A KMT2A fusion acquired at seventy-three: KMT2A::ELL AML arising out of polycythaemia vera, and a menin inhibitor used to hold the line

Not with an acute illness. This patient had been followed for a chronic myeloproliferative neoplasm for four years. What changed, in early 2024 and while on hydroxyurea, was progressive anaemia, thrombocytopenia, worsening fatigue and increasing splenomegaly 1.

The fatigue and anaemia are the same concept-pallor-and-fatigue-anaemia that brings a pale infant to a paediatrician, and they mean the same thing: the marrow is failing. The difference is that here it took years, in a patient already under surveillance, and the trigger for action was a rising transfusion requirement rather than a sudden collapse.

Hydroxyurea was stopped and a marrow biopsy showed grade 2-3 reticulin fibrosis with megakaryocytic atypia - post-polycythaemic myelofibrosis. He was started on pacritinib. About five weeks later a repeat marrow showed more than 20% myeloblasts. That interval, between myelofibrosis in July 2024 and overt leukaemia in August 2024, is the whole transformation.

Counts at transformation were profoundly low, not high: WBC 0.9 x 10^9/L, haemoglobin 6.7 g/dL, platelets 14 x 10^9/L, with weekly red cell transfusions 1. There is no hyperleukocytosis here. The contrast with the infant entries in this arm, where white counts ran from 104 to 1,755 x 10^9/L, is the single most visible difference between de novo and post-myeloproliferative KMT2A-rearranged disease.

Cytogenetics showed 46,XY,t(11;19)(q23.3;p13.1) and FISH confirmed a KMT2A::ELL fusion in 23.6% of analysed cells, alongside derivative chromosomes 9, 14 and 16 (concept-cytogenetics-karyotype-kmt2a, concept-fish-split-signal-kmt2a). RNA fusion analysis confirmed the rearrangement (concept-molecular-confirmation-fusion-transcript).

The band position matters. 19p13.1 carries ELL; 19p13.3 carries MLLT1. Both give a t(11;19) on a karyotype report, and the difference between them is the difference between two distinct fusions - one of which appears elsewhere in this arm, in a newborn who died on day four. A reader comparing the two entries is looking at the same-looking translocation producing two different diseases in patients seventy-three years apart in age.

Next-generation sequencing showed JAK2 V617F persisting at 42.8% variant allele frequency alongside the new fusion, plus a GATA2 M388T variant the authors classify as of uncertain significance because some GATA2 alterations are germline predisposition variants. The authors read the persistence of JAK2 alongside acquisition of KMT2A::ELL as favouring branching clonal evolution with late clonal dominance, and state plainly that single-cell sequencing would be needed to establish the architecture.

He was not a candidate for intensive induction or transplant, on grounds of age, coronary artery disease, paroxysmal atrial fibrillation, chronic kidney disease and diabetes. First-line treatment was azacitidine 75 mg/m2 days 1-7 with venetoclax on standard ramp-up to 400 mg daily, with allopurinol prophylaxis. He completed about 13 cycles - roughly 13 months - before documented progression 1.

On progression, with persistent cytopenias and continued transfusion dependence, treatment was switched to revumenib 160 mg orally twice daily, on the basis of the KMT2A::ELL fusion. Antimicrobial prophylaxis was fluconazole, aciclovir and cefpodoxime; serial ECGs were done because of the known QT risk, and the QT intervals remained stable. He had no differentiation syndrome, no tumour lysis and no significant non-haematological toxicity.

Outcome as reported: at four months on revumenib he remained clinically stable, with persistent transfusion dependence but no progressive blast burden and no new organ involvement. The authors are explicit that this is disease stabilisation, not complete haematologic remission, and that residual marrow fibrosis limited count recovery even where blast burden was controlled. They also note the absence of measurable-residual-disease monitoring, which limits how precisely the molecular response can be described.

Because a KMT2A fusion is not only something a child is born with.

1. The same lesion, acquired seventy-three years later, in a different context. KMT2A rearrangements appear in roughly 3% of de novo AML and around 9% of therapy-related AML, and are distinctly uncommon in post-myeloproliferative AML, where TP53- and ASXL1-driven transformation dominates. This patient acquired KMT2A::ELL on top of a JAK2-mutant clone that had been there for years.
2. It contributes ELL to the arm. ELL is a canonical partner in this graph and a transcription-elongation factor, and no infant case with this partner was found; the partner appears here in an adult secondary leukaemia.
3. What a menin inhibitor looks like when it does not produce a remission. Elsewhere in this arm revumenib cleared marrow and CSF. Here it stabilised disease in a frail, fibrotic, transfusion-dependent patient - no complete remission, no count recovery, but no progression either, for four months. That is a real and commonly encountered outcome and the graph should carry it next to the success.
4. The counts are the opposite of the infant presentations. Pancytopenia, not hyperleukocytosis; a five-week transformation inside a four-year illness, not a week of cough.