Twin concordance: near-total for infant KMT2A-rearranged leukaemia, about 5% for childhood ALL overall
also called monozygotic twin concordance · intraplacental metastasis · twin leukaemia
When one identical twin baby gets KMT2A-rearranged leukaemia, the other almost always gets it too, usually within weeks. For older children with other kinds of leukaemia the co-twin is affected only about one time in twenty, and that gap is the single strongest clue that the KMT2A fusion does almost all of the work by itself.
The same thing, for a scientist
Concordant leukaemia in monozygotic twins is caused by intraplacental spread of a single clone through the shared monochorionic circulation, not by shared germline risk: Ford et al. showed that each of three infant twin pairs carried the *same unique, non-constitutive* KMT2A (HRX) rearrangement, which can only mean one founding cell in one fetus seeded the co-twin [E1]. The quantitative contrast is the informative part. Across childhood ALL as a whole the concordance rate in monozygotic twins is approximately 5%, which forced the field to posit obligatory postnatal second events [E2][E3]. In infant KMT2A-rearranged leukaemia the rate was believed to be essentially 100% with very short latency, and the first published exception did not appear until 2009 [E4]. A near-unity penetrance in the twin setting says that in a fetal haematopoietic cell of the right type, the KMT2A fusion is close to sufficient. That is simultaneously the best argument for targeting the fusion machinery and the worst news for any strategy that depends on a second, druggable cooperating lesion.
Concordance is metastasis, not heredity1 deeper
The clonal-identity argument is airtight and does not rely on statistics. The twins share a genomic breakpoint that is absent from their own non-leukaemic tissue, so the rearrangement is somatic and arose once 1. Vascular anastomoses in a single monochorionic placenta then carry the clone from the index fetus to the co-twin. Greaves et al. review the full case series, including dizygotic and dichorionic pairs, and show the pattern tracks chorionicity rather than zygosity, exactly as the transplacental model predicts 2. Concordance in twins therefore measures the probability that a transplanted preleukaemic clone progresses to disease in a second host, which is a cleaner readout of fusion sufficiency than anything available in singletons.
Why the 5% figure is not the number that matters here
The widely quoted ~5% concordance is dominated by ETV6::RUNX1 and other non-infant childhood ALL, where a long preleukaemic latency and a requirement for postnatal secondary hits are well established 23. Applying that figure to infant KMT2A-r disease is a category error; the correct comparison is infant KMT2A-r twins specifically, where concordance is far higher and latency far shorter.
The exception that proves the clone can be cleared
Chuk et al. reported the first discordant monozygotic infant pair: twin A presented at 9 months with KMT2A::MLLT1 (MLL-ENL) ALL, while twin B carried a demonstrable clonal KMT2A::MLLT1-positive population in marrow and blood that subsequently disappeared, temporally associated with virus-induced cytopenias, and twin B remained healthy three years later 4. This is the cleanest published human evidence that a KMT2A-fusion-positive clone is not an irreversible commitment and can be eliminated, plausibly immunologically. It is a single case and the immune mechanism is inferred rather than demonstrated, but it is the observation that most directly answers the PI's question of whether fusions are 'in flux'.
Subclonal diversification already begins before birth
Twin pairs also bound when diversification starts. Kempski et al. described concordant infant acute monoblastic leukaemia in which the two twins shared a common stem line but had extensively divergent chromosomal subclones, so chromosomal instability was imposed early and evolution can be partitioned into pre- and postnatal phases 5. Kotecha et al. similarly found a trisomy X subpopulation in only one of a t(1;11) concordant pair 6. Intraclonal diversity is therefore present at or near birth even in a disease whose coding mutation burden at diagnosis is near zero.
| metric | monozygotic twin concordance |
|---|---|
| value | 5 |
| unit | percent |
| population | childhood ALL overall (all subtypes) |
| region | pooled case series |
| year | 2003 |
| source | Greaves et al., Blood 2003 (PMID 12791663); contrast with near-100% believed for infant KMT2A-rearranged pairs, first exception PMID 19411627 |
| notes | No formally estimated, denominator-based concordance rate exists for infant KMT2A-rearranged twins; the near-100% figure is a case-series impression subject to strong publication bias toward concordant pairs. |
- Near-unity penetrance in the twin setting argues the fusion is close to sufficient, which means there is probably no cooperating lesion to co-target and combination strategies for infants will have to combine against the same node rather than against two independent nodes.
- If spontaneous immune clearance of fusion-positive clones happens [E4], then any therapy that is profoundly lymphodepleting in infants may be removing the one endogenous control on residual preleukaemic cells.
- There is no denominator-based concordance estimate for infant KMT2A-r twins. A registry-linkage study (national twin registries crossed with childhood cancer registries) could produce one, and it would almost certainly come in below 100% because discordant pairs are not written up.
- Should the healthy co-twin of an infant KMT2A-r case be screened, and if a clone is found, what should be done? Chuk et al. show the clone can clear on its own [E4], so surveillance may be the correct action and treatment may be harm. This is a live clinical-ethics question with no guideline.
- If immune clearance is real, which immune compartment does it? A fetal or neonatal NK/T-cell response against a fusion-positive clone would be the single most exploitable finding in this entire beat, and it currently rests on one case report.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Concordant infant twin leukaemias share a unique, non-constitutive KMT2A rearrangement, proving a single in-utero founding clone spread between fetuses.
In utero rearrangements in the trithorax-related oncogene in infant leukaemias · Ford AM, Ridge SA, Cabrera ME, Mahmoud H, Steel CM, Chan LC, Greaves M · Nature 1993What this evidence cannot show
Three twin pairs; rearrangements characterised by Southern blot rather than nucleotide-level sequencing, so 'unique' means an identical non-germline restriction pattern rather than an identical base-resolution junction.
Evidence that argues against this (1)
- Base-resolution sequencing shows 'a single founding clone' is true only of the initiating lesion. Whole-genome sequencing of leukaemic cells from two monozygotic twin pairs with ALL found that SHARED prenatal coding single-nucleotide variants were limited to the putative initiating lesions; every other non-synonymous variant was distinct between the two tumours and therefore secondary and postnatal, on a background of non-coding changes that were almost entirely discordant. A shared non-germline restriction pattern on a Southern blot cannot see any of that, so the 1993 result establishes a shared initiating event and not a shared leukaemia. Ma Y et al. 2013
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review
Concordance in twin leukaemia is explained by intraplacental clonal spread, and the modest overall concordance rate in childhood ALL implies additional postnatal events are required.
Leukemia in twins: lessons in natural history · Greaves MF, Maia AT, Wiemels JL, Ford AM · Blood 2003What this evidence cannot show
A review, used here as a map of the case-series literature rather than as primary data. Twin case reports are strongly subject to publication bias toward concordant pairs, which inflates any naive concordance estimate.
Evidence that argues against this (1)
- The concordance rate the review reasons from is itself an ascertained number, and a population-based count is far lower. Merging the California Birth Master Statistical File (1982-2017) with the California Cancer Registry (1988-2022) gave 255 twin pairs in which at least one sibling had leukaemia and only 2 concordant pairs: 0.8% for all leukaemias, 0.5% for lymphoid (1 of 199), with estimated monozygotic rates of 2.6% and 1.4%. The authors attribute the ~10% figure in the twin-cohort literature to ascertainment bias in purpose-built cohorts. Crucially the same dataset does NOT deflate the infant claim - AML under one year of age was 25% concordant raw and 100% on the monozygotic estimate - so the aggregate must not be transferred to infants, and the review's inference from a 'modest overall rate' to required postnatal events rests on the aggregate. Nickels EM et al. 2025
- E3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
An approximately 5% concordance rate in identical twins, together with protracted latency, is the basis for requiring postnatal secondary events in most childhood leukaemia.
Chromosome translocations and covert leukemic clones are generated during normal fetal development · Mori H, Colman SM, Xiao Z, Ford AM, Healy LE, Donaldson C, Hows JM, Navarrete C, Greaves M · Proceedings of the National Academy of Sciences of the USA 2002What this evidence cannot show
The 5% figure is stated by the authors as a background premise rather than derived in this paper, and it aggregates fusion subtypes with very different biology. Confidence lowered from 'strong' to 'weak' (coordinator, 2026-09-18, confidence-split audit). This project's own entry `epi-fusion-positive-newborns-versus-leukemia-risk` records that the Mori 2002 cord-blood frequency is **unreproduced**: Lausten-Thomsen et al. screened 1,417 cord bloods and confirmed none. Two entries carried the paper at 'strong' while two carried it at 'weak', and the graph's own position is the latter. Confidence attaches to how well a claim is supported, not to how well known a paper is.
Evidence that argues against this (1)
- The 5% figure traces to a non-population-based questionnaire study, which is worth stating because the entry uses it as a premise. Twins with cancer were ascertained through three cooperative clinical trials groups, a cancer-twin registry and one large paediatric hospital, then sent a postal questionnaire; 3 of 197 twin pairs (76 monozygotic) were concordant for leukaemia, giving a monozygotic case-wise rate of 5% that the authors themselves describe as substantially lower than previously reported. So 5% is already a downward revision of an earlier, more biased number, it aggregates leukaemia subtypes with very different biology, and its denominator depends on which families answered a postal survey. Buckley JD et al. 1996
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Concordance of KMT2A-rearranged leukaemia in infant monozygotic twins is not universal: a fusion-positive preleukaemic clone was documented in an unaffected co-twin and then spontaneously cleared.
Discordance of MLL-rearranged (MLL-R) infant acute lymphoblastic leukemia in monozygotic twins with spontaneous clearance of preleukemic clone in unaffected twin · Chuk MK, McIntyre E, Small D, Brown P · Blood 2009What this evidence cannot show
A single case. The proposed immune-mediated clearance is a temporal association with viral illness and cytopenias, not a demonstrated mechanism; no immunological assays are reported. Three years of follow-up does not exclude later disease.
Evidence that argues against this (1)
- The prevailing case-series view before this report was that infant KMT2A-r twin concordance approaches 100% with very short latency, which the same authors state explicitly as the belief their case overturns. Greaves MF et al. 2003
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Concordant infant twin leukaemia can show a shared stem line with extensive divergent chromosomal subclones, dating the onset of intraclonal diversification to the prenatal period.
Prenatal chromosomal diversification of leukemia in monozygotic twins · Kempski H, Mensa-Bonsu KA, Kearney L, Jalali GR, Hann I, Khurshid M, Greaves M · Genes, Chromosomes & Cancer 2003What this evidence cannot show
One twin pair with acute monoblastic leukaemia, described by the authors as unique in the extent of its chromosomal diversity, so it is explicitly not representative.
Evidence that argues against this (1)
- Directly disputes dating the diversification to the prenatal period. Whole-genome sequencing of two twin ALL pairs placed ALL non-synonymous variants other than the initiating lesion as distinct between the twins and therefore postnatal, with non-coding changes almost entirely discordant. Extensive divergent subclones in a concordant pair are equally compatible with rapid POSTNATAL diversification from a shared, near-clonal prenatal founder; cytogenetic divergence at diagnosis carries no timestamp by itself. Ma Y et al. 2013
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Genetic divergence between concordant twins is also seen with rarer KMT2A partners, here t(1;11)(p32;q23) with a trisomy X subpopulation in one twin only.
Pre-natal, clonal origin of t(1;11)(p32;q23) acute lymphoblastic leukemia in monozygotic twins · Kotecha RS, Murch A, Kees U, Cole CH · Leukemia Research 2011What this evidence cannot show
Single case report; clonal identity inferred from shared cytogenetics rather than a sequenced genomic junction.
Evidence that argues against this (1)
- Base-rate check on what a rare-partner concordant pair can support. In a population-based registry of 255 twin pairs with leukaemia, only 2 pairs were concordant, and lymphoid-leukaemia concordance at ages 1-10 was 0.7% (2.1% on the monozygotic estimate). Concordant pairs with unusual partner genes are therefore heavily selected for publication, and a single such pair constrains the mechanism in that pair rather than showing that divergence is a general property of rarer KMT2A partners. Nickels EM et al. 2025
- ← depends onThe prevention window: the break happens in a fetus, and we know when it CLOSES but not when it opens
twin data bound the event before the shared monochorionic circulation
shared non-constitutive breakpoint in twins independently proves a single prenatal founding clone
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
near-total concordance argues the fusion is close to sufficient in the fetal target cell
concordance and latency in infant KMT2A-r pairs differ sharply from childhood ALL overall
high twin concordance is the main reason to expect high, not low, penetrance for KMT2A carriers
short latency and high concordance are the two halves of the near-sufficiency argument
- ← supportsThree pairs of twins who both got leukaemia: the patients who proved the fusion happens before birth
the three concordant twin pairs whose shared, non-constitutional KMT2A breakpoints are the founding observation behind the concordance epidemiology
Counter-evidence (Nickels 2025, PMID 39815817): population-based California registry linkage gives twin leukaemia concordance of 0.8% raw (MZ estimate 2.6%), far below the case-series figures -- but AML under one year was 25% raw and 100% on the MZ estimate, so the infant stratum survives while the general twin premise does not
Counter-evidence (Ma 2013, PMID 23569245): whole-genome sequencing of twin ALL pairs puts every non-initiating non-synonymous variant as distinct and POSTNATAL, so shared prenatal genetics is limited to the initiating lesion and cytogenetic divergence carries no prenatal timestamp
Counter-evidence (Moura 2015, PMID 24991719): a series of 77 neonatal leukaemias, KMT2A-AFF1 and KMT2A-MLLT3 prominent among the non-Down-syndrome cases, puts the lower bound on latency at effectively zero -- a seven-week twin presentation is a point on a distribution selected for being early