Two clones, two antibodies: a four-month-old with refractory KMT2A::AFF1 mixed-phenotype leukaemia treated with blinatumomab and gemtuzumab together

The report gives five presenting features and they are, unusually, all the textbook ones at once: asthenia, pallor, cutaneous haemorrhages, hepatosplenomegaly, and obstructive respiratory signs of the upper airway 1.

• The asthenia and pallor are concept-pallor-and-fatigue-anaemia - a four-month-old who is tired and pale, which is exactly what a parent notices and exactly what is easiest to attribute to something else.
• The cutaneous haemorrhages are concept-bruising-petechiae-thrombocytopenia, and here they sat on top of grade 3 disseminated intravascular coagulation.
• The hepatosplenomegaly is concept-hepatosplenomegaly-infant-leukemia.
• The white cell count reached 762 x 10^9/L with spontaneous tumour lysis syndrome, without renal failure - concept-hyperleukocytosis-emergency arriving before any treatment had been given.

A non-SARS coronavirus, HCoV-NL63, was found on the nasopharyngeal swab. The report does not connect it to anything; it is recorded here because it is the kind of incidental positive that can explain away a sick infant for another day.

Marrow was hypercellular with 80% lymphoblasts and 9% cells resembling myeloblasts with fine granulations but peroxidase-negative, plus atypical dysgranulopoiesis 1.

Flow cytometry resolved two distinct populations, both CD34+ and CD135+ (FLT3):

The authors note the crossover explicitly: the two populations form a continuum of expression rather than two clean compartments. This is the lineage-ambiguity that concept-flow-cytometry-immunophenotype-kmt2a-r describes as an intrinsic property of KMT2A-rearranged disease, seen at its most extreme.

FISH found KMT2A rearrangement in 93% of nuclei (concept-fish-split-signal-kmt2a) and RT-PCR confirmed KMT2A::AFF1 (MLL-AF4) (concept-molecular-confirmation-fusion-transcript). A FLT3 missense mutation was present at 43% allelic frequency. CSF status was CNS3 - overt CNS disease at diagnosis (concept-cns-involvement-at-diagnosis).

Chemotherapy failed, three times. On the Interfant-06 high-risk arm she was a poor prednisone responder with 127.9 x 10^9/L blasts at day 8. The CSF normalised after three intrathecal injections, but remission was not achieved: an M2 marrow (9% blasts) at the end of induction, progression to M3 (32% blasts) fifteen days into the IB consolidation, then failure of high-dose cytarabine with amsacrine, with flow showing both clones still present (80% B-lymphoid, 20% myeloid) 1.

Then two antibodies, in a deliberate order. After parental consent and national regulator authorisation, the team attacked each clone separately. Blinatumomab (CD19/CD3) went first, at 5 then 15 micrograms/m2/day, precisely so that T cells could expand before gemtuzumab ozogamicin - a CD33-directed calicheamicin conjugate that is cytotoxic to those same T cells - was added from day 11, at 3 mg/m2 on days 11, 15 and 18, with triple intrathecal therapy every 15 days.

Toxicity: generalised seizures at day 25 in a febrile context lasting 5 minutes, with all radiological, biological and microbiological investigation negative; blinatumomab was stopped, restarted at day 30 and continued to day 73. The only other grade 3/4 toxicity was an undocumented febrile neutropenia. No sinusoidal obstruction syndrome and no liver toxicity, which is the feared combination when gemtuzumab precedes a busulfan-containing transplant.

It worked. CR1 at day 38 of blinatumomab/gemtuzumab, with MRD undetectable by flow and by genomic PCR built on the KMT2A breakpoint. Transplant at 9 months of age after busulfan, fludarabine and thiotepa with defibrotide prophylaxis; full donor chimerism at months 1 and 2; ciclosporin stopped at month 3; MRD negative again at month 6 (concept-hsct-in-kmt2a-r-leukemia).

Then it relapsed. Eleven months after transplant, at 21 months of age, with invasion of the sinuses and orbits, CNS2 CSF and 54% marrow blasts. The relapse clone was pro-B with the original phenotype (CD19 bright, CD10-, partial CD33, no monocytic antigens) - the myeloid clone did not come back. KMT2A rearrangement and the FLT3 mutation persisted; a monoallelic complete IKZF1 deletion had been acquired; and FLT3 (CD135) was expressed six times more strongly than at diagnosis.

Salvage was intensive chemotherapy with twice-daily midostaurin - chosen on the FLT3 finding - plus triple intrathecal therapy, achieving marrow and extramedullary CR2 with undetectable breakpoint MRD, followed three months later by autologous CD19 CAR T cells with grade 2 cytokine release syndrome and haematological and neurological toxicity.

Outcome as reported: 36 months old, in persistent MRD-negative CR2 twelve months after the CAR-T infusion, with no known sequelae.

This is the graph's only entry in which a named therapeutic strategy is chosen from first principles at the bedside and then tested on one child, and the reasoning is written down.

1. Lineage plasticity is not an abstraction. Two CD34+CD135+ populations, a continuum between them, and a relapse in which only one of the two came back. risk-lineage-switch-cd19-directed-therapy says that CD19-directed pressure can select a myeloid escape; here the pre-existing myeloid clone was co-targeted on purpose, and the relapse was lymphoid.
2. The sequencing of two immunotherapies was a mechanistic argument. Blinatumomab needs functional effector T cells; gemtuzumab kills them. Giving blinatumomab an 11-day head start is a hypothesis about drug interaction, executed in a patient, with the reasoning stated.
3. A second lesion changed the second-line plan. FLT3 mutant at 43% at diagnosis, CD135 six-fold brighter at relapse, midostaurin at salvage. The fusion is the driver; it was not the only thing that mattered.
4. Survival here is a sequence, not an event. Failed induction, failed consolidation, failed high-dose cytarabine, two antibodies, a transplant, a relapse, midostaurin, CAR-T. A child alive at 36 months in CR2 has been through eight lines of treatment before her third birthday.

It is also the closest thing in this arm to a long-term survivor of refractory KMT2A::AFF1 infant leukaemia - with the honest qualification that 36 months of age, 12 months after CAR-T, is not the same as cure.

The abstract of this paper describes the patient as "a 4-year-old infant". The full text says "Four-month-old little girl" and reports transplant at 9 months of age and relapse at 21 months of age, so the abstract's "4-year-old" cannot be right; it is a typographical error for four months. Anyone extracting ages from abstracts - including automated pipelines - will get this one wrong. This entry uses the full text 1.