COG AAML0531/1031 (n=2,080): MLLT10 fusions are adverse, and the 5-prime partner barely changes it
also called Hylkema 2024 Blood Advances · MLLT10 structural variants paediatric AML
Across two large childhood leukemia trials, any fusion involving the gene MLLT10 went with very poor outcomes. Which gene MLLT10 was joined to made almost no difference, which is a rare case of the partner clearly not mattering.
The same thing, for a scientist
A retrospective transcriptome and karyotype study of 2,080 children and young adults enrolled on COG AAML0531 (NCT00372593) and AAML1031 (NCT01371981) identified 127 patients (6.1%) with MLLT10 fusions: KMT2A::MLLT10 n=104 (81.9%), PICALM::MLLT10 n=13 (10.2%), and X::MLLT10 n=10 (7.9%; DDX3X and TEC twice each, then DDX3Y, CEP164, SCN2B, TREH, NAP1L1 and XPO1 once each). MLLT10-fused patients had 5-year EFS 18.6% versus 49% (P<.001), OS 38.2% versus 65.7% (P<=.001) and relapse risk 76% versus 38.6% (P<.001). Crucially, within MLLT10, the 5-prime partner made little difference: KMT2A::MLLT10 EFS from study entry 19.5% versus 12.7% for other MLLT10 partners (P=.628) and OS 40.4% versus 27.6% (P=.361); PICALM::MLLT10 EFS 9.2% versus 20% for X::MLLT10 (P=.788). Methylation signatures did differ - PICALM::MLLT10 and X::MLLT10 resembled NUP98::NSD1, whereas KMT2A::MLLT10 affected distal regulatory elements - yet outcome did not.
The symmetric experiment
Every other entry in this beat holds KMT2A fixed and varies the 3-prime partner. This one holds MLLT10 fixed and varies the 5-prime gene, including swapping KMT2A out entirely for PICALM, XPO1, DDX3X and others 1. The result is a null: outcome is a property of MLLT10 involvement, not of what is fused to it. That is a partner-independence result of the same shape as the menin-interface class claim, but for the opposite half of the fusion.
Why it cuts against the menin story specifically
PICALM::MLLT10 and the X::MLLT10 fusions do not contain KMT2A at all, so they cannot bind menin through the KMT2A N-terminus, yet they produce indistinguishable clinical outcomes to KMT2A::MLLT10 1. If the disease behaves the same whether or not a menin-binding N-terminus is present, then whatever determines outcome in this subgroup is not the menin interface. The comparison is small (13 and 10 patients) and outcome equivalence is not mechanistic equivalence - but the direction is uncomfortable for a class claim built on menin.
Epigenetics separated where outcome did not
The non-KMT2A MLLT10 fusions carried a DNA hypermethylation signature resembling NUP98::NSD1, while KMT2A::MLLT10 primarily affected distal regulatory elements 1. So the molecular assays could tell the fusions apart and the survival curves could not. That is a useful warning in the other direction too: a demonstrated molecular difference between partners does not license a prognostic claim.
| metric | 5-year event-free survival, MLLT10 fusions versus other paediatric AML, and by 5-prime partner |
|---|---|
| value | 18.6 |
| unit | % (any MLLT10 fusion) |
| comparator value | 49 |
| comparator unit | % (no MLLT10 fusion) |
| population | children and young adults with AML on COG AAML0531 and AAML1031 |
| n total | 2080 |
| n mllt10 | 127 |
| region | United States (Children's Oncology Group) |
| year | 2024 |
| prespecified | no - retrospective analysis of two completed trials |
| source | PMID 38306602 |
five prime partners — the numbers behind the chart
{
"KMT2A::MLLT10": 104,
"PICALM::MLLT10": 13,
"X::MLLT10": 10
}x partners
[ "DDX3X (2)", "TEC (2)", "DDX3Y", "CEP164", "SCN2B", "TREH", "NAP1L1", "XPO1" ]
within mllt10 — the numbers behind the chart
{
"kmt2a_efs_pct": 19.5,
"other_efs_pct": 12.7,
"efs_p": 0.628,
"kmt2a_os_pct": 40.4,
"other_os_pct": 27.6,
"os_p": 0.361,
"picalm_efs_pct": 9.2,
"x_efs_pct": 20,
"p": 0.788
}other endpoints — the numbers behind the chart
{
"os_mllt10_pct": 38.2,
"os_other_pct": 65.7,
"relapse_mllt10_pct": 76,
"relapse_other_pct": 38.6
}trials
[ "NCT00372593", "NCT01371981" ]
- PICALM::MLLT10 and X::MLLT10 patients look identical clinically to KMT2A::MLLT10 patients but have no menin-binding KMT2A N-terminus, so a menin inhibitor offered on clinical resemblance would be given to patients for whom the mechanism cannot apply.
- Conversely, if KMT2A::MLLT10 responds to menin inhibition and the outcome-matched non-KMT2A fusions do not, the trial that shows it will need to enrol 13-patient comparator arms, which no sponsor will power.
- Do PICALM::MLLT10 patients, who have no KMT2A N-terminus, respond to menin inhibition? If they do not while KMT2A::MLLT10 patients do, outcome equivalence with mechanistic divergence would be demonstrated directly.
- Is the within-MLLT10 null anything more than a power failure at n=13 and n=10? A pooled international MLLT10 registry could answer it.
- Does the NUP98::NSD1-like methylation signature in non-KMT2A MLLT10 fusions predict response to the same agents that work in NUP98::NSD1 disease?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Among 2,080 children and young adults on COG AAML0531 and AAML1031, 127 (6.1%) had MLLT10 fusions with 5-year EFS 18.6% versus 49% without (P<.001), OS 38.2% versus 65.7% and relapse risk 76% versus 38.6%; within MLLT10 the 5-prime partner did not significantly change outcome (KMT2A::MLLT10 EFS 19.5% versus 12.7% for other MLLT10 partners, P=.628; OS 40.4% versus 27.6%, P=.361; PICALM::MLLT10 EFS 9.2% versus 20% for X::MLLT10, P=.788), although PICALM::MLLT10 and X::MLLT10 carried a hypermethylation signature resembling NUP98::NSD1 while KMT2A::MLLT10 affected distal regulatory elements.
Structural variants involving MLLT10 fusion are associated with adverse outcomes in pediatric acute myeloid leukemia. · Abla O, Ries RE, et al. · Blood Adv 2024What this evidence cannot show
The within-MLLT10 comparisons run on 13 PICALM::MLLT10 and 10 X::MLLT10 patients, so the P values of .628, .361 and .788 are nulls with almost no power and cannot support an equivalence claim. Retrospective analysis of two trials; fusion detection by transcriptome profiling and/or karyotyping, so ascertainment differs between patients. EFS measured from study entry rather than diagnosis for the within-MLLT10 comparison.
Evidence that argues against this (2)
- In childhood KMT2A-r AML the 3-prime partner does separate outcome strongly (partner-based high-risk group 5-year EFS 30.3% versus 54.0%, P<.0001, n=1,130), so partner independence within MLLT10 is a local finding about one already very adverse partner and not a general result. van Weelderen RE et al. 2023
- In an international 756-patient series, t(10;11)(p12;q23)/KMT2A::MLLT10 and t(10;11)(p11.2;q23)/KMT2A::ABI1 carried different independent hazard ratios (1.5 versus 2.5), showing that partners resolved at sub-band level can separate even when a band-level grouping does not. Balgobind BV et al. 2009
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KMT2A::MLLT10 is one of the five partners in the I-BFM high-risk group for childhood KMT2A-rearranged AML, which is consistent with the very poor MLLT10-associated outcomes seen here.
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 cohorts with likely patient overlap in era and in participating groups; consistency between them is not independent replication.
Evidence that argues against this (1)
- In adult KMT2A-r AML cohorts of 172 and 205 patients, translocation subtype including MLLT10 showed no independent effect on survival, so the MLLT10 adverse signal is not established across age. Issa GC et al. 2021
- related to →KMT2A::MLLT10 (MLL-AF10), t(10;11)(p12;q23)
KMT2A::MLLT10 n=104 of 127 MLLT10 fusions; 5-year EFS 19.5%
MLLT10 fusions 5-year EFS 18.6% versus 49% (P<.001), n=2,080
PICALM::MLLT10 has no KMT2A N-terminus and no menin-binding motif yet matches KMT2A::MLLT10 outcome
- supports →One karyotype band, two partner genes: why t(11;19) and t(10;11) can cancel a real partner effect
a t(10;11) karyotype can be a fusion containing no KMT2A at all