Oncofusion

start / graph / epi-kmt2a-mllt1-outcome-by-age-band-and-the-infant-hole

Epidemiologymoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q26b-mllt1-control-arm

KMT2A::MLLT1 outcome by age band: favourable in adult ALL, high-risk in childhood AML, and completely unstratified in the infants who are most of the disease

also called MLL-ENL survival by age · t(11;19) prognosis infants · KMT2A::MLLT1 outcome adults

In adults and older children with lymphoblastic leukaemia this fusion carries a comparatively good outlook, and in children with myeloid leukaemia a bad one. In babies under one - who are more than half of all patients with this fusion - the two largest trials ever run did not report an outcome figure for it at all.

The same thing, for a scientist

This entry extends rather than challenges the graph's existing partner-prognosis entry, which is built entirely on non-infant cohorts. Adding the age bands it does not cover: (1) In a consecutive Italian series, KMT2A::MLLT1 was found in 4/150 (2.6%) paediatric, 8/993 (0.8%) adult and 0/70 elderly ALL patients; across all 12, CR was achieved in 11 (92%) and 48-month overall and event-free survival were 73.3% and 67%, with the authors concluding that in adults too this fusion marks a rare entity with a favourable prognosis. (2) In 20 KMT2A-rearranged adult AML patients, MLLT1 and MLLT3 partners derived better benefit from allogeneic transplant than other partners. (3) For infants under one year - the modal patient for this fusion - Interfant-99 (n=482, 4-year EFS 47.0%) and Interfant-06 (n=651, 6-year EFS 46.1%, OS 58.2%) both report the prognostic model at the level of 'KMT2A rearrangement' and do not report a partner-stratified outcome. That is an absence of testing, not a negative result, and it is the largest gap in this fusion's clinical description.

Go deeper
What this entry adds and what it does not touch

The graph already holds epi-partner-mllt1-prognostic-sign-flips-between-aml-and-all, which covers childhood AML (I-BFM), non-infant childhood ALL and a cross-age B-ALL analysis. Nothing here contradicts it. The point of this entry is that all three of those cohorts are non-infant or age-mixed, while more than half of KMT2A::MLLT1 patients are infants. So the sign-flip finding is real and is bounded to bands that exclude the modal patient.

Two bands are added below: adult/paediatric ALL treated as one series, and adult AML at transplant. One band is documented as empty.

Adult and paediatric ALL: favourable, on twelve patients

A consecutive Italian series screened 150 paediatric, 993 adult and 70 elderly ALL patients and found the MLL/ENL fusion in 4 (2.6%), 8 (0.8%) and 0 respectively 1. Eight of the twelve had WBC >50 x 10^9/L. Immunophenotype was B-precursor in 7 and T-precursor in 3 of the 10 evaluable.

Outcome across all twelve: CR in 11/12 (92%), and at 48 months overall survival 73.3% and event-free survival 67% 1. The authors' conclusion is that 'also in adults, the MLL/ENL fusion identifies a rare leukemic entity with a favorable prognosis'.

A second finding in the same series bears on this beat directly. In the five patients with parallel RT-PCR at CR, MLL/ENL transcript status did not correlate with outcome: both PCR-negative patients relapsed and only one of three PCR-positive patients did, and transcript was detected repeatedly during follow-up in five long survivors without a relapse following. The authors propose an MLL/ENL-expressing preleukaemic stem cell analogous to RUNX1::RUNX1T1 1. If that is right, the fusion transcript is not a reliable measure of residual disease in this partner group - which also constrains any attempt to measure reciprocal-allele status from remission material.

Adult AML: a partner effect at transplant, on twenty patients

Twenty KMT2A-rearranged adult AML patients were characterised by RT-PCR and sequencing against a control cohort of 138 AML patients. The KMT2A-rearranged cohort's overall survival resembled intermediate-risk AML, and MLLT1 and MLLT3 partners showed better benefit from allogeneic transplant than other partners; expression of the KMT2A 3' region, EVI1 and FLT3 associated with survival in transplanted patients 2.

Twenty patients across more than eighty possible partners means the MLLT1 subgroup here is a handful of people. This is recorded because it is the only adult-AML partner-stratified transplant observation located, not because it can carry weight.

Infants: the band with no number

Interfant-99 enrolled 482 infants under one year on a hybrid ALL/AML regimen; 4-year EFS was 47.0% (SE 2.6, 95% CI 41.9-52.1), and the independent prognostic factors reported were 'all types of rearrangements in the MLL gene', very high WBC, age under 6 months and poor prednisone response 3. Partner identity is not among them, and the phrase 'all types' is the study's own statement that it did not separate them.

Interfant-06 enrolled 651 infants; 6-year EFS was 46.1% (SE 2.1) and OS 58.2% (SE 2.0), and 'KMT2A rearrangement was the strongest prognostic factor for EFS, followed by age, WBC count, and prednisone response' 4. Again the model operates at the level of rearranged-versus-germline.

This is an absence of testing, not a negative result. Neither trial reports having compared partners and found them equivalent; both report the rearrangement as a single covariate. A reader who takes 'partner did not appear in the model' as evidence that partner does not matter in infants is reading an untested variable as a null result, which is the specific error this project has flagged before with Interfant-06's breakpoint analysis.

One practical consequence: because KMT2A::MLLT1 is a minority partner in infant ALL and the two Interfant cohorts are the only ones large enough to support a partner-stratified survival curve for it, the infant number does not exist anywhere and cannot be assembled from published data.

Where the reciprocal question sits in this

For t(4;11), reciprocal expression has a large measured effect on infant outcome: AF4-MLL was expressed in only 19 of 43 t(4;11)+ patients, and those patients had 4-year EFS 62.4% versus 11.7% (P=0.001) and OS 73.7% versus 25.2% (P=0.016), retaining significance in a Cox model adjusted for Interfant-06 risk stratification 5.

That is the size of effect the 153-versus-101 KMT2A::MLLT1 comparison is powered to look for, and it is why the missing infant number matters: the t(4;11) result was obtained in infants, and the KMT2A::MLLT1 arm's cleanest natural experiment sits in a band where no partner-stratified outcome has been published at all.

The direction is also worth noting because it is the opposite of what an 'enforceable remainder' model would predict: in t(4;11), the reciprocal-positive patients did better.

Facts and figures
metricsurvival by age band and lineage
fusionfusion-kmt2a-mllt1
relation to existing entryextends epi-partner-mllt1-prognostic-sign-flips-between-aml-and-all by adding the adult ALL, adult AML and infant bands; does not contradict it
t411 reciprocal expression comparator
p0.001p: 0.001efs_4y_negative_pct11.7efs_4y_negative_pct: 11.7n43.0n: 43efs_4y_positive_pct62.4efs_4y_positive_pct: 62.4
bands
[
  {
    "band": "paediatric + adult ALL, pooled series",
    "n": 12,
    "cr_pct": 92,
    "os_48mo_pct": 73.3,
    "efs_48mo_pct": 67,
    "source_pmid": "21953510"
  },
  {
    "band": "adult AML at allogeneic transplant",
    "n_cohort": 20,
    "finding": "MLLT1 and MLLT3 partners derived better benefit from allo-HSCT",
    "source_pmid": "29384595"
  },
  {
    "band": "infant ALL (<1 year)",
    "n_interfant99": 482,
    "efs_4y_pct_whole_cohort": 47,
    "n_interfant06": 651,
    "efs_6y_pct_whole_cohort": 46.1,
    "os_6y_pct_whole_cohort": 58.2,
    "partner_stratified_outcome_reported": false
  }
]
prevalence in all screening
{
  "paediatric": "4/150 (2.6%)",
  "adult": "8/993 (0.8%)",
  "elderly": "0/70"
}
t411 reciprocal expression comparator — the numbers behind the chart
{
  "efs_4y_positive_pct": 62.4,
  "efs_4y_negative_pct": 11.7,
  "p": 0.001,
  "n": 43
}
unreadable after four routes
[
  "31283407"
]
How this could bite us back
  • If KMT2A::MLLT1 is genuinely favourable in lymphoblastic disease, then adding a targeted agent to front-line therapy in this subgroup risks trading a good outcome for added toxicity, and the trial would need a non-inferiority design rather than a response-rate endpoint.
  • The one series that looked found that the fusion transcript did not track relapse, so an MRD-guided intensification strategy keyed to MLL/ENL PCR could intensify treatment in patients who are cured and spare patients who are not.
  • Every partner-stratified outcome statement for this fusion comes from a cohort that excludes infants, who are the majority of patients. A risk-stratification rule built on those numbers would be applied mostly to people it was never derived in.
Still open
  1. What is the 6-year EFS of infants with KMT2A::MLLT1 on Interfant-06? The data exist in the trial database, the partner was recorded at diagnosis, and no published figure gives it.
  2. Does the adult-ALL favourability survive a larger series? Twelve patients with 73.3% 48-month OS is a striking number and the confidence interval on it is wide enough to include the whole-cohort average.
  3. Is the MLL/ENL transcript really a poor MRD marker, or was the discordance in five patients chance? If the preleukaemic-stem-cell interpretation is right, it would mean reciprocal-allele status cannot be assayed from remission marrow in this fusion.
  4. Why is the elderly count zero in 70 patients? If KMT2A::MLLT1 genuinely disappears above a certain age while KMT2A::AFF1 does not, that is a cell-of-origin statement and it is testable in any large adult ALL registry.
Every source behind this entry (5)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In consecutive screening of 150 paediatric, 993 adult and 70 elderly ALL patients, KMT2A::MLLT1 was found in 4 (2.6%), 8 (0.8%) and 0 respectively; across all 12 patients CR was 11/12 (92%) and 48-month overall and event-free survival were 73.3% and 67%; MLL/ENL transcript status at CR did not correlate with outcome.

    PMID 21953510DOI
    What this evidence cannot show

    Twelve patients, single country, diagnosed across a decade of changing protocols, with no comparator arm. Survival estimates on n=12 have confidence intervals the paper does not print. The MRD sub-analysis rests on five patients with parallel RT-PCR. Read from abstract only in this session. The cohort is dominated by adults, so the 'favourable prognosis' conclusion is a statement about adult ALL and not about infants.

    Evidence that argues against this (1)
    • In childhood KMT2A-rearranged AML the same partner sits in the high-risk group, so 'favourable' is lineage-conditional and does not generalise across the disease. van Weelderen RE et al. 2024
  2. E2weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    Among 20 KMT2A-rearranged adult AML patients characterised against 138 AML controls, overall survival resembled intermediate-risk AML and the MLLT1 and MLLT3 partners showed better benefit from allogeneic haematopoietic stem-cell transplant; KMT2A 3' region, EVI1 and FLT3 expression associated with survival in transplanted patients.

    MLL-rearranged acute myeloid leukemia: Influence of the genetic partner in allo-HSCT response and prognostic factor of MLL 3' region mRNA expression. · Burillo-Sanz S, Morales-Camacho RM, Caballero-Velazquez T, et al. · European Journal of Haematology 2018
    PMID 29384595DOI
    What this evidence cannot show

    Twenty patients split across multiple partner genes, so the MLLT1 subgroup is a handful; no multiplicity adjustment is described in the abstract; transplant allocation was not randomised and partner may proxy for lineage, age and fitness. Read from abstract only.

    Evidence that argues against this (1)
    • A far larger paediatric AML analysis places 19p13.3/KMT2A::MLLT1 in the high-risk partner group with 5-year EFS 30.3% versus 54.0%, which is not the profile of a partner that responds especially well to intensification. van Weelderen RE et al. 2024
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Interfant-99 enrolled 482 infants under one year with ALL and reported 4-year event-free survival of 47.0% (SE 2.6, 95% CI 41.9-52.1); the independently prognostic factors were all types of MLL rearrangement, very high white cell count, age under 6 months and poor prednisone response - partner identity was not separated.

    PMID 17658395DOINCT00015873
    What this evidence cannot show

    Read from abstract only; the full text was not retrieved this session, so it cannot be excluded that a partner-stratified table appears in the paper without reaching the abstract. What can be said is that partner identity is not among the reported independent prognostic factors and that the rearrangement was modelled as a single category. Median follow-up 38 months.

    Evidence that argues against this (1)
    • A partner-stratified analysis does exist for non-infant childhood ALL and finds a large difference by partner, so pooling partners in infants is a modelling choice rather than a demonstrated equivalence. Attarbaschi A et al. 2023
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Interfant-06 enrolled 651 infants under one year with ALL, with 6-year event-free survival 46.1% (SE 2.1) and overall survival 58.2% (SE 2.0); KMT2A rearrangement was the strongest prognostic factor for EFS, followed by age, white cell count and prednisone response, with no partner-stratified outcome reported.

    PMID 31283407DOI
    What this evidence cannot show

    Read from abstract only, and this paper is unreadable by all four PMC routes plus the publisher landing page (Europe PMC gives no PMCID; the DOI resolves to an ASCO page returning HTTP 403). A partner-stratified table could exist in the full text or supplement without appearing in the abstract, so this is a bounded statement about the abstract and the reported prognostic model, not a verified statement about the whole paper.

    Evidence that argues against this (1)
    • The Interfant-06 MRD companion paper stratifies outcome finely by end-of-induction and end-of-consolidation MRD and by consolidation type, showing the cohort can support fine stratification when the investigators choose a variable - which makes the absence of a partner split a choice rather than a power limit. Stutterheim J et al. 2021
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 124 infant B-ALL cases diagnosed and treated on Interfant-99/06, the reciprocal fusion AF4-MLL was expressed in only 19 of 43 (45%) t(4;11)-positive patients, and those patients had significantly better 4-year event-free survival (62.4% versus 11.7%, P=0.001) and overall survival (73.7% versus 25.2%, P=0.016), retaining significance in a Cox model adjusted for Interfant-06 risk stratification.

    PMID 30679323DOI
    What this evidence cannot show

    This is t(4;11)/KMT2A::AFF1, not KMT2A::MLLT1, and reciprocal expression is measured by transcript rather than by genomic reciprocal-allele status - a patient can carry the allele and not express it, so the comparison is expression-positive versus expression-negative, not reciprocal-positive versus reciprocal-negative. n=43 with 19 positive. Read from abstract only.

    Evidence that argues against this (1)
    • Reciprocal-allele status in KMT2A::MLLT1 is determined by breakpoint geography rather than by transcript expression, since half of these patients break upstream of the partner gene and cannot form a reciprocal fusion at all, so the two situations are not measured on the same scale. Hu H et al. 2023
Connected to