756 children with 11q23/MLL-rearranged AML: 5-year EFS runs from 11% to 92% depending on the partner
also called Balgobind 2009 Blood · I-BFM 11q23 AML partner study
Doctors from eleven countries pooled 756 children with this leukemia and asked whether the partner gene mattered. It mattered enormously: survival ranged from about one in ten to more than nine in ten depending on which partner was involved.
The same thing, for a scientist
This retrospective international study is the source of the partner-based risk classification now used in paediatric AML. Outcome data on 756 children with 11q23/MLL-rearranged AML were collected from 11 collaborative groups with central karyotype review before subgroup assignment. Whole-cohort 5-year EFS was 44% (+/-5), with subgroup values spanning 11% (+/-5) to 92% (+/-5). Multivariable analysis identified four subgroups as independent predictors: t(1;11)(q21;q23) HR 0.1 (P=.004, favourable); t(6;11)(q27;q23) HR 2.2 (P<.001); t(10;11)(p12;q23) HR 1.5 (P=.005); t(10;11)(p11.2;q23) HR 2.5 (P=.005). The authors could not confirm the favourable prognosis of t(9;11)(p22;q23) reported from single-institution data, and concluded that partner screening is needed for treatment stratification at diagnosis.
The numbers, and what is fragile about them
The headline 11%-92% range is the widest partner effect anywhere in this literature 1. It is also the number most vulnerable to small arms: t(1;11) is a rare partner and its HR of 0.1 rests on very few children, which is why the point estimate is so extreme. The reproducible signals are the adverse ones - t(6;11)/KMT2A::AFDN and both t(10;11) breakpoints - and those are the ones that survived into the I-BFM high-risk group 1, 2.
The t(9;11) non-replication
A single-institution series of 298 children had reported t(9;11) as the most favourable genetic factor in AML, with EFS 66%+/-15 for 15 t(9;11) patients versus 25.9%+/-11.2 for the other 27 MLL-rearranged cases 3. With 756 patients and central karyotype review, that advantage did not reproduce 1. This is the single most instructive episode in the partner-outcome literature: the partner claim that was most widely believed is the one that failed to replicate at scale. See epi-partner-mllt3-t9-11-favourability-claim-and-its-non-replication.
| metric | 5-year event-free survival by 11q23 translocation subgroup |
|---|---|
| value | 44 |
| unit | % (whole cohort, +/-5) |
| range low PCT | 11 |
| range high PCT | 92 |
| population | children with 11q23/MLL-rearranged AML |
| n total | 756 |
| not confirmed | t(9;11)(p22;q23) favourable prognosis |
| region | international, 11 collaborative groups |
| year | 2009 |
| prespecified | no - retrospective subgroup analysis |
| source | PMID 19528532 |
independent subgroups — the numbers behind the chart
[
{
"karyotype": "t(1;11)(q21;q23)",
"partner": "KMT2A::MLLT11",
"hr": 0.1,
"p": 0.004,
"direction": "favourable"
},
{
"karyotype": "t(6;11)(q27;q23)",
"partner": "KMT2A::AFDN",
"hr": 2.2,
"p": 0.001,
"direction": "adverse"
},
{
"karyotype": "t(10;11)(p12;q23)",
"partner": "KMT2A::MLLT10",
"hr": 1.5,
"p": 0.005,
"direction": "adverse"
},
{
"karyotype": "t(10;11)(p11.2;q23)",
"partner": "KMT2A::ABI1",
"hr": 2.5,
"p": 0.005,
"direction": "adverse"
}
]- A drug targeting the shared menin-KMT2A interface is being offered to a population whose untreated prognosis already varies from 11% to 92% EFS; single-arm response rates in such a population are close to uninterpretable without partner stratification.
- If the favourable t(1;11) subgroup is real, exposing those children to a novel agent on a class rationale risks trading a 92% outcome for an unknown one.
- How much of the 11%-92% spread survives adjustment for additional cytogenetic aberrations and for RAS-pathway comutation, which were not measured here?
- Are the two t(10;11) breakpoints (MLLT10 at p12, ABI1 at p11.2) genuinely different in outcome, or is the difference an artefact of band-level assignment?
- Would molecular rather than cytogenetic partner assignment narrow or widen the observed range?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Among 756 children with 11q23/MLL-rearranged AML pooled from 11 collaborative groups with central karyotype review, 5-year EFS was 44% (+/-5) overall with subgroup values from 11% (+/-5) to 92% (+/-5); t(1;11)(q21;q23) (HR 0.1, P=.004), t(6;11)(q27;q23) (HR 2.2, P<.001), t(10;11)(p12;q23) (HR 1.5, P=.005) and t(10;11)(p11.2;q23) (HR 2.5, P=.005) were independent prognostic subgroups, and the favourable prognosis of t(9;11)(p22;q23) could not be confirmed.
Novel prognostic subgroups in childhood 11q23/MLL-rearranged acute myeloid leukemia: results of an international retrospective study. · Balgobind BV, Raimondi SC, et al. · Blood 2009What this evidence cannot show
Retrospective; patients were treated on many different protocols across 11 groups and an unstated span of years, so protocol and era are uncontrolled. Subgroups defined by karyotype band rather than by molecular partner, so t(10;11) and t(11;19) each conflate more than one gene. The extreme ends of the 11%-92% range come from small subgroups and the confidence intervals around them are wide. No adjustment for co-occurring mutations, which later work shows carry substantial prognostic weight.
Evidence that argues against this (1)
- A follow-up study by the same collaboration on 733 of these children showed that additional cytogenetic aberrations carry independent prognostic weight of comparable size to the partner (5-year OS 47% with ACA versus 62% without, P<.001; trisomy 8 favourable, trisomy 19 adverse), so part of the between-partner spread may be co-occurring cytogenetics rather than the partner itself. Coenen EA et al. 2011
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The adverse partners identified here (AFDN at 6q27, MLLT10 at 10p12, ABI1 at 10p11.2) were carried forward into the I-BFM high-risk fusion-partner group, which in 1,130 children independently predicted inferior EFS, CIR and OS.
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
Confirmation in an overlapping era and partly overlapping study groups; this is a re-test of an imported classification, not an independent discovery.
Evidence that argues against this (1)
- A 159-patient KMT2A-r AML clinicogenomic series found that KRAS mutation, not partner, was the independent adverse factor (2-year OS 24.6% versus 50.5%, P=.001), with the KMT2A::MLLT4/AFDN subgroup showing only a trend to inferior OS and also carrying the highest KRAS mutation frequency - so an apparent AFDN effect may be a KRAS effect. Wu Y et al. 2025
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
The t(9;11) favourable-prognosis claim that this study failed to confirm came from a single institution: among 42 children with MLL rearrangements, the 15 with t(9;11) had EFS 66%+/-15 versus 25.9%+/-11.2 for the other 27 (P=.004).
Favorable impact of the t(9;11) in childhood acute myeloid leukemia. · Rubnitz JE, Raimondi SC, et al. · J Clin Oncol 2002What this evidence cannot show
Single institution, four consecutive trials, 42 MLL-rearranged patients of whom 15 had t(9;11). Karyotype-based partner assignment. t(9;11) was tightly associated with M5 morphology (21 of 23 cases), so morphology and partner are not separable in this dataset.
Evidence that argues against this (1)
- The 756-patient international study with central karyotype review could not confirm this favourable prognosis, and the HARMONY adult database found OS for t(9;11)/KMT2A::MLLT3 no different from other translocations (P=0.756). Hernández-Sánchez A et al. 2024
source of the partner risk classification that I-BFM re-tested
5-year EFS 11%-92% across 11q23 partner subgroups, n=756
could not confirm the favourable prognosis of t(9;11)/KMT2A::MLLT3
- related to →KMT2A / MLL1 — the gene that breaks
outcome varies by which partner joins the retained KMT2A N-terminus
the 756-patient partner model contained no mutational covariates
- ← related toThe protocol and era confound: the partner datasets span 1983 to 2016 and outcome moved under them
partner effect estimated by pooling 11 collaborative groups with no era adjustment
partner-specific phenotype within one family: TCF3::ZNF384 confers dexamethasone resistance where EP300::ZNF384 does not - the same partner-dependent divergence this cohort measured for KMT2A partners on outcome