The partner-matched control arm is not 241 versus 219 - it is 153 versus 101, and in that subset the two groups make an identical fusion protein
also called MLLT1 reciprocal-status control arm · 153 vs 101 · ACER1-MLLT1 intergenic breaks versus MLLT1 intron 1 breaks
Published breakpoint data on 302 patients splits this leukaemia into groups finer than anyone has used. Two of those groups - 153 patients and 101 patients - build exactly the same fusion protein and differ only in whether a mirror-image gene also exists, which is as close to a controlled human experiment as this field is ever likely to get.
The same thing, for a scientist
The reciprocal-negative arm of KMT2A::MLLT1 has been framed in this graph as 241 spliced against 219 balanced cases, from the recombinome's 460 [E1]. A 2023 Leukemia paper mined the same Frankfurt archive at finer resolution on 302 t(11;19) patients and reports four breakpoint classes on the MLLT1 side: **50.7% (n=153) break 5' upstream of MLLT1 but 3' downstream of ACER1**; **33.4% (n=101) break within MLLT1 intron 1**; and 15.9% (n=48) break downstream of exon 2, giving partial (6.0%) or complete (9.9%) exclusion of the YEATS domain, with about 5.3% (n=16) breaking between exons 6 and 7 [E2]. The first two classes together are the 84.1% (n=254) that retain the YEATS domain [E2]. **That is the comparison to run.** Both the 153 and the 101 produce a fusion transcript whose partner contribution begins at MLLT1 exon 2 - the intergenic-break class by read-through and splicing to the first bona fide splice acceptor [E1], the intron-1 class by ordinary intronic splicing - so the direct oncoprotein is architecturally the same molecule in both arms, YEATS domain included. Only the 153 are reciprocal-negative. The 48 patients with downstream breaks, who make a genuinely different oncoprotein, are excluded rather than averaged in. Nobody has run this comparison, and the paper that supplies the numbers was asking a different question.
What this comparison would settle
Whether the presence of a reciprocal KMT2A allele changes outcome, HOXA signature, MRD kinetics or menin-inhibitor response when the partner gene, the breakpoint class on the KMT2A side and the direct fusion protein are all held constant. Every published reciprocal-status comparison confounds at least one of those. In t(4;11) the reciprocal-positive and reciprocal-negative groups differ in breakpoint position, in HOXA expression and possibly in the direct product's own composition 3, and the spliced class there is 1 case in 1,262 1 so the comparison cannot be made at all. Pooled analyses across partners 4 confound reciprocal status with partner identity, which is itself prognostic. The 153-versus-101 split has none of those problems.
What it would NOT settle, stated plainly
Four confounds survive, and two of them are serious.
1. Age is untested between these two specific classes. The paper reports that patient age group is not significantly correlated with YEATS-domain retention status (Chi-square p = 0.33, df = 2) 2 - but that tests 254 versus 48, not 153 versus 101. The recombinome separately shows that KMT2A-side breakpoint preference changes with age at about six months 1, and MLLT1 is 52.6% infant 1. If intergenic breaks are commoner in infants than intron-1 breaks, the comparison is an age comparison. This is the first thing to check and it is one cross-tabulation.
2. Lineage is associated with breakpoint position in this partner. Significantly fewer AML than ALL patients retain the YEATS domain (Chi-square p < 0.0001, df = 1) 2. Again that is the 254/48 split rather than 153/101, but it establishes that MLLT1 breakpoint position is not lineage-neutral, and lineage flips this partner's prognostic sign 5.
3. MLLT1 gene dosage may differ. In the intergenic-break class the MLLT1 coding unit is never interrupted, so a second intact copy travels to der(11); in the intron-1 class one MLLT1 allele is split. Whether MLLT1's own promoter accompanies the gene is not resolved by 'upstream of exon 1'. ENL is a selective dependency in its own right 6.
4. Reciprocal-capable is not reciprocal-expressing. An intron-1 break permits a reciprocal allele; it does not guarantee an in-frame, transcribed one. In the recombinome's complex-rearrangement set, 386 of 426 reciprocal fusions were out of frame or non-functional at the DNA level 1. Assigning the 101 to a 'reciprocal-positive' arm on breakpoint position alone will mislabel an unknown fraction, and 3' KMT2A deletion supplies a second route to the same error 7.
What data would be needed and who holds it
All of it already exists in one place. The 302-patient breakpoint dataset was provided by R. Marschalek of the Diagnostic Center of Acute Leukemia (DCAL), Institute of Pharmaceutical Biology, University of Frankfurt, who is a named author on the 2023 Leukemia paper and its stated source of chromosomal breakpoint data and analysis 2; the same archive underlies the 460-case recombinome row 1. The recombinome states that the database offers the possibility to analyse specific patient cohorts and that the authors will share these data with requesting scientists 1. The analysis needed is: (i) cross-tabulate breakpoint class against age group and lineage for the 254 YEATS-retaining cases; (ii) link the 153 and 101 to outcome through the national study groups that referred them; (iii) where RNA exists, confirm reciprocal transcript status per patient rather than inferring it. Steps (i) and (iii) require no new consent and no new sample.
Why this changes the arithmetic already in the graph
The graph's existing entry frames the arm as 241 of 460 1. The two datasets are consistent - 52.4% against 50.7% for the intergenic class, in cohorts of 460 and 302 drawn from the same archive - but the 302-patient version is the more useful one because it separates the intron-1 balanced cases (101) from the downstream-break cases (48), and only the first is a matched comparator. Pooling all 219 non-spliced cases into one arm mixes 101 patients who make the same oncoprotein as the index arm with 48 who make a different one. The headline number for this beat should therefore be 153 versus 101, with 48 excluded and stated. This entry does not withdraw the 241/219 figures; it says which question each pair answers.
| status | planned |
|---|---|
| result summary | Re-derives the partner-matched reciprocal-status comparison from published breakpoint classes. Proposed arms: 153 intergenic-break (ACER1-MLLT1, reciprocal-negative) versus 101 MLLT1-intron-1 (reciprocal-capable), both producing a YEATS-containing MLL-ENL from an exon-2 junction; 48 downstream-break patients excluded because their direct oncoprotein differs. Prior framing of 241 versus 219 mixes the 48 into the comparator arm. |
| data holder | Diagnostic Center of Acute Leukemia (DCAL), Institute of Pharmaceutical Biology, University of Frankfurt (R. Marschalek) |
| first query needed | 2x2 of breakpoint class (intergenic vs MLLT1 intron 1) against age group, and against lineage, within the 254 YEATS-retaining cases |
| power note | With arms of 153 and 101 and a binary outcome around 50%, this design has roughly 80% power to detect an absolute difference of about 19 percentage points at alpha 0.05 two-sided (this graph's calculation, not published). It is not powered for a small effect, and a null result should be reported as a bounded null rather than as no difference. |
inputs
[ "PMID 36435883 breakpoint classes, n = 302 t(11;19) patients", "PMID 37019990 Table 1 and spliced-fusion counts, n = 460 KMT2A::MLLT1 patients", "DCAL Frankfurt archive (not public; authors state they will share on request)" ]
arms — the numbers behind the chart
{
"reciprocal_negative_intergenic": 153,
"reciprocal_capable_intron1": 101,
"excluded_downstream_break": 48,
"cohort_total": 302
}matched on
[ "fusion partner gene", "MLLT1 exon-2 junction", "YEATS domain retention", "direct oncoprotein architecture" ]
unmatched confounds
[ "age group (untested for this split)", "lineage (associated with breakpoint position in this partner)", "wild-type MLLT1 dosage", "reciprocal-capable is not reciprocal-expressing" ]
- If this comparison comes back null, the graph's reciprocal-remainder hypothesis loses the one population that could have supported it with the partner held constant, and the honest conclusion would be that the t(4;11) reciprocal-status association is a proxy for breakpoint position or HOXA status rather than for the reciprocal product.
- If it comes back positive, the immediate clinical question is whether menin inhibitors should be dosed or combined differently in reciprocal-positive patients - and the reciprocal product carries no menin-binding motif, so a menin inhibitor is the wrong tool for it. The therapeutic implication of a positive result is a combination, not a dose change.
- Either result is only as good as the reciprocal-status label. A breakpoint-derived label misclassifies patients with 3' KMT2A deletion and patients whose reciprocal junction is out of frame, both in the direction of diluting a true effect.
- Are the 153 and the 101 balanced for age and lineage? The chi-square tests published for this cohort are on the 254-versus-48 YEATS split and say nothing about the split that matters here. Until this is answered the comparison is not known to be clean.
- Do the 153 and 101 have linkable outcome data? The breakpoint archive is diagnostic; the outcomes sit with the national study groups that referred the samples, and whether the link exists is not stated in either paper.
- Does the intergenic break class carry a second functional MLLT1 transcription unit? That question is answerable from the breakpoint sequences already in the archive and would decide whether the arms are matched for wild-type ENL dosage.
- Is there an equivalent fine-grained split for any other KMT2A partner? EPS15 (12 of 51 spliced) and ELL (10 of 119) are the only other partners with a double-figure spliced class, and both are far too small; if the MLLT1 comparison is null, there is no second population in which to check it.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Read in full text (PMC10169636): of 460 KMT2A::MLLT1 cases, about 50% (n = 241) of breakpoints localise upstream of MLLT1 exon 1 and only direct and no reciprocal transcripts are expressed; in spliced fusions the most 3' KMT2A exon splices to the partner's exon 2, the first exon with a bona fide splice acceptor; AFF1 has 1 spliced case in 1,262; of 426 complex rearrangements 40 reciprocal fusions were in frame and 386 were out of frame or non-functional at the genomic DNA level; MLLT1 cases are 242 infant, 129 paediatric, 84 adult and 5 unclassified of 460; KMT2A breakpoint preference in ALL switches from intron 11 to intron 9 at about six months of age; and the authors state the database offers the possibility to analyse specific patient cohorts and that they will share these data with requesting scientists.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023What this evidence cannot show
Read in full text via Europe PMC fullTextXML (body element present). Single-centre referral series, 2003-2022. The 241 figure is a count of breakpoint positions with the transcript consequence inferred from gene structure, not a per-patient reciprocal-transcript assay.
Evidence that argues against this (1)
- The two cohorts are not independent - 460 and 302 are both drawn from the Frankfurt archive - so agreement between 52.4% and 50.7% is reproducibility of a database extract rather than replication in a second population. Hu H et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Read in full text (PMC11246743): from breakpoint data on a clinical cohort of 302 t(11;19) patients, 50.7% (n = 153) harbour genomic breaks 5' upstream of the MLLT1 gene but 3' downstream of ACER1, another 33.4% (n = 101) harbour breaks within the first MLLT1 intron, so 84.1% (n = 254) retain the YEATS domain in the resulting MLL-ENL fusion protein, while the remaining 15.9% (n = 48) break downstream of exon 2 giving partial inclusion (6.0%) or exclusion (9.9%) of the YEATS domain, with about 5.3% (n = 16/302) breaking between exons 6 and 7; patient age group is not significantly correlated with YEATS-domain retention status (Chi-square p = 0.33, df = 2) while significantly fewer AML than ALL patients retain the YEATS domain (Chi-square p < 0.0001, df = 1); the YEATS domain is coded from ENL exons 2-4 (amino acids 5-140); and the chromosomal breakpoint data and analysis were provided by R. Marschalek of the Diagnostic Center of Acute Leukemia, University of Frankfurt.
The ENL YEATS epigenetic reader domain critically links MLL-ENL to leukemic stem cell frequency in t(11;19) Leukemia. · Hu H, Saha N, Yang Y, Ahmad E, et al. · Leukemia 2023What this evidence cannot show
Read in full text by the FOURTH route only: Europe PMC fullTextXML returned HTTP 500 and efetch db=pmc returned HTTP 429 with no body, and the PMC HTML page at https://pmc.ncbi.nlm.nih.gov/articles/PMC11246743/ returned 61,379 characters of text. Europe PMC reports isOpenAccess N for this record. The breakpoint analysis is a re-mining of a previously published cohort, not a new ascertainment. The paper never uses the words spliced, reciprocal or der(19); the reciprocal-status reading of its 153/101 split is this graph's inference from combining it with the recombinome's mechanism description. Crucially, the two chi-square tests reported are on the 254-versus-48 YEATS split, NOT on the 153-versus-101 split, so they do not establish that the two arms of the proposed comparison are balanced for age or lineage.
Evidence that argues against this (1)
- The same paper shows that the two classes are NOT interchangeable for everything: breaks that keep or lose the YEATS domain give oncoproteins with different in vivo leukaemogenicity and different leukemic stem cell frequency, so breakpoint position in this partner has real biological consequences independent of reciprocal status, and there is no guarantee the 153 and 101 are alike in ways the paper did not measure. Hu H et al. 2023
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the Interfant cohort, HOXA cluster genes were expressed exclusively in the reciprocal-expressing t(4;11) patients, who also had markedly better event-free and overall survival.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, Roy A, et al. · Haematologica 2019What this evidence cannot show
Abstract-level reading for this entry. n = 43 for the t(4;11) subgroup. The collinearity of HOXA status with reciprocal status means the two cannot be separated in this cohort, which is precisely the defect the MLLT1 comparison is meant to avoid.
Evidence that argues against this (1)
- A partner-stratified paediatric cohort reporting outcome by KMT2A partner measured neither HOXA nor reciprocal status, so the collinearity has no independent check. Attarbaschi A et al. 2023
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a paediatric cohort of 90 MLL-rearranged acute leukaemia patients, 37.8% (34/90) displayed a single direct MLL fusion gene, 15.6% (14/90) carried a single reciprocal fusion and 27.8% (25/90) had both reciprocal MLL fusion alleles, and children with reciprocal MLL fusions showed relatively favourable outcomes compared with children with complex translocations or a single direct MLL fusion allele (66.1% versus 24.6% and 27.6%, P = 0.001).
Relatively favorable prognosis for MLL-rearranged childhood acute leukemia with reciprocal translocations. · Yang L, Ding L, Liang J, Chen J, et al. · Pediatr Blood Cancer 2018What this evidence cannot show
Abstract-level reading. This is the only clinical outcome analysis in the literature stratified by reciprocal-fusion status, and it POOLS all KMT2A partners, so partner identity and reciprocal status are confounded - the exact confound the MLLT1 comparison removes. n = 90 with 77 treated, chemotherapy era, median follow-up 2.5 years, no MLLT1 breakout, no drug-response endpoint.
Evidence that argues against this (1)
- A functional study in human t(4;11) lines found the reciprocal product dispensable: loss of AF4-MLL had no effect on growth of RS4;11 or SEMK2-M1 cells, while knockdown of MLL-AF4 strongly inhibited both. Kumar AR et al. 2011
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 1,256 children with KMT2A-rearranged AML, the 19p13.3/KMT2A::MLLT1 group (n = 56) was independently associated with inferior event-free survival (HR 1.6, P = .01) and higher cumulative incidence of relapse (HR 1.8, P < .01), while 19p13.1/KMT2A::ELL (n = 75) showed no significant association on any endpoint, and the median age at diagnosis in the MLLT1 group was 6.8 years against under 4 years for most other KMT2A-r groups.
Optimized cytogenetic risk-group stratification of KMT2A-rearranged pediatric acute myeloid leukemia. · van Weelderen RE, Harrison CJ, Klein K, Jiang Y, et al. · Blood Adv 2024What this evidence cannot show
Numbers relayed from a full-text reading performed by a delegated worker in this session and not independently re-read by the author of this entry; the PMID, title and DOI were independently re-fetched from PubMed esummary on 2026-09-19. Retrospective, cytogenetics-based, so breakpoint class and reciprocal status were not and could not have been measured.
Evidence that argues against this (1)
- In childhood non-infant T-ALL, t(11;19)(q23;p13.3) had the highest 5-year EFS in a 629-patient cohort at 91.2% (n = 34), the opposite prognostic direction, so the adverse AML finding does not generalise across lineage. Attarbaschi A et al. 2023
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
ENL, but not its paralogue AF9, is required in acute myeloid leukaemia: CRISPR-Cas9 depletion of ENL is anti-leukaemic while AF9 depletion is not.
ENL links histone acetylation to oncogenic gene expression in acute myeloid leukaemia. · Wan L, Wen H, Li Y, Lyu J, et al. · Nature 2017What this evidence cannot show
This paper concerns WILD-TYPE ENL in MLL-rearranged AML cell lines (MOLM-13, MV4;11), not the KMT2A::MLLT1 fusion protein. The dosage argument in this entry is an inference from it.
Evidence that argues against this (1)
- The companion paper states explicitly that it reports on wild-type ENL rather than the fusion, so neither 2017 Nature paper tests the MLL-ENL chimera and neither can support a fusion-specific dosage claim. Erb MA et al. 2017
- E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a T-ALL with a cytogenetically cryptic t(11;19), MLL sequences distal to the breakpoint were deleted from the genome, precluding formation of a reciprocal ENL/MLL fusion protein - a route to reciprocal-negativity independent of breakpoint position within MLLT1.
Complex MLL rearrangement in a patient with T-cell acute lymphoblastic leukemia. · Chervinsky DS, Sait SN, Nowak NJ, et al. · Genes Chromosomes Cancer 1995What this evidence cannot show
Abstract only; no PMCID in Europe PMC. n = 1, 1995 methods. The frequency of this mechanism in KMT2A::MLLT1 is unknown and no series has counted it.
Evidence that argues against this (1)
- The recombinome catalogues the converse case - deletion of 5'-KMT2A with reciprocal fusion genes retained - in only two patients out of 426 complex rearrangements, suggesting deletion-driven misclassification in either direction is uncommon, though the paper gives no MLLT1-specific count. Meyer C et al. 2023
the partner-matched, direct-product-matched comparison the hypothesis requires
refines 241 vs 219 to 153 vs 101; the 48 downstream-break cases make a different oncoprotein and should not be in the comparator arm
breakpoint-class structure of this fusion's patient population
the arm assignment rests on the spliced-fusion mechanism holding
lineage flips this partner's prognostic sign, so lineage must be balanced across the arms
one documented leak in the reciprocal-negative arm
one documented leak in the reciprocal-positive arm
supplies the age denominators the control arm must be balanced against
the measured absence the proposed analysis would fill
why the comparison must be restricted to exon-2 junctions
the in vitro/in vivo counterpart of the human comparison, and it has never been run either
the protocol that would run the comparison
the patient comparison has no in vitro counterpart, and on breakpoint priors two or three lines should already be reciprocal-negative
the control-arm population is 88% lymphoid, where partner assignment is reliable
the reader is retained in both arms, so YEATS status is not a confounder in the 153-versus-101 comparison
- ← related tot(9;11) is the reciprocal-POSITIVE arm: 2 of 612 patients are spliced, against 241 of 460 for t(11;19)
cross-partner denominator for the within-partner test