Fever and headache thirty-two months after transplant: KMT2A::AFDN AML relapsing into the CNS, and a menin inhibitor that cleared the spinal fluid

With an infection that would not settle. The record is persistent high fever to 39.8 C, spontaneous bleeding from the gums, and a respiratory infection 1.

Those first two are concept-fever-and-infection-at-presentation and concept-bruising-petechiae-thrombocytopenia - marrow failure presenting as its two commonest consequences, a susceptibility to infection and a failure of haemostasis. The white count was 104.16 x 10^9/L, over the conventional hyperleukocytosis threshold (concept-hyperleukocytosis-emergency), and platelets were 117 x 10^9/L.

This entry is here partly because the patient is an adult. Everything else in this arm is an infant, and the graph as a whole is about infant disease. But KMT2A::AFDN AML in an adult is the same fusion biology, and the presenting picture is recognisably the same: fever, bleeding, a very high white count.

Marrow showed 84% blasts. Immunophenotyping found 66.9% of immature cells positive for CD33, CD64, CD117, CD14 and MPO - the monocytic pattern of AML-M5 that concept-flow-cytometry-immunophenotype-kmt2a-r associates with KMT2A rearrangement on the myeloid side 1.

The karyotype was not a clean t(6;11). It was complex - 46-52,XY with gains and losses across many chromosomes and, relevantly, del(11)(q23) present in two copies - which is exactly the situation in which a karyotype alone cannot name a fusion. KMT2A rearrangement was confirmed as positive and PCR detected KMT2A::AFDN (concept-cytogenetics-karyotype-kmt2a, concept-molecular-confirmation-fusion-transcript). The PCR assay had a stated limit of detection of 1 copy per 10^4 cells, and it was subsequently used as the MRD measure throughout.

Mutations in TET2, RUNX1 and ASXL1 were also present - a co-mutation profile more typical of adult myeloid disease than of infant leukaemia, and a reminder that the same fusion arrives in different genomic company depending on the age of the patient.

First line (revumenib was not available). Hydroxyurea 2-3 g/day for 7 days for the white count, then IA induction (idarubicin 12 mg/m2 days 1-3, cytarabine 200 mg/m2 days 1-7) to morphological complete remission - but with flow MRD at 0.41% and KMT2A::AFDN at 21.32 copies per 100 ABL1 copies. Consolidation with IA and high-dose cytarabine brought MRD to 0.31% and the transcript to undetectable. Intrathecal cytarabine and dexamethasone were given once per consolidation cycle as CNS prophylaxis, with normal CSF throughout 1.

Transplant. Busulfan-cyclophosphamide conditioning, peripheral blood stem cells at 3.4 x 10^6 CD34+ cells/kg. Immunosuppression stopped at 9 months with no chronic GVHD. At 24 months post-transplant, marrow MRD was negative by flow and PCR with complete donor chimerism (concept-hsct-in-kmt2a-r-leukemia).

Relapse, at 32 months post-transplant. Fever and headache. Marrow blasts 48.46%, transcript 26.54%, donor chimerism down to 72%. MRI showed diffuse abnormal signal in the cervical vertebral bodies, adjacent soft tissue and osseous structures. CSF flow cytometry showed 64% blasts - overt CNS relapse (concept-cns-involvement-at-diagnosis), in a patient who had received CNS prophylaxis and whose CSF had been normal.

Salvage. Azacitidine 75 mg/m2 days 1-7, venetoclax 100 mg daily days 1-14, and revumenib 160 mg twice daily from day 6 - started on day 6 rather than day 1 specifically to lower the tumour burden first and reduce differentiation-syndrome risk, and dose-adjusted for concomitant posaconazole, a strong CYP3A inhibitor. Donor lymphocyte infusion at 2.5 x 10^7 CD3+ T cells/kg two days after the azacitidine-venetoclax started. One cycle of intrathecal cytarabine and dexamethasone was given during the first consolidation because of the CNS burden, and further intrathecal treatment was withheld once remission was achieved.

Result. Complete haematologic remission 34 days after starting, MRD-negative by flow and PCR, CSF MRD negative, full donor chimerism, and a PET-CT showing no hypermetabolic lesion and no cervical bone destruction. Toxicity: grade IV myelosuppression with neutrophils under 0.5 x 10^9/L for 19 days, infectious fever, grade II nausea and vomiting, grade II hepatic impairment. No differentiation syndrome and no QTc prolongation. He remained disease-free on revumenib maintenance at six months 1.

This is the only entry in the arm in which a menin inhibitor is actually given, and it is given into the compartment the graph worries about most.

1. CNS disease responded. CSF blasts went from 64% to MRD-negative. The authors are careful, and so is this entry: the patient also received one intrathecal treatment, so revumenib's CNS penetration cannot be established from this case, and the paper says so explicitly. What the case does establish is that this combination was compatible with clearing a heavy CNS burden.
2. The relapse came at 32 months, after 24 months of documented MRD negativity. Falling donor chimerism (100% to 72%) accompanied it. A patient can be molecularly clean for two years and relapse in the third, in the CNS, with a headache.
3. The sequencing of revumenib was deliberate. Six days of azacitidine-venetoclax first to reduce burden, then the menin inhibitor - a strategy to avoid differentiation syndrome, which occurred in 16-27% of patients in AUGMENT-101 and did not occur here.
4. It contributes the AFDN partner to this arm in an adult, alongside a paediatric-adjacent AFDN case with a completely different phenotype.

The standing caution: the response was to a four-part regimen given alongside a donor lymphocyte infusion, so the contribution of the menin inhibitor specifically cannot be isolated. AUGMENT-101 gave revumenib monotherapy a CR/CRh rate of 24.5% (12/49) in relapsed or refractory KMT2A-rearranged AML, which is the honest single-agent prior.