Retrospective paired-target MRD on stored Interfant marrows, stratified by IGHV usage
also called paired MRD experiment · IGHV6-1 MRD failure test
Use marrow samples that were already collected and already tested two ways, and ask whether the antibody-gene test failed more often in the babies whose leukaemia used the restricted antibody gene segment. No new patients and no new samples are needed.
The same thing, for a scientist
Interfant-06 measured MRD in 249 KMT2A-rearranged infants with the KMT2A genomic breakpoint as the main target and IG/TR as additional targets, and reported 19 patients positive above cut-off by KMT2A and negative by IG/TR against 12 the other way [E1]. Separately, KMT2A::AFF1 shows proximally biased IGHV usage with IGHV6-1 in up to 50% of samples, a median of three DNJH stems per sample and short CDR3 [E2]. **These two datasets have never been joined.** The experiment is to take the stored diagnostic and follow-up material from an Interfant cohort, determine IGHV segment usage and DNJH-stem count at diagnosis from the archived diagnostic DNA, and test whether ASO-PCR target failure - defined as loss of quantifiability, discordance with the KMT2A target of at least one log, or inability to design a target at all - is concentrated in patients with IGHV6-1 usage, with multiple DNJH stems, or with short CDR3.
Design
Cohort. Archived diagnostic and protocol-timepoint DNA from KMT2A-rearranged infants on an Interfant protocol with both a KMT2A breakpoint target and at least one IG/TR target, plus - critically - the patients in whom an IG/TR target could not be established, who are absent from every published concordance analysis by construction 1.
Exposure. From diagnostic DNA: IGHV segment usage, number of clonally related DNJH stems, CDR3 nucleotide length. Measured by IGH amplicon sequencing with marker-DNJ-stem assignment, the method that detected clonal evolution in a third of 280 adults with BCP-ALL 3.
Outcome. Three pre-specified ASO-PCR failure endpoints: (a) no IG/TR target obtainable at diagnosis; (b) achieved quantitative range worse than 5x10^-4; (c) at least one log lower than the paired KMT2A breakpoint result at any protocol timepoint.
Analysis. Logistic regression of each failure endpoint on IGHV6-1 usage, stem count and CDR3 length, adjusting for age at diagnosis, partner gene and timepoint. Pre-specify the direction: the hypothesis predicts failure is more common with IGHV6-1 usage, more stems and shorter CDR3.
Controls. The same analysis in KMT2A-germline infants from the same protocol, who have Ig/TR targets and no KMT2A fusion 4, and in a non-KMT2A paediatric BCP-ALL comparator. If the association appears in all three groups it is a property of Ig repertoire biology and not of this subtype.
What each result would mean
- Failure concentrated in IGHV6-1 users: the proximal bias is a clinically actionable assay hazard, and a subtype-specific primer-design strategy is warranted.
- Failure concentrated in multi-stem patients and not in IGHV6-1 users: the problem is oligoclonality and ongoing recombination, and the fix is sequencing-based tracking of the whole clonal family rather than a single ASO target 3.
- No association: the mechanism described in KMT2A::AFF1 does not translate into assay failure, and the Interfant preference for the genomic breakpoint rests on target absence and regenerating-marrow false positivity rather than on repertoire biology. This is a real possible outcome and would be worth publishing.
The experiment is cheap because every input already exists. Its binding constraint is access to the Interfant biobank and the linked MRD laboratory records, not money or technology.
| design | retrospective cohort on archived Interfant diagnostic and follow-up DNA with paired KMT2A-breakpoint and IG/TR MRD results |
|---|---|
| readout | three pre-specified ASO-PCR failure endpoints (no target obtainable; quantitative range worse than 5e-4; >=1 log below the paired KMT2A result) regressed on IGHV6-1 usage, DNJH-stem count and CDR3 length |
| controls | KMT2A-germline infants from the same protocol; non-KMT2A paediatric BCP-ALL comparator |
| cost tier | low |
| timeline | 12-18 months, dominated by biobank access and governance rather than by laboratory work |
| feasibility risk | whether the trial records retain the patients in whom no IG/TR target could be established, and whether archived infant DNA supports IGH amplicon sequencing at depth |
tests hypotheses
[ "hyp-ig-mrd-fails-where-the-repertoire-is-proximally-biased" ]
- If the association is found, the immediate clinical implication is that certain infants should be moved off Ig/TCR MRD - which in centres without breakpoint access means moving them onto nothing better, so a positive result creates an obligation the health system may not be able to meet.
- A retrospective analysis of assay failures in a trial's own laboratories asks those laboratories to surface their own failures, which is a governance problem as much as a scientific one.
- A negative result would be read as reassurance that Ig/TCR MRD is fine in this subtype, when the study would only have excluded one of several mechanisms by which it can fail.
- Do the Interfant MRD laboratory records retain assay-design failures, or only the assays that succeeded? If only the latter, the most informative endpoint is unavailable and the study shrinks to a discordance analysis.
- Is archived infant diagnostic DNA from 2006-2016 of sufficient quality for IGH amplicon sequencing, and does consent cover it?
- Should the exposure be measured on DNA or on RNA? The KMT2A::AFF1 finding is an RNA-seq read-count measurement and the MRD assay is DNA-based, so a DNA-side replication of the proximal bias is a prerequisite step.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Interfant-06 measured MRD in 249 infants with KMT2A-rearranged ALL using the KMT2A genomic breakpoint as the main PCR target and immunoglobulin and T-cell receptor rearrangements as additional targets, with results available for both targets in 223 samples, and found 19 patients positive above protocol cut-offs by KMT2A and not IG/TR against 12 in the reverse direction.
Clinical Implications of Minimal Residual Disease Detection in Infants With KMT2A-Rearranged Acute Lymphoblastic Leukemia Treated on the Interfant-06 Protocol. · Stutterheim J, van der Sluis IM, de Lorenzo P, et al. · J Clin Oncol 2021What this evidence cannot show
Read in full text. The paper does not report how many patients had no usable IG/TR target, which is the group this experiment most needs; whether that information survives in the trial records is unknown and is the main feasibility risk.
Evidence that argues against this (1)
- Overall concordance was 94% with no significant discrepancy at end of induction, so the failure events this experiment is powered on are rare and the study may be underpowered unless it pools Interfant-99 and Interfant-06. Stutterheim J et al. 2021
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 48 adult, 36 paediatric and 37 infant KMT2A::AFF1 BCP-ALL cases, proximal IGHV genes were strongly favoured with IGHV6-1 significantly elevated in all age classes and present in the B-cell receptors of up to 50% of KMT2A-rearranged samples, with a median of three DNJH stems per sample and shorter nucleotide CDR3 regions than most other BCP-ALL subtypes.
Proximally biased V(D)J recombination in the clonal evolution of IGH alleles in KMT2A::AFF1 BCP-ALL of all age classes. · Müller H, Dicker F, Bär C, et al. · Hemasphere 2024What this evidence cannot show
Read in full text. RNA-seq repertoire analysis of diagnostic samples only, with no remission samples and no MRD assay performed, so the exposure variable this experiment uses is well measured and its link to assay failure is entirely untested.
Evidence that argues against this (1)
- A different group's deep BCR sequencing of t(4;11) infant samples found no significantly expanded VDJ clones at all, which would make IGHV segment usage hard to define per patient and could render the exposure variable unmeasurable in some infants. Agraz-Doblas A et al. 2019
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
High-throughput IGH sequencing with clonally related sequences identified by a shared DNJ-stem detected clonal evolution at diagnosis in one third of 280 adults with BCP-ALL, with high prevalence and degree of evolution driven by ongoing DH/VH-DJH recombination associated with the presence of KMT2A rearrangements.
Insights into IGH clonal evolution in BCP-ALL: frequency, mechanisms, associations, and diagnostic implications. · Darzentas F, Szczepanowski M, Kotrová M, et al. · Front Immunol 2023What this evidence cannot show
Adults, and the marker-DNJ-stem method is introduced by the authors rather than independently validated. Applying it to archived infant diagnostic DNA assumes the material supports amplicon sequencing at adequate depth, which for stored infant samples is not guaranteed.
Evidence that argues against this (1)
- The same paper attributes much BCP-ALL clonal evolution to VH replacement, which was enriched in Ph-like and DUX4 rather than KMT2A-rearranged disease, so stem count may not be the right exposure variable for this subtype. Darzentas F et al. 2023
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 167 infants with KMT2A-germline ALL on Interfant-06, MRD was measured by real-time quantitative PCR in 73 patients at end of induction, end of consolidation and before OCTADAD, and high end-of-induction MRD was associated with worse 6-year disease-free survival (61.4% versus 87.9% for undetectable MRD, P = 0.02).
Minimal residual disease and outcome characteristics in infant KMT2A-germline acute lymphoblastic leukaemia treated on the Interfant-06 protocol. · Stutterheim J, de Lorenzo P, van der Sluis IM, et al. · Eur J Cancer 2022What this evidence cannot show
This group has no KMT2A fusion target by definition, so it serves as a control for Ig/TR assay behaviour without a paired comparator - meaning failure endpoint (c) cannot be assessed in them, only (a) and (b).
Evidence that argues against this (1)
- KMT2A-germline infant ALL is biologically different, with age under 6 months favourable rather than adverse, so it is an imperfect control for repertoire biology in the rearranged subtype. Stutterheim J et al. 2022
the retrospective test on stored Interfant material
- related to →MRD by genomic breakpoint: the only assay in this disease that is fusion-specific by construction
uses the paired breakpoint target as the reference measurement