IDH and TET2 mutations are mutually exclusive in AML - so 'IDH2 enriched, TET2 depleted' in one cohort is ONE observation, not two
also called IDH-TET2 mutual exclusivity · epistasis of the hypermethylator lesions
Two different mutations, in IDH and in TET2, produce the same over-methylated DNA, and leukaemias almost never carry both - one is enough. So if a group of patients has more of one, it will automatically look like it has less of the other, whether or not anything interesting is going on.
The same thing, for a scientist
**A statistical caution with a mechanistic cause, and the graph needs it because two of this beat's headline numbers come from the same table.** In a large adult AML cohort, IDH1/2 mutations were **mutually exclusive** with mutations in TET2, and TET2 loss-of-function produced epigenetic defects similar to IDH mutation, with mutant IDH impairing TET2 catalytic function in cells - the two lesions are alternative routes to one hypermethylator phenotype, and a cell that has taken one gains nothing from the other [E1]. **The consequence for this beat is arithmetic.** In the FoundationOne Heme series, KMT2A-altered AML is **IDH2-enriched (16.0% vs 10.4%)** and **TET2-depleted (10.1% vs 16.4%)** in the same 3,863 patients [E2]. **Under mutual exclusivity those two rows are not independent**: patients whose hypermethylator slot is filled by IDH2 cannot also be counted as TET2-mutant, so an excess of one mechanically produces a deficit of the other. **Cycle 44 read the TET2 depletion as a finding and this beat's brief read the IDH2 enrichment as a second, converging finding. They are at most one and a half findings, and possibly one.** What would separate them is a joint analysis - the observed count of IDH2-mutant-and-TET2-wild-type against the count expected from the marginals under exclusivity - and **the paper publishes only marginals, so it cannot be done from the published data** [E2]. **This does not make either row false.** The total hypermethylator-lesion burden in KMT2A-altered AML is 26.1% (IDH2 plus TET2, treating them as exclusive) against 26.8% in KMT2A-wild-type - **essentially identical** - which is the arithmetic the exclusivity model predicts and which nobody has stated. **If that near-identity is the whole story, the methylome co-mutation axis in KMT2A-rearranged AML is a redistribution and not an enrichment**, and the live threat to the address argument is smaller than it looks.
The arithmetic, stated so it can be checked or refuted
Under strict mutual exclusivity, the fraction of patients carrying some hypermethylator lesion of this class is the sum of the individual frequencies.
| KMT2A-altered (n = 521) | KMT2A-wild-type (n = 3,342) | |
|---|---|---|
| IDH2 | 16.0% | 10.4% |
| TET2 | 10.1% | 16.4% |
| IDH1 | 6.7% | 7.6% |
| IDH2 + TET2 | 26.1% | 26.8% |
| IDH2 + TET2 + IDH1 | 32.8% | 34.4% |
The totals differ by 0.7 and 1.6 percentage points. With n = 521 in the smaller arm, a difference of that size is nowhere near detectable. So the hypermethylator-lesion burden of KMT2A-altered AML is indistinguishable from that of KMT2A-wild-type AML; what differs is which gene carries it.
Three caveats that keep this from being a finished result, and they are not small:
- The sums are only valid if exclusivity is strict. Figueroa reports exclusivity in a 2010 cohort without an exclusivity test statistic in the abstract, and co-mutated cases are reported in later series. If 2% of patients carry both, the sums are 2 points too high in whichever arm the double mutants sit.
- This is not a statistical test. No interval is computed on the sums here and no formal comparison is made, because the covariance between the two rows is unknown without patient-level data. It is an observation about the marginals, offered as a hypothesis, and it is labelled as one.
- DNMT3A, also flat (16.0% vs 17.2%), is a fourth route to an altered methylome and is not in the sums because it is not in the exclusivity relationship.
Why this belongs in the graph as its own entry
Because two entries written in two consecutive cycles - gene-tet2 (cycle 44) and gene-idh2 (this cycle) - each treat their row as independent evidence about KMT2A biology, and a reader taking both at face value would conclude that the methylome is doubly implicated. It may be singly implicated, or not at all.
And there is a general form worth stating, in the family of rule 84: when two variables are known to be mutually exclusive, a table showing one up and the other down in the same arm contains less information than two independent rows would. The null is not "no difference in either" - the null is "the same total, differently distributed", and against that null this comparison has not been tested.
What would settle it. Patient-level data from the FoundationOne series, or any KMT2A cohort with both genes genotyped, and one 2x2 of hypermethylator-lesion-present versus absent by KMT2A status. That is the measurement exp-idh2-mutant-expression-versus-kmt2a-fusion-occupancy cannot substitute for and that no amount of mechanism can replace.
| relationship | mutual exclusivity between IDH1/IDH2 and TET2 mutation in AML |
|---|---|
| mechanistic cause | both lesions converge on impaired TET2 catalytic output and a hypermethylator phenotype; the second lesion adds nothing |
| consequence for this graph | The IDH2 enrichment and the TET2 depletion in the FoundationOne Heme KMT2A comparison are not independent observations. Cited together as converging evidence, they double-count. |
| what would settle it | One 2x2 of hypermethylator-lesion-present versus absent by KMT2A status, from patient-level data in any cohort genotyping both genes. |
| system | Adult AML cohorts (Figueroa 2010; FoundationOne Heme 2026). Nothing in infant or paediatric disease and nothing functional. |
hypermethylator lesion burden
{
"kmt2a_altered_idh2_plus_tet2_pct": 26.1,
"kmt2a_wildtype_idh2_plus_tet2_pct": 26.8,
"difference_pct_points": -0.7,
"with_idh1_added_kmt2a_altered_pct": 32.8,
"with_idh1_added_kmt2a_wildtype_pct": 34.4,
"status": "HYPOTHESIS from marginals under an exclusivity assumption. Not a statistical test, no interval, covariance unknown.",
"source_pmid": "41514668"
}- This entry is a brake, and brakes get ignored. If a future beat cites the IDH2 enrichment and the TET2 depletion together as two lines of evidence that the methylome is co-opted in KMT2A-rearranged AML, it will be double-counting, and the graph will carry a stronger claim than the data support - which is how a stratification hypothesis reaches a protocol.
- Is the total hypermethylator-lesion burden genuinely equal between KMT2A-altered and KMT2A-wild-type AML? The marginals say 26.1% against 26.8% and only patient-level data can confirm it.
- How strict is IDH-TET2 exclusivity in modern panel cohorts? The 2010 claim predates routine NGS and no test statistic is available from its abstract.
- If the burden is equal and only the gene differs, is there ANY biological consequence of IDH2 rather than TET2 carrying it in a KMT2A-rearranged cell? IDH2 additionally inhibits histone demethylases, so the two are not epigenetically identical even if they are statistically exchangeable.
- Do the same arithmetic on DNMT3A and ASXL1 and the picture may change again - ASXL1 is significantly depleted in KMT2A-altered AML (10.5% vs 16.6%) and this entry does not account for it.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In a large adult AML cohort, IDH1/2 mutations were mutually exclusive with mutations in the alpha-ketoglutarate-dependent enzyme TET2, TET2 loss-of-function mutations were associated with epigenetic defects similar to those of IDH1/2 mutants, and expression of mutant IDH impaired TET2 catalytic function in cells - establishing the two lesions as alternative routes to one hypermethylator phenotype.
Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. · Figueroa ME, Abdel-Wahab O, Lu C, et al. · Cancer Cell 2010What this evidence cannot show
Adult AML, 2010, before routine next-generation sequencing panels; the cohort size and the exclusivity test statistic are not given in the abstract and the full text was not read in this beat. 'Mutually exclusive' in a cohort of moderate size at these mutation frequencies is a claim with limited power - a handful of co-mutated cases can be consistent with the data. No KMT2A-rearranged subgroup is analysed. Methylation by HELP rather than bisulfite.
Evidence that argues against this (1)
- Later series do report IDH and TET2 co-mutated AML cases, so exclusivity is a strong tendency rather than an absolute rule, and any arithmetic that adds the two frequencies as though they never overlap will overstate the total hypermethylator burden by the size of the overlap. Batayneh O et al. 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In the 3,863-case FoundationOne Heme AML series, KMT2A-altered cases show IDH2 at 16.0% against 10.4% and TET2 at 10.1% against 16.4%, so summing the two mutually exclusive hypermethylator lesions gives 26.1% in KMT2A-altered against 26.8% in KMT2A-wild-type - a difference of 0.7 percentage points, far below what n = 521 can resolve - and the paper reports marginals only, with no joint IDH2-by-TET2 cross-tabulation that would allow the exclusivity to be tested in these arms.
Characterizing KMT2A Rearrangement in Acute Myeloid Leukemia: A Comprehensive Genomic Study. · Batayneh O, Moein M, Naji NS, et al. · Cancers (Basel) 2026What this evidence cannot show
THE SUM IS NOT A STATISTICAL TEST. It is valid only under strict mutual exclusivity, no interval is computed on it, and the covariance between the two rows cannot be estimated from published marginals. It is offered as a hypothesis about what the table may mean, explicitly not as a result. Read in full text from PMC12784994 on 2026-09-21 including Table 2. Adult AML, median age 62; 43.1% of the KMT2A-altered arm are duplications rather than fusions.
Evidence that argues against this (1)
- If exclusivity is not strict in this cohort, or if the double mutants are concentrated in one arm, the near-identity of the two sums is an artefact of the assumption rather than a property of the data - and there is no way to check from the published table, which is exactly why this entry's confidence is moderate and not strong. Figueroa ME et al. 2010
The reason the TET2 depletion and the IDH2 enrichment in one cohort are not independent evidence: summed, the hypermethylator burden is 26.1% vs 26.8%.
Applies the same brake to the enrichment as to the depletion; neither row should be cited as independent of the other.
The mechanistic cause of the exclusivity: both lesions converge on impaired TET output, so the second adds nothing.
- related to →Is the CpG address a contested resource? On current evidence, no - and the beat that went looking says so
Shrinks the live threat: if the hypermethylator burden is equal and only the gene differs, the methylome axis is a redistribution rather than an enrichment.
The deflating possibility written into the falsifiers: the summed hypermethylator burden is 26.1% vs 26.8% and there may be nothing to explain.
One half of the exclusivity relationship that makes the IDH2 and TET2 rows non-independent.