The predictable clinical error this beat exists to prevent: a menin inhibitor given for a KMT2A-rearranged sarcoma because of the gene name
also called menin inhibitor misapplied in KMT2A-rearranged solid tumour · gene-name-driven off-label menin inhibition
A menin-blocking drug is approved for KMT2A-rearranged leukaemia, and a molecular report on a sarcoma can say 'KMT2A rearrangement' in the same words. In these sarcomas the part of KMT2A that menin holds on to is not in the fusion at all, so the drug has nothing to grip.
The same thing, for a scientist
Delivery and target-identification risk. Menin inhibitors are approved and widely available for relapsed/refractory KMT2A-rearranged acute leukaemia, and molecular tumour boards routinely reason from a fusion gene name to a targeted agent. KMT2A-rearranged sarcoma generates a report reading 'KMT2A rearrangement' or 'YAP1::KMT2A fusion' - and in every arrangement with a published breakpoint the KMT2A contribution begins at exon 2, 4 or 5, so the menin-binding motif at aa 6-25 (exon 1) is not present in the fusion protein. The risk is not that the drug will be weakly active; it is that there is no target. Three supporting observations make the risk concrete rather than theoretical: the only closed menin-inhibitor study in solid tumours was terminated by the sponsor for pre-defined efficacy signals not being met; the only sarcoma in which menin inhibition has been tested preclinically (Ewing, where menin is a wild-type dependency) showed a three-order-of-magnitude potency gap versus a KMT2A-fusion leukaemia line with no tumour regressions; and menin behaves as a tumour suppressor in mesenchymal tissue in MEN1 syndrome, so the pharmacological direction may be wrong as well as ineffective. Mitigation is one line in a report: state the KMT2A exon range, not just the gene name.
How the error happens, step by step
- An RNA panel reports
YAP1::KMT2A fusionorKMT2A rearrangement detectedin a soft tissue sarcoma. - A board recognises KMT2A rearrangement as the indication for an approved menin inhibitor.
- Nobody reads the exon range, because in leukaemia the exon range never mattered - every leukaemic breakpoint retains exon 1 by construction.
- The drug is given to a patient whose fusion protein contains a 49-residue zinc finger of KMT2A and nothing else 1.
Steps 3 is the whole risk. The exon range is in the report already. Nobody has had a reason to look at it before, because no KMT2A rearrangement anyone cared about lacked exon 1.
Why 'try it, there is nothing else' is not a good argument here
There genuinely is no standard therapy for this entity, which is exactly the condition under which a low-probability targeted attempt looks reasonable. Three things argue against it:
- The only solid-tumour menin study that has read out was terminated for lack of efficacy at 41 patients 2. The class has no demonstrated solid-tumour activity to borrow.
- The one sarcoma with a reported menin dependency responded poorly - VTP-50469 slowed four of seven Ewing xenografts with no regressions, at an in vitro IC50 of 10 nM in MV4;11 against over 3 micromolar in Ewing lines 3.
- Menin may be acting as a suppressor in this lineage. Sarcomas arising in MEN1 syndrome patients show loss of heterozygosity at the MEN1 locus and loss of menin expression 4. Inhibiting menin in a mesenchymal tumour is therefore not merely a long shot; the direction may be wrong.
And the opportunity cost is real: an approved drug given on a gene name displaces enrolment in a trial that might address the actual driver, which by the functional data available is YAP1/TEAD.
Mitigation
- Report and read the KMT2A exon range, not the gene name. Exon 1 present or absent is the whole question, and it is one field.
- Ask which species was detected. A reciprocal
KMT2A::YAP1transcript with a 5' breakpoint at KMT2A exon 3, 4 or 6 does begin at exon 1, and those are reported in a substantial minority of YAP1-arm cases - so the answer is not uniform across transcripts within one tumour. - Do the co-immunoprecipitation before the patient.
exp-menin-pulldown-on-yky-versus-kmt2a-aff1is a low-cost experiment that would convert this risk from an argument into a measurement. - Do not treat the absence of a trial as an untested opportunity. The exon map is a positive reason to expect no binding site, which is a stronger prior than ignorance.
| kind | delivery |
|---|---|
| severity | high for the individual patient (an ineffective therapy with real toxicity, displacing enrolment in a trial of the actual driver); low population impact because the entity is rare |
| mechanism of the error | a molecular report naming 'KMT2A rearrangement' without the exon range, read by a board whose experience of KMT2A rearrangement comes entirely from leukaemia, where exon 1 is retained by construction |
| mitigation | report and read the KMT2A exon range; ask which transcript species was detected, because reciprocal 5'-KMT2A transcripts in the YAP1 arm DO begin at exon 1; run exp-menin-pulldown-on-yky-versus-kmt2a-aff1 before treating a patient on this rationale |
| status of the risk | predicted from coordinates, never measured; no menin binding assay has been run on any sarcoma KMT2A fusion |
affects
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib", "drug-enzomenib" ]
supporting observations
[ "NCT05731947 (revumenib, colorectal and other solid tumours, n=41) TERMINATED: pre-defined efficacy signals not met", "VTP-50469 in Ewing sarcoma: no regressions in 7 xenografts; IC50 10 nM in MV4;11 versus >3 micromolar in Ewing lines", "menin shows loss of heterozygosity and loss of expression in MEN1-syndrome sarcomas, i.e. suppressor behaviour in mesenchyme" ]
- This entry could itself do harm in the other direction. It argues from coordinates against a drug in a disease with no standard therapy, and if a second menin surface on KMT2A exists the argument collapses. The correct action is the co-immunoprecipitation, not a policy.
- Warning against gene-name-driven prescribing in this entity should not generalise into scepticism about menin inhibitors in KMT2A-rearranged leukaemia, where the interface is present, the drug is approved and the evidence is strong.
- Naming YAP1/TEAD as the rational alternative axis is itself an untested inference: no TEAD inhibitor has been given in this entity, and YAP1's transactivation region is partly replaced in the fusion allele.
- Has this already happened? A menin inhibitor given off-label for a KMT2A-rearranged solid tumour would most likely appear as a single case report or not at all, and no such report was found in the literature searched.
- Do the commercial panels that detect these fusions report the KMT2A exon range prominently, or only in a supplementary field? The mitigation depends entirely on that formatting.
- Is there any KMT2A-rearranged solid tumour whose dominant transcript DOES retain exon 1 - and would that patient be the one legitimate menin-inhibitor candidate outside haematology?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Derived here (2026-09-21) from coordinates: the KMT2A menin-binding motif lies at aa 6-25 within exon 1, and every published KMT2A-rearranged sarcoma arrangement with a stated breakpoint begins at KMT2A exon 2 (VIM), exon 4 (LDB1) or exons 4-6 (YAP1), so none of their dominant fusion transcripts can encode the motif the approved menin inhibitors compete with.
Ensembl REST (ENST00000534358) and UniProtKB (Q03164) records, KMT2A exon boundaries intersected with the feature table · Ensembl (EMBL-EBI); UniProt Consortium · Ensembl REST / UniProtKB 2026What this evidence cannot show
A coordinate prediction, not a binding measurement; no menin pulldown has ever been performed on any sarcoma KMT2A fusion protein, and no such protein has been observed by any method. The risk statement is therefore 'predicted to have no target', not 'measured to have no target'.
Evidence that argues against this (1)
- Reciprocal KMT2A::YAP1 transcripts with 5' breakpoints at KMT2A exons 3, 4 and 6 are reported in a substantial minority of YAP1-arm cases, and those DO begin at exon 1. If such a species is translated in a given patient's tumour, that patient's tumour contains a menin-competent KMT2A after all - which is a reason to check the transcript rather than a reason to give the drug. Warmke LM et al. 2025
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
The only menin-inhibitor study in non-haematologic malignancy to have closed, NCT05731947 (revumenib in colorectal cancer and other solid tumours, 41 patients enrolled), was terminated by the sponsor after Phase 1b because pre-defined signals of efficacy were not met in the Phase 1b dose expansion cohort.
A Phase 1/2 Study to Investigate the Safety, Tolerability, Pharmacokinetics, Pharmacodynamics, and Efficacy of SNDX-5613 in Patients With Colorectal Cancer and Other Solid Tumors · Syndax Pharmaceuticals · ClinicalTrials.gov 2023What this evidence cannot show
A registry record read via the raw ClinicalTrials.gov API, with no publication, no response data and no biomarker analysis. The trial did not select for KMT2A rearrangement, so it bounds the class's demonstrated solid-tumour activity rather than its activity in a KMT2A-rearranged solid tumour specifically. The whyStopped text is the sponsor's own wording.
Evidence that argues against this (1)
- A negative unselected solid-tumour trial is weak evidence about a biomarker-selected population, and one menin-inhibitor solid-tumour study remains open on a GIST-specific rationale. The class is not formally excluded from solid tumours; it is simply undemonstrated there. Kura Oncology et al. 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In seven Ewing sarcoma xenograft models and Ewing sarcoma cell lines, the menin-MLL1 inhibitor VTP-50469 slowed tumour progression significantly in four models with no evidence of regression, at an in vitro IC50 of 10 nM in the MLL-rearranged leukaemia line MV4;11 against greater than 3 micromolar in the Ewing lines.
Evaluation of VTP-50469, a menin-MLL1 inhibitor, against Ewing sarcoma xenograft models by the pediatric preclinical testing consortium. · Kurmasheva RT, Bandyopadhyay A, Favours E, et al. · Pediatr Blood Cancer 2020What this evidence cannot show
Ewing sarcoma, in which the menin dependency is a wild-type MLL1/menin dependency downstream of EWSR1::FLI1 rather than a fusion-interface dependency, so it is not a model of KMT2A-rearranged sarcoma. Immunodeficient-mouse xenografts. The load-bearing number is a viability IC50 ratio, not a target-engagement measurement.
Evidence that argues against this (1)
- Other groups report that Ewing sarcoma tumorigenicity is critically dependent on MLL1 and menin, and that EWS-FLI1 and menin converge on ATF4. So a genetic menin dependency in a sarcoma is on record while the pharmacological result is close to negative - and the pharmacological arm may have failed on exposure or on the assay rather than on the biology. Svoboda LK et al. 2017
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In five patients with MEN1 syndrome and a sarcoma, all five sarcomas showed loss of heterozygosity of the MEN1 gene and loss of menin expression, leading the authors to suggest sarcomas may be a phenotypic expression of MEN1 syndrome - i.e. menin behaving as a tumour suppressor in mesenchymal tissue.
Sarcomas arising in MEN1 patients: demonstrating LOH of the MEN1 locus and loss of menin expression. · van Leeuwaarde RS, Halfdanarson TR, Sudhakar SM, et al. · Fam Cancer 2024What this evidence cannot show
Five patients, a case series, and germline MEN1 syndrome rather than sporadic sarcoma - so it establishes a suppressor role for menin in a hereditary context, not in the KMT2A-rearranged sarcomas at issue here. Loss of heterozygosity plus loss of expression is strong evidence of a two-hit suppressor mechanism but does not by itself show that pharmacological menin inhibition would promote mesenchymal tumours. Read from the abstract.
Evidence that argues against this (1)
- Menin is a context-dependent protein that acts as a suppressor in endocrine tissue and as a required co-factor in KMT2A-fusion leukaemia, and menin inhibitors have been given to many patients without a reported sarcoma signal. A germline two-hit suppressor phenotype over a lifetime is not the same risk as weeks of partial pharmacological occupancy in an adult - the same scope distinction this project applies to gnomAD constraint. Jiménez JA et al. 2021
the entity in which the error would occur
- confers resistance to →Revumenib (SNDX-5613, Revuforj) — first approved menin–KMT2A interface inhibitor
no menin-binding motif in the fusion, so no target - predicted from coordinates, never measured
- confers resistance to →Ziftomenib (KO-539, Komzifti) — once-daily menin inhibitor approved in NPM1-mutant AML
same coordinate argument applies to every menin inhibitor in the class
the registry bound and the terminated solid-tumour trial that make the risk concrete
the measurement that would replace this coordinate argument
menin shows suppressor behaviour in MEN1-syndrome sarcomas, so the pharmacological direction may be wrong in mesenchyme
an MBM-retaining KMT2A fusion in a sarcoma is exactly the gene-name-driven prescribing risk this entry names
MBM-retaining KMT2A fusion in a leiomyosarcoma - the gene-name-driven prescribing risk
complete leukaemic N-terminal module in an angiosarcoma - the gene-name-driven prescribing risk
MBM retained without CXXC in an MPNST - and the partner's own pharmacology points in two opposite directions
complete leukaemic N-terminal module in an MPNST - and the partner is predicted not to be translated at all
MBM and CXXC retained in a uterine sarcoma - the gene-name-driven prescribing risk
MBM and CXXC retained in an osteosarcoma - and the partner is a candidate tumour suppressor, so the therapeutic logic inverts
SHARPENS THE RISK RATHER THAN EASING IT: the risk entry warns against prescribing off a gene name. These seven cases are exactly the rows most likely to be misread as a rationale, because they retain the drug-relevant motif and three look leukaemia-like on paper. Nothing here is a treatment rationale -- menin is a tumour suppressor in mesenchyme, and the only non-haematologic menin trial with a result was terminated after phase 1b for not meeting pre-defined efficacy signals.