Stratify an existing menin-inhibitor cohort by reciprocal fusion status
also called does AF4-MLL status predict menin-inhibitor response · one RT-PCR on banked samples
Roughly half of these children carry a second fusion protein that the menin drug cannot touch. Nobody records which half a patient is in. The test is one cheap PCR on samples that already exist.
The same thing, for a scientist
A retrospective analysis, not a new trial. In any menin-inhibitor cohort containing t(4;11) patients, assay banked diagnostic material for the AFF1::KMT2A reciprocal transcript and compare response rate, depth and duration between reciprocal-positive and reciprocal-negative strata. The assay is established: reciprocal transcript status was determined by RT-PCR in 124 Interfant-treated infants and carried a large prognostic signal on conventional therapy, 4-year EFS 62.4% versus 11.7% [E1]. Reciprocal-positive patients carry a KMT2A-derived oncoprotein with no menin-binding motif and an intact WIN/SET module [E2], so a menin inhibitor cannot engage it. This is the cheapest available test of `hyp-menin-primacy-was-argued-from-one-of-two-fusion-products`, and it also supplies the first clinical stratification variable this graph has proposed that is derived from fusion-protein architecture rather than from partner identity.
Design
Population. t(4;11)/KMT2A::AFF1 patients from any menin-inhibitor cohort with banked diagnostic RNA. Paediatric and adult may be pooled with age as a covariate; infant cases analysed separately if n permits.
Exposure. AFF1::KMT2A reciprocal transcript, detected/not detected by RT-PCR using the published assay design, run blind to outcome.
Endpoints. Overall response (CR/CRh), MRD negativity, duration of response, and time to progression. Duration is the priority: this graph's NUP98 work found that the informative clinical number in menin-inhibitor trials is a durability rather than a rate.
Covariates. Age, prior lines, blast burden, co-mutations (RAS pathway, FLT3, TP53), and KMT2A breakpoint intron, since breakpoint position determines reciprocal expression and may act independently.
Power. With a 45% prevalence of reciprocal positivity, detecting a hazard ratio of 2 for duration of response needs roughly 80-120 t(4;11) patients with banked material. No single trial has that; a pooled analysis across AUGMENT-101, the ziftomenib programme and the paediatric cohorts might.
Decision rule. A duration difference in the direction of worse outcomes for reciprocal-positive patients, surviving adjustment, supports the hypothesis and argues for recording reciprocal status at enrolment. No difference kills the hypothesis and strengthens the menin argument.
The confound that has to be handled explicitly
Reciprocal-positive patients do markedly better on conventional therapy 1. Any menin-inhibitor cohort is a relapsed/refractory population, so reciprocal-positive patients reaching it have already failed a therapy they were expected to do well on - a selection effect that could mask or invert the signal. The analysis must therefore condition on prior therapy and report the interaction, not just the main effect. A cleaner but slower alternative is a frontline cohort.
| design | Retrospective biomarker stratification of banked diagnostic RNA from menin-inhibitor-treated t(4;11) patients by AFF1::KMT2A reciprocal transcript status |
|---|---|
| readout | CR/CRh rate, MRD negativity, duration of response and time to progression by reciprocal status |
| controls | Blinded assay; adjustment for age, prior lines, co-mutations and KMT2A breakpoint intron; explicit reporting of the prior-therapy interaction |
| prevalence assumption PCT | 45 |
| estimated n required | 80-120 t(4;11) patients with banked material, pooled across programmes |
| cost tier | low |
| timeline | 6-9 months, dominated by material access rather than by assay time |
| priority endpoint | duration of response, not response rate |
- If reciprocal-positive patients respond better rather than worse, the finding would make reciprocal status an enrichment biomarker and could narrow eligibility for a drug that is currently available to all KMT2A-rearranged patients.
- A negative result would be read as closing the question, when a relapsed/refractory cohort is the setting least able to detect the effect because of the selection confound.
- Introducing any new stratification variable into paediatric eligibility risks delaying access for the stratum found to do worse, in a disease where the alternative is chemotherapy with an 11.7% four-year event-free survival.
- Do any menin-inhibitor programmes have banked diagnostic RNA from t(4;11) patients with consent permitting this analysis?
- Is reciprocal status stable between diagnosis and relapse? If the der(4) allele is lost during clonal evolution, diagnostic material may misclassify the relapsed disease actually being treated.
- Would a protein-level assay for the reciprocal product be feasible on banked material, or is RT-PCR the only option? Every clinical number on this axis is currently transcript-based.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Reciprocal AFF1::KMT2A transcript status was determined by RT-PCR in 124 Interfant-treated infants with B-ALL, was positive in 45% of t(4;11) cases, and predicted 4-year event-free survival of 62.4% versus 11.7% with significance retained in a risk-adjusted Cox model.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al. · Haematologica 2019What this evidence cannot show
Retrospective, conventional-therapy era, and transcript rather than protein. None of these patients received a menin inhibitor, so the prognostic effect measured is on a different treatment background from the one proposed here.
Evidence that argues against this (1)
- A CRISPR-engineered t(4;11) in human HSPCs with both products present shows lineage plasticity rather than a fixed phenotype, offering an explanation for the prognostic difference that is about cell state rather than about the reciprocal protein's activity. Benz T et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The reciprocal product retains the KMT2A C-terminus through the final exon - including the WIN motif and SET domain - while losing the N-terminal menin-binding motif, so it is architecturally outside the reach of a menin inhibitor.
The role of reciprocal fusions in MLL-r acute leukemia: studying the chromosomal translocation t(4;11). · Wilhelm A, Marschalek R · Oncogene 2021What this evidence cannot show
The exon composition is from that study's engineered constructs; patient-level composition varies with breakpoint, though not in a way that restores the menin-binding motif.
Evidence that argues against this (1)
- Marschalek's own data support a hit-and-run model in which 48 hours of reciprocal expression suffices and the protein is dispensable later, which would make its menin-blindness therapeutically irrelevant in a relapsed population. Wilhelm A et al. 2021
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
NUP98::KMT2A, which shares the menin-blind, SET-retaining architecture of the reciprocal product, is reported resistant to menin inhibition - the only existing clinical-adjacent observation on this architecture.
Transforming activities of the NUP98-KMT2A fusion gene associated with myelodysplasia and acute myeloid leukemia. · Fisher JN, Thanasopoulou A, Juge S, et al. · Haematologica 2020What this evidence cannot show
A mouse model of a fusion seen in two patients; extrapolating from it to t(4;11) reciprocal-positive patients is an architectural analogy, not evidence.
Evidence that argues against this (1)
- NUP98::KMT2A has reduced HOXA output and a very long latency, so it is a weak oncoprotein and its menin-inhibitor resistance may simply reflect that its leukaemia was never menin-driven, rather than that menin-blind architecture confers resistance. Kaltenbach S et al. 2010
same stratification applied to trial data rather than to a diagnostic archive
Specifies which assay that proposed reanalysis would have to use, and what it would miss.
Supplies the effect size and the era caveat that make the proposed reanalysis worth running.
That reanalysis would have to record breakpoint intron as well, or it cannot separate these models.
Establishes that the proposed reanalysis has no precedent anywhere in the indexed literature.
The retrospective stratification that would decide this hypothesis on banked material.
Stratifies menin-inhibitor response in t(4;11) patients by reciprocal AFF1::KMT2A transcript status on banked diagnostic material.
The cohorts with banked t(4;11) material are menin-inhibitor cohorts; duration of response is the priority endpoint.