Oncofusion

start / graph / drug-enasidenib

Drugmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q81-methylome-axis

Enasidenib - an approved mutant-IDH2 inhibitor sitting next to the one methylome co-mutation that IS enriched in KMT2A-rearranged AML, and never tested there

also called enasidenib · AG-221 · CC-90007 · Idhifa

Enasidenib blocks the mutant form of IDH2, the enzyme whose faulty product drives DNA over-methylation, and it has been approved for adults with IDH2-mutant leukaemia since 2017. IDH2 mutations turn up more often in KMT2A-rearranged AML than in other AML - and nobody has ever reported giving this drug to that group specifically.

The same thing, for a scientist

**Enasidenib (CHEMBL3989908, AG-221/CC-90007, first approval 2017) is a first-in-class oral selective inhibitor of mutant IDH2 enzymes.** In the first-in-human phase 1/2, among patients with relapsed or refractory IDH2-mutant AML the **overall response rate was 40.3% with a median response duration of 5.8 months**, median overall survival 9.3 months, and 19.7 months among the 34 patients (19.3%) attaining complete remission; grade 3-4 drug-related events were indirect hyperbilirubinaemia (12%) and **IDH-inhibitor-associated differentiation syndrome (7%)**; no MTD was reached from 50 to 650 mg/day and 100 mg once daily was selected [E1]. Crucially, **responses were associated with cellular differentiation and maturation, typically without aplasia - the drug works by inducing differentiation, not by cytotoxicity** [E1]. That mechanism is worth stating plainly in this graph, because **it is the same mechanism revumenib works by**, and the two share differentiation syndrome as their signature toxicity. **Enasidenib is relevant to this beat for exactly one reason and it is a hypothesis, not a result:** IDH2 alteration is enriched in KMT2A-altered AML at 16.0% versus 10.4% (recomputed Fisher odds ratio 1.63, p = 4.1e-4), so a subgroup exists in which the approved drug and the menin inhibitor address different lesions in the same patient [E2]. **No trial, no case series and no case report of any IDH inhibitor in KMT2A-rearranged leukaemia was found in either the literature or the registry** on 2026-09-21 - see `absence-no-idh-inhibitor-has-been-reported-in-a-kmt2a-rearranged-cohort` for the bounded query set [E3]. **And the paediatric position is worse than untested: this graph found no paediatric enasidenib licence or trial in KMT2A-rearranged disease, and the one registry record pairing an IDH inhibitor with KMT2A is an adult master protocol with a minimum age of 18** [E3].

Go deeper
Why this entry exists as more than a stub

The dispatch question was: IDH inhibitors exist and are approved; if IDH2 is enriched in KMT2A-rearranged AML, is there any trial or case series in that subgroup? The answer is no, and a bounded, dated no is a result. But an approved drug with a matching biomarker and no data in a subgroup is also a decision, and the decision has a rationale that should be written down rather than assumed:

  1. The IDH2 enrichment is one paper, one assay, 2026, retrospective, with no outcome data and with 43% of the KMT2A arm being duplications rather than fusions 2.
  2. IDH1 in the same table is flat (6.7% vs 7.6%, p = 0.53), which is what a chance or lesion-type explanation predicts and what a 2-HG mechanism does not.
  3. Enasidenib is licensed by IDH2 genotype, not by KMT2A genotype, so any IDH2-mutant KMT2A-rearranged patient is already eligible for it on label. The absence is therefore an absence of reporting, not an absence of access - and a subgroup analysis of existing enasidenib trials by KMT2A status is a desk exercise somebody could do this year.
  4. Both enasidenib and revumenib cause differentiation syndrome - 7% grade 3-4 for enasidenib 1, 10% any-grade in the SAVE triplet. Combining them stacks the class toxicity, and nobody has.
The differentiation-syndrome convergence, which is the real finding here

Enasidenib's responses come with "cellular differentiation and maturation, typically without evidence of aplasia" 1. Menin inhibitors do the same thing by a completely different route - releasing the HOXA/MEIS1 differentiation block rather than relieving 2-HG inhibition of dioxygenases - and produce the same clinical syndrome.

Two consequences the graph should hold:

  • A pharmacodynamic one. If both agents act by forcing differentiation of the same blast population, their effects may be non-additive - the cell can only differentiate once. That is a specific, testable prediction against the intuition that two mechanisms are better than one.
  • A safety one. Differentiation syndrome is the dose-limiting and occasionally fatal toxicity of both classes. Stacking them in an IDH2-mutant KMT2A-rearranged patient is the combination with the clearest biological rationale and the clearest toxicity rationale against it, and no protocol has attempted to navigate that.
Facts and figures
chembl IDCHEMBL3989908
targetgene-idh2
modalitysmall molecule (selective inhibitor of MUTANT IDH2 enzymes; not a degrader and not active against wild-type dependency)
statusApproved (ChEMBL max_phase 4, first approval 2017) for IDH2-mutant relapsed/refractory AML in adults. No paediatric indication identified in this beat and no regulatory registry queried.
mechanism summaryInhibits the neomorphic activity of mutant IDH2, lowering 2-hydroxyglutarate and relieving competitive inhibition of alpha-KG-dependent dioxygenases including the TET 5mC hydroxylases and the KDM histone demethylases. Clinically it acts by inducing differentiation and maturation rather than by cytotoxicity, typically without aplasia.
shown for fusionNO - not in any KMT2A fusion, any KMT2A-rearranged cohort, any case series or any case report, in either the literature or the trial registry as of 2026-09-21. The sole rationale is a co-occurrence figure.
differentiation syndrome convergenceEnasidenib and menin inhibitors both act by releasing a differentiation block and both cause differentiation syndrome (enasidenib 7% grade 3-4; SAVE triplet 10% any grade, 5% grade 3). A combination stacks the class toxicity and may be non-additive if both release the same block.
trials
[
  "NCT01915498",
  "NCT03013998"
]
key ae
[
  "IDH-inhibitor-associated differentiation syndrome, 7% grade 3-4 - the same class toxicity as menin inhibitors",
  "indirect hyperbilirubinaemia, 12% grade 3-4"
]
resistance
[
  "Not characterised in this beat. The phase 1/2 abstract reports no resistance mechanism."
]
name list for absence bounding
[
  "enasidenib",
  "AG-221",
  "CC-90007",
  "Idhifa"
]
How this could bite us back
  • Differentiation syndrome is the dose-limiting and occasionally fatal toxicity of both IDH inhibitors and menin inhibitors. The combination with the clearest biological rationale in this beat is also the one with the clearest reason to be dangerous, and no protocol has attempted it.
  • Relieving 2-HG inhibition restores TET activity and lowers CpG methylation - which, if the address argument is right, RESTORES unmethylated CpG and could expand rather than contract the fusion's available address space. An IDH2 inhibitor in a KMT2A-rearranged patient is not obviously anti-leukaemic on this graph's own model, and might be the wrong direction.
  • The entire rationale rests on one retrospective commercial-panel series with no outcome data, an internally inconsistent p-value, and a KMT2A arm that is 43% duplications. That is not enough to move a patient, and this entry should not be cited as though it were.
Still open
  1. Do the existing enasidenib trial cohorts contain KMT2A-rearranged patients, and how did they do? This is a retrospective subgroup analysis of data that already exists and would need no new patient.
  2. Is IDH2 mutation in a KMT2A-rearranged AML clonally dominant or subclonal? An IDH inhibitor in a subclone is a different proposition from one in the founding clone, and no clonal-architecture data exist for this co-occurrence.
  3. If both enasidenib and a menin inhibitor work by forcing differentiation, are their effects additive or does the blast differentiate once? This is testable in an IDH2-R140Q KMT2A-rearranged model that nobody has built.
  4. Would lowering 2-HG restore TET activity enough to change CpG-island methylation at fusion-bound sites, and would that raise or lower fusion occupancy? The prediction is unclear because 2-HG also inhibits the fusion's competing CXXC readers KDM2A and KDM2B.
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In the first-in-human phase 1/2 of enasidenib (AG-221/CC-90007), a first-in-class oral selective inhibitor of mutant IDH2, patients with relapsed or refractory IDH2-mutant AML had an overall response rate of 40.3% with median response duration 5.8 months and median overall survival 9.3 months (19.7 months among the 34 patients attaining complete remission); grade 3-4 enasidenib-related events included indirect hyperbilirubinaemia (12%) and IDH-inhibitor-associated differentiation syndrome (7%); no maximum tolerated dose was reached from 50 to 650 mg/day; and responses were associated with differentiation and maturation rather than cytotoxicity.

    Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. · Stein EM, DiNardo CD, Pollyea DA, et al. · Blood 2017
    PMID 28588020DOI
    What this evidence cannot show

    Adults with advanced IDH2-mutant myeloid malignancy; single-arm phase 1/2 dose-escalation and expansion (NCT01915498), efficacy reported for the relapsed/refractory AML subgroup only. No KMT2A-rearranged subgroup is reported and KMT2A status is not stated as collected. Read from abstract only in this beat. Nothing paediatric.

    Evidence that argues against this (1)
    • Enasidenib's benefit is defined entirely by IDH2 genotype, and the only reason to consider it in KMT2A-rearranged disease is a single retrospective co-occurrence figure from a commercial panel with no outcome data - which is a weaker rationale than the drug's own label already provides. Batayneh O et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    IDH2 alteration is significantly more frequent in KMT2A-altered than KMT2A-wild-type adult AML in a 3,863-case hybrid-capture series (16.0% of 521 vs 10.4% of 3,342), giving a recomputed Fisher odds ratio of 1.630 and p = 4.1e-4 - which matches the paper's Table 2 value of 0.0004 and not its abstract's p < 0.0001.

    Characterizing KMT2A Rearrangement in Acute Myeloid Leukemia: A Comprehensive Genomic Study. · Batayneh O, Moein M, Naji NS, et al. · Cancers (Basel) 2026
    PMID 41514668DOI
    What this evidence cannot show

    Read in full text from PMC12784994 on 2026-09-21 including Table 2; the paper disagrees with itself on this p-value and the table is the value that reproduces. Adult AML, median age 62; 43.1% of the KMT2A-altered arm are duplications rather than fusions and the table is not stratified by lesion type. No outcome data. Nothing about paediatric or infant disease.

    Evidence that argues against this (1)
    • IDH1, which produces the same oncometabolite by the same neomorphic mechanism, shows no enrichment in the same table (6.7% vs 7.6%, recomputed p = 0.53), so the mechanistic story usually attached to this finding does not survive its own internal control. Xu W et al. 2011
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry

    As of 2026-09-21, no trial, case series or case report of any IDH1 or IDH2 inhibitor in KMT2A-rearranged leukaemia has been published or registered: a PubMed query over six inhibitor names crossed with four KMT2A-rearrangement terms returns 4 records, all of them broad reviews, and a ClinicalTrials.gov query for the same inhibitors with KMT2A or MLL returns 1 record - the Beat AML master protocol NCT03013998, an adult study with a minimum age of 18 in which KMT2A and IDH2 are separate biomarker-assignment criteria rather than a joint eligibility criterion.

    NCT03013998
    What this evidence cannot show

    Bounded by titles and abstracts on the PubMed side (rule 80) - an IDH-inhibitor case inside a KMT2A cohort paper would be invisible. Bounded by six inhibitor names on the literature side (rule 72): enasidenib, ivosidenib, olutasidenib, vorasidenib, AG-221, AG-120, FT-2102. NOTE AN INSTRUMENT FAILURE CAUGHT HERE: the identical query with t(4;11) added inside the parenthesised OR group returned 0 rather than 4, because PubMed's parser breaks on translocation parentheses - the zero was an artefact and is not the figure quoted. Positive control through the identical shape: the same inhibitor clause crossed with NPM1 returns 12, the inhibitor clause alone returns 644, and the KMT2A clause alone returns 3,160. The Beat AML record was retrieved individually from the raw API with eligibility and intervention lists; its enasidenib arm (AG-221 with azacitidine, substudy S3) is assigned on IDH2 mutation.

    Evidence that argues against this (1)
    • Enasidenib is licensed by IDH2 genotype irrespective of KMT2A status, so IDH2-mutant KMT2A-rearranged patients have been treated with it and simply not reported as a subgroup - the absence is of reporting, not of exposure, and a retrospective subgroup analysis of existing enasidenib cohorts would settle it without a new trial. Stein EM et al. 2017
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Enasidenib is ChEMBL3989908, a small molecule with maximum phase 4 and a first approval recorded in 2017, carrying the synonyms AG-221 and CC-90007.

    ChEMBL molecule record CHEMBL3989908 (ENASIDENIB) · ChEMBL, EMBL-EBI · ChEMBL database 2026
    What this evidence cannot show

    Fetched 2026-09-21. ChEMBL's first_approval is one global year with no indication, jurisdiction or age restriction attached; no regulatory registry was queried in this beat, so nothing here bounds paediatric labelling (rule 75). Enasidenib carries a boxed warning for differentiation syndrome in its US labelling, which is not represented in the ChEMBL record.

    Evidence that argues against this (1)
    • A ChEMBL max_phase of 4 says a drug is approved somewhere for something; it cannot tell you whether it is available to a child with KMT2A-rearranged leukaemia, which is the question this graph needs answered and which only a regulatory registry can answer. ChEMBL et al. 2026
Connected to