Infant BCP-ALL (n=175): KMT2A::AFF1 2-year EFS 36% versus 37% for every other partner (P=0.9)
also called Dmitry Rogachev infant BCP-ALL cohort · AFF1 versus non-AFF1 infant EFS
In 175 babies with B-cell leukemia studied with modern sequencing, those whose KMT2A was fused to AFF1 did exactly as well, or badly, as those with any other partner. The two survival numbers were 36% and 37%.
The same thing, for a scientist
A single-centre cohort of 175 infants aged 0-365 days with BCP-ALL diagnosed 2010-2023, genotyped by karyotype, FISH and RNA-seq with Sanger validation, found no survival difference by KMT2A partner gene. KMT2A::AFF1 had 2-year EFS 36% (95% CI 21-59) against 37% (95% CI 23-60) for all other partners pooled (log-rank P=0.9). The same study found the KMT2A-germline group (n=33, 17.7%) contained NUTM1-r (n=9) and PAX5-r (n=10) cases with 2-year EFS of 80% (52-100) and 100% respectively. This is the most recent and most molecularly complete test of the partner question in infants, and it is flatly null.
The numbers are the point
36% versus 37%, P=0.9 1. Not a trend, not an underpowered near-miss - the point estimates are one percentage point apart. The confidence intervals are wide (21-59 and 23-60), so this does not exclude a moderate effect, but it is the cleanest quantified statement of partner equivalence in infants that this beat located.
AFF1 is the comparator that matters
Because the menin-interface bet in this graph is anchored on KMT2A::AFF1, a result in which AFF1 is indistinguishable from the pooled remainder is directly relevant: in infants, AFF1 is neither privileged nor penalised. Set against the I-BFM AML classification, where AFF1 sits in the high-risk group 2, the partner's meaning is evidently conditional on disease context rather than intrinsic.
What 'all other partners' hides
Pooling every non-AFF1 partner into one comparator arm is exactly the move that would conceal a partner effect if two non-AFF1 partners moved in opposite directions. In non-infant ALL the two extremes of the partner-by-lineage table are both non-AFF1 3. So this null constrains AFF1-versus-average, not partner-versus-partner.
| metric | 2-year event-free survival, KMT2A::AFF1 versus other KMT2A partners |
|---|---|
| value | 36 |
| unit | % (KMT2A::AFF1, 95% CI 21-59) |
| comparator value | 37 |
| comparator unit | % (all other KMT2A partners, 95% CI 23-60) |
| p value | 0.9 |
| population | infants aged 0-365 days with BCP-ALL, diagnosed 2010-2023 |
| n total | 175 |
| n kmt2a germline | 33 |
| region | Russia (Dmitry Rogachev National Medical Research Center) |
| year | 2025 |
| prespecified | no - retrospective single-centre cohort |
| source | PMID 40000924 |
germline subgroups
[
{
"lesion": "NUTM1-r",
"n": 9,
"efs_2y_pct": 80
},
{
"lesion": "PAX5-r",
"n": 10,
"efs_2y_pct": 100
}
]- If AFF1 and non-AFF1 infants truly have identical outcomes, then a menin inhibitor trial in infants gains no power from partner stratification - but it also loses the only cheap way of detecting that the drug works differently by partner.
- What are the per-partner curves rather than AFF1-versus-pooled? MLLT1 and MLLT3 arms reported separately would test whether pooling hid an effect.
- Does the null hold for MRD clearance as well as EFS in this cohort, given that a larger cross-age study found AFF1 associated with slower clearance?
- How many infants in the non-AFF1 arm had each partner? Without that, the pooled comparator cannot be interpreted.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Among 175 infants aged 0-365 days with BCP-ALL genotyped by karyotype, FISH and RNA-seq, survival did not differ by KMT2A partner gene: KMT2A::AFF1 2-year EFS 36% (95% CI 21-59) versus 37% (95% CI 23-60) for other partners, log-rank P=0.9.
Genetic Diversity in KMT2A -r and KMT2A -Wt Groups: Assessing the Prognostic Value of Markers in BCP-ALL Among Infants. · Ilyasova K, Zerkalenkova E, et al. · Int J Lab Hematol 2025What this evidence cannot show
Single centre, 175 infants across 2010-2023, so treatment era varies within the cohort and protocols are not stated in the abstract. The comparator is all non-AFF1 partners pooled, which would mask offsetting partner effects. Confidence intervals are wide (each spans roughly 35-40 percentage points), so moderate differences are not excluded. Two-year rather than five-year EFS.
Evidence that argues against this (2)
- In 465 KMT2A-rearranged B-ALL cases spanning 1 month to 89 years, AFF1 as fusion partner was significantly associated with slower MRD clearance (P=7.0E-04 univariate), and fusion partner remained independent of age and developmental maturity in multivariable analysis (P=0.03) - so an early-response partner effect involving AFF1 exists even where a survival effect does not. Hartmann AM et al. 2026
- In childhood AML, KMT2A::AFF1 is assigned to the high-risk partner group whose 5-year EFS is 24 points below the non-high-risk group, so AFF1 is not prognostically neutral across diseases. van Weelderen RE et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KMT2A::AFF1 (4q21) is one of the five partners in the I-BFM high-risk fusion-partner group in childhood KMT2A-rearranged AML.
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
Different lineage and different cohort; the contrast with the infant B-ALL null is a juxtaposition of two studies, not a tested interaction.
Evidence that argues against this (1)
- Two large adult AML series found no partner effect on survival at all, so the AML high-risk assignment is itself not universal across age within the myeloid lineage. Issa GC et al. 2021
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In non-infant childhood KMT2A-r ALL, the extremes of the partner-by-lineage outcome table are both non-AFF1 partners (t(9;11)/MLLT3 T-ALL 41.7% and t(11;19)/MLLT1 T-ALL 91.2%), so pooling non-AFF1 partners into one arm can average away a real partner effect.
Outcomes of Childhood Noninfant Acute Lymphoblastic Leukemia With 11q23/KMT2A Rearrangements in a Modern Therapy Era: A Retrospective International Study. · Attarbaschi A, Möricke A, et al. · J Clin Oncol 2023What this evidence cannot show
The two extreme cells have n=9 and n=34 and the comparison is retrospective and not pre-specified; they are cited here to show that pooling is hazardous, not to assert those specific values in infants.
Evidence that argues against this (1)
- CCG 1953, which compared the three major partners separately rather than pooling, also found no significant difference between them in infants, so the pooling criticism does not by itself rescue a partner effect in this age group. Hilden JM et al. 2006
AFF1 2-year EFS 36% versus 37% for other partners, P=0.9, n=175
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
AFF1 indistinguishable from pooled other partners in infants
second, RNA-seq-era infant cohort reaching the same null
the cleanest infant partner null
that entry already records this subtype's n=9 2-year EFS of 80% as the KMT2A-wild-type side of the same 175-infant cohort; this entry does not restate it and builds instead on the 85-case international series, which PubMed indexes with NO abstract