rationale for post-transplant menin-inhibitor maintenance, which is now in five registered trials.

Go deeper
The number that defines the problem1 deeper

In Interfant-06, high-risk infants — KMT2A-rearranged, under 6 months, with WBC ≥300 ×10⁹/L and/or prednisone poor response — were intended to proceed to stem-cell transplant in first remission.

Only 46% of them received one, because many had early events. Their 6-year EFS was 20.9% (SE 3.4) 1.

Read that carefully. It is not a statement that transplant does not work. It is a statement that the majority of the children a transplant was designed for relapsed or died before the transplant could happen. Those are different problems requiring different solutions, and the field routinely conflates them when it debates whether HSCT is worthwhile in infant ALL.

The practical consequence: an agent that does nothing except keep a high-risk infant in MRD-negative remission for three more months is doing something that Interfant-06's chemotherapy could not do for more than half of these children. That is the honest framing of blinatumomab (which produces MRD-negative remission in 93%), of inotuzumab (80% ORR, 40% one-year survival, 9 of 25 proceeding to HSCT or CAR-T), and of menin inhibitors in the relapsed setting.

Transplant is also, increasingly, the thing patients have already failed

In the European real-world CD19 CAR-T cohort of children under 3 (76% KMT2A-rearranged), 66% had relapsed after a prior allogeneic HSCT 2. So by the time an infant reaches cellular therapy, transplant is typically already behind them and already failed. That inverts the classic sequencing assumption that transplant is the definitive consolidation at the end of the line.

What transplant buys, and what it costs an infant

What it buys: a graft-versus-leukemia effect that is mechanistically independent of every drug resistance mechanism in this beat. A donor T cell does not care about MEN1 interface mutations, PRC1.1 loss, RAS co-mutation or adaptive transcriptional states. In a disease whose defining feature is that the leukemia keeps finding ways around targeted agents, an immune effector mechanism with no shared escape route is genuinely valuable.

What it costs a baby: total body irradiation or busulfan-based conditioning in the first year of life is among the most damaging exposures in paediatric medicine. Growth failure, endocrine failure, infertility, neurocognitive impairment and second malignancy are not rare late effects here; they are expected ones. This graph covers the late-effects picture separately under concept-growth-endocrine-late-effects-infant-leukemia and concept-neurocognitive-late-effects-infant-leukemia. Graft-versus-host disease in an infant, and the immunosuppression required to control it, sits on top of that.

The risk-benefit is therefore extremely sensitive to the alternative. If blinatumomab-containing frontline therapy genuinely delivers 4-year DFS above 80%, the population for whom transplant's late-effects cost is justified shrinks substantially — and that is one of the most consequential secondary questions Interfant-21 will answer.

Post-transplant maintenance: the fastest-moving idea in this space

If relapse after transplant is the failure mode, and a well-tolerated oral agent can suppress residual disease without myelosuppression, post-transplant maintenance is the obvious application. Five registered studies are pursuing it 3:

  • NCT06575296 — revumenib post-transplant maintenance, KMT2A-r or NPM1-m, childhood cohorts listed. Phase 1, n=27, recruiting (City of Hope).
  • NCT06440135 — ziftomenib maintenance after allogeneic HCT. Phase 1, n=22, recruiting (MGH).
  • NCT07563010 — revumenib versus placebo as post-HSCT maintenance, KMT2Ar/NPM1m/NUP98r AML. Phase 2 randomised, n=146, not yet recruiting (CIBMTR). The only randomised maintenance study.
  • NCT07101497 — BN104 post-HSCT maintenance. Phase 2, n=60, recruiting.
  • NCT07559695 — real-world observational study of menin inhibitors as post-HSCT maintenance.

And one that will not happen: NCT07498465, the COG study of revumenib as post-HSCT maintenance in paediatric ALL, AML and MPAL, is WITHDRAWN with zero enrolment. That was the trial that would have characterised long-duration menin exposure in children, including the endocrine safety signal the adult labels raise. Its absence is the largest single gap in paediatric menin-inhibitor safety.

Facts and figures
modalityallogeneic haematopoietic stem cell transplantation
indication in infant allhigh-risk stratum only (KMT2A-rearranged, age <6 months, WBC >=300 x10^9/L and/or prednisone poor response), in first remission
key constraintin Interfant-06 only 46% of high-risk infants assigned to transplant received one; 6-year EFS in that group was 20.9% (SE 3.4)
mechanism of benefitgraft-versus-leukemia effect, mechanistically independent of every pharmacological resistance route in this disease
post transplant relapse frequency66% of infants and young children referred for CD19 CAR-T had relapsed after prior allogeneic HSCT
withdrawn pediatric maintenance trialNCT07498465 (COG), enrolment 0
pediatric eligibletrue
maintenance trials
[
  "NCT06575296",
  "NCT06440135",
  "NCT07563010",
  "NCT07101497",
  "NCT07559695"
]
major late effects
[
  "growth failure",
  "endocrine failure including thyroid and gonadal",
  "infertility",
  "neurocognitive impairment",
  "second malignancy",
  "chronic graft-versus-host disease"
]
How this could bite us back
  • Late effects are the defining cost: total body irradiation or busulfan conditioning in the first year of life produces growth failure, endocrine failure, infertility and neurocognitive impairment as expected outcomes, not rare complications.
  • Essential-function toxicity of maintenance: menin is required for long-term HSC function under transplantation stress in mouse models, so giving a menin inhibitor to a child whose donor graft is still establishing itself targets the exact function the graft depends on.
  • Attrition rather than failure: more than half of the infants for whom transplant was indicated never received it, which means the intervention's real-world value is capped by disease kinetics rather than by biology.
  • Graft-versus-host disease in an infant, plus the immunosuppression to control it, on top of B-cell aplasia from prior CD19-directed therapy, produces a compound immunodeficiency with no age-specific management evidence.
  • Sequencing trap: 66% of infants reaching CAR-T had already failed transplant, so treating transplant as the definitive last step is empirically wrong and plans should be built assuming it may not hold.
Still open
  1. Is the problem transplant efficacy or transplant access? Interfant-06 cannot separate them, and the answer determines whether the field should improve conditioning or improve bridging. Everything in this beat argues for bridging.
  2. If blinatumomab-containing frontline therapy delivers the DFS the single-arm data suggest, which infants still need a transplant at all? This is arguably the most important late-effects question in paediatric oncology right now and Interfant-21 will inform it.
  3. Does post-transplant menin maintenance prevent relapse, and is it safe for years in a child? NCT07563010 is the only randomised test and the paediatric version of the question lost its trial when NCT07498465 was withdrawn.
  4. Does menin inhibition after transplant impair donor haematopoietic reconstitution? Menin is required for long-term HSC function under competitive transplantation stress in mice, which is precisely the stress a fresh graft is under.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In Interfant-06, high-risk infants were assigned to stem-cell transplant in first remission and had a 6-year event-free survival of 20.9% (SE 3.4), but only 46% actually received a transplant because many had early events.

    PMID 31283407DOI
    What this evidence cannot show

    Taken from the primary abstract; the full text was not retrieved, so no formal comparison of transplanted versus non-transplanted high-risk infants, and no confidence intervals, are asserted here. Comparing outcomes between those who did and did not reach transplant is confounded by guarantee-time bias in any case: reaching transplant requires surviving long enough to reach it.

    Evidence that argues against this (3)
    • The Interfant-99 transplant analysis found a real benefit, but only in a narrow subgroup - which cuts against blanket high-risk assignment as much as it does against blanket avoidance. Of 277 KMT2A-rearranged infants in first remission, 37 received HSCT and 240 chemotherapy only; the disease-free-survival advantage (p=.03) was restricted to the 97 of 297 (33%) with age under 6 months PLUS either poor steroid response at day 8 or WBC >=300 x10^9/L, where HSCT was associated with a 64% reduction in the risk of relapse or death in CR (HR 0.36, 95% CI 0.15-0.86). In all remaining patients there was no advantage of HSCT over chemotherapy at all. Mann G et al. 2010
    • A protocol that deliberately restricted HSCT did better than Interfant-06, not worse. The JPLSG MLL-10 trial confined the transplant option to its high-risk group and used early intensification with high-dose cytarabine instead, reporting 3-year EFS of 66.2% (SE 5.6) for KMT2A-rearranged infants and 94.4% (SE 5.4) for the intermediate-risk group, and concluded that risk stratification was able to ELIMINATE HSCT for a subset of KMT2A-rearranged infants. The Interfant-06 finding that only 46% of assigned infants reached transplant may therefore be less a delivery failure than evidence that transplant was assigned to the wrong denominator. Tomizawa D et al. 2020
    • THE LARGEST DATASET IN THIS DISEASE POINTS THE OTHER WAY, and this entry previously did not cite it. Among 256 children with t(4;11) - the largest such subgroup reported - ANY type of transplantation in first remission was associated with significantly worse disease-free survival (HR 1.61, 95% CI 1.10-2.35, p=0.014) and overall survival (HR 1.76, 95% CI 1.08-2.45, p=0.004) than chemotherapy alone, and even HLA-matched related or matched unrelated donor grafts 'tended to be associated with a worse outcome'. The study is retrospective, spans 1983-1995 under therapy no longer given, and is subject to confounding by indication in the direction that would deflate benefit - but it means the efficacy question is live rather than closed, and an entry that frames the constraint as reaching transplant rather than transplant working should say so. Pui CH et al. 2002
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In an international real-world cohort of 38 children under 3 years screened for tisagenlecleucel (76% KMT2A-rearranged), 25 (66%) had relapsed after a prior allogeneic haematopoietic stem cell transplant.

    PMID 36084658DOI
    What this evidence cannot show

    This cohort is selected for CAR-T referral, so a high rate of prior transplant is partly a referral artefact rather than a population statistic. It nonetheless establishes that post-transplant relapse is a routine clinical situation in this age group.

    Evidence that argues against this (1)
    • 66% prior transplant describes a CAR-T REFERRAL population, not the transplant rate in KMT2A-r infants. Children are screened for tisagenlecleucel precisely because a transplant has failed, so post-HSCT relapse is an entry criterion in effect. The prospective picture is different and moving the other way: in the JPLSG MLL-10 trial, transplant was deliberately restricted to high-risk infants (KMT2A-r plus age under 180 days or CNS leukaemia); 43 of 56 HR patients were transplanted in first remission, achieving 3-year EFS 56.8% (95% CI 42.4-68.8) and OS 80.2% (95% CI 67.1-88.5), and the investigators still concluded that 'considering the high rate of potentially life-threatening toxicities and the risk of late effects, its indication should be further restricted or even eliminated in the future'. Reading a referral cohort's prior-HSCT fraction as a statement about practice inverts the direction of the field. Takachi T et al. 2021
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.trial_registry

    Five registered studies are investigating menin inhibitors as post-transplant maintenance, including the only randomised one (NCT07563010, revumenib vs placebo, phase 2, n=146, CIBMTR, not yet recruiting) and a recruiting phase 1 with childhood cohorts (NCT06575296, n=27); the Children's Oncology Group study of revumenib as post-HSCT maintenance in paediatric patients (NCT07498465) was withdrawn with zero enrolment.

    NCT06575296
    What this evidence cannot show

    Registry snapshot of 2026-09-18. The other identifiers are NCT06440135, NCT07563010, NCT07101497, NCT07559695 and the withdrawn NCT07498465. None has reported. The registry does not state why NCT07498465 was withdrawn, and inferring a reason would be speculation.

    Evidence that argues against this (1)
    • Counting registered studies overstates what this programme can deliver, and for children it delivers nothing this decade. Checked against the raw ClinicalTrials.gov v2 API on 2026-09-19: the ONLY randomised study, NCT07563010 (CIBMTR, phase 2, revumenib vs placebo as post-allo-HSCT maintenance in KMT2Ar/NPM1m/NUP98r AML, estimated n=146, primary endpoint relapse-free survival), is NOT_YET_RECRUITING with a planned start of 2026-12, primary completion 2031-06, and a MINIMUM AGE OF 18 YEARS — it excludes children entirely. The recruiting paediatric-inclusive study, NCT06575296 (City of Hope, phase 1, estimated n=27, minimum age 2 years), has 'Incidence of adverse events' as its SOLE primary outcome with primary completion 2028-04-29, so by design it cannot produce an efficacy estimate. A count of five studies is not five answers. Center for International Blood and Marrow Transplant Research 2026
Connected to