Children's Cancer Hospital Egypt (n=150 KMT2A-r): EFS 50% to 17% across four partner groups in a single-centre cohort
also called Fayez 2025 Acta Oncologica · CCHE paediatric KMT2A-r AML partner outcome
A single large hospital in Egypt followed 150 children with this leukemia and grouped them by partner gene. Survival ran from half to one in six depending on the group, though extra chromosome changes mattered too.
The same thing, for a scientist
A retrospective analysis of 610 children with intermediate-risk AML treated at Children's Cancer Hospital Egypt between January 2008 and December 2021 identified 150 (24.6%) with KMT2A rearrangements, assigned to four fusion-partner groups. t(9;11)(p22;q23)/KMT2A::MLLT3 was the most prevalent partner at 40.8%. KMT2A rearrangement was an independent predictor of relapse (CIR 46% versus 30%, P=0.006) and was associated with hyperleukocytosis (P=0.029), CNS disease (P=0.003), monocytic differentiation (P=0.001), additional cytogenetic abnormalities (P=0.04) and complex karyotype (P=0.001). Outcome varied by fusion partner group, with EFS ranging from 50% to 17%, OS from 54% to 27% and CIR from 75% to 38%. Within the KMT2A-r group, additional cytogenetic abnormalities and complex karyotype independently worsened outcome (5-year OS 34% versus 53%, P=0.027, and 26% versus 51%, P=0.004).
Why a single-centre cohort from outside the usual collaborations matters here
Every other positive partner result in this beat comes from I-BFM, Ponte-di-Legno or COG - overlapping collaborations that share patients, classifications and protocols. This one is independent of all of them, uses a single protocol at a single institution, and still finds a partner range 1. That removes the pooled-multi-group confound entirely, which is the main methodological criticism of the large positives.
But the cohort is not independent of the confounds it demonstrates
KMT2A rearrangement was itself associated with additional cytogenetic abnormalities (P=0.04) and complex karyotype (P=0.001), and both independently worsened outcome within the KMT2A-r group by 19 and 25 OS points 1. That is the same confound the I-BFM collaboration identified in 733 children in 2011 2. Whether the partner groups here differ in ACA rate is not reported, so the partner range and the ACA effect cannot be separated from the published summary.
What the reported range does and does not establish
EFS 50% to 17%, OS 54% to 27%, CIR 75% to 38% across four groups 1. Per-group n are not given in the abstract, and with 150 patients across four groups and a modal partner at 40.8% the smallest group is likely under 20 patients. The range is wide and the arms are small - which is the standing condition of this entire literature.
| metric | event-free survival, overall survival and cumulative incidence of relapse by fusion-partner group |
|---|---|
| value | EFS 50% to 17% |
| unit | % across four partner groups |
| os range PCT | 54 to 27 |
| cir range PCT | 75 to 38 |
| population | children with intermediate-risk AML, single centre, 2008-2021 |
| n intermediate risk aml | 610 |
| n kmt2a r | 150 |
| kmt2a r PCT | 24.6 |
| partner groups | 4 |
| modal partner | t(9;11)(p22;q23) / KMT2A::MLLT3 |
| modal partner PCT | 40.8 |
| region | Egypt (Children's Cancer Hospital Egypt) |
| year | 2025 |
| prespecified | no - retrospective single-centre cohort |
| per group n reported | false |
| source | PMID 40965326 |
kmt2a r vs negative — the numbers behind the chart
{
"cir_pct": 46,
"comparator_cir_pct": 30,
"p": 0.006
}within kmt2a r — the numbers behind the chart
{
"aca_os_5y_pct": 34,
"no_aca_os_5y_pct": 53,
"aca_p": 0.027,
"complex_os_5y_pct": 26,
"non_complex_os_5y_pct": 51,
"complex_p": 0.004
}- Outcomes here are lower across the board than in the European and North American cohorts, so a menin inhibitor evaluated against these baselines would look more effective for reasons that have nothing to do with the drug or the partner.
- If additional cytogenetic abnormalities carry much of the partner range, then a menin inhibitor selected on partner will be given to patients whose real adverse factor is karyotypic complexity, which menin inhibition does not address.
- Which four partner groups were used, and what are the per-group n? Without them the 50%-to-17% range cannot be weighted or compared with other cohorts.
- Do the partner groups differ in additional cytogenetic abnormality rate? If they do, the partner range in this cohort may be an ACA range.
- Does the partner ordering here match the I-BFM high-risk/non-high-risk assignment, which would be an independent external validation of that classification?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Among 610 children with intermediate-risk AML at a single centre, 150 (24.6%) had KMT2A rearrangements assigned to four fusion-partner groups with t(9;11)/KMT2A::MLLT3 most prevalent (40.8%); outcome varied by partner group with EFS 50% to 17%, OS 54% to 27% and CIR 75% to 38%, while KMT2A rearrangement independently predicted relapse (CIR 46% versus 30%, P=0.006) and additional cytogenetic abnormalities and complex karyotype independently worsened outcome within the KMT2A-r group (5-year OS 34% versus 53%, P=0.027; 26% versus 51%, P=0.004).
Balancing benefit and burden: treatment intensification in paediatric KMT2A rearrangements acute myeloid leukaemia. · Fayez H, Elsherif M, et al. · Acta Oncol 2025What this evidence cannot show
Single centre, retrospective, 2008-2021, so era varies within the cohort. Per-group n and the identity of the four partner groups are not given in the abstract; with 150 patients and a 40.8% modal partner the smaller groups are likely under 20. No P value is reported for the partner comparison itself in the abstract - only the range. Restricted to intermediate-risk AML by the cohort definition, so it is not a whole-AML population. Whether the partner groups differ in additional cytogenetic abnormality rate is not stated, so partner and ACA effects are not separated.
Evidence that argues against this (2)
- In 733 children with 11q23/MLL-rearranged AML, additional cytogenetic aberrations independently predicted outcome with an effect size comparable to the partner (5-year OS 47% with ACA versus 62% without, P<.001), so a partner range reported without ACA adjustment may largely reflect differing ACA rates between partner groups. Coenen EA et al. 2011
- Two adult AML series of 172 and 205 patients found no partner effect on survival at all, so a single-centre paediatric range does not settle the question across the disease. Issa GC et al. 2021
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Additional cytogenetic aberrations are an established independent prognostic factor in paediatric 11q23/MLL-rearranged AML: in 733 children, ACA were present in 47% and associated with 5-year OS 47% versus 62% (P<.001), with trisomy 8 favourable within the ACA group (OS 61% versus 39%, P=.003) and trisomy 19 adverse (OS 24% versus 50%, P<.001).
Prognostic significance of additional cytogenetic aberrations in 733 de novo pediatric 11q23/MLL-rearranged AML patients: results of an international study. · Coenen EA, Raimondi SC, et al. · Blood 2011What this evidence cannot show
Retrospective multi-group pooling from the same collaboration that defined the partner risk classification, so the ACA and partner analyses share a cohort and are not independent. Complex karyotype was significant in univariate analysis only.
Evidence that argues against this (1)
- The partner-based risk group remained independently prognostic in a later 1,130-patient analysis after adjustment for end-of-induction MRD, indicating the partner effect is not entirely reducible to cytogenetic background. van Weelderen RE et al. 2023
- ← related toThe protocol and era confound: the partner datasets span 1983 to 2016 and outcome moved under them
a single-centre single-protocol cohort, which removes the pooling confound
EFS 50% to 17% across four partner groups, n=150, single centre
an independent single-centre paediatric cohort that also finds a partner range
additional cytogenetic abnormalities cost 19 and complex karyotype 25 OS points within KMT2A-r