Measure the RAS allele fraction in CSF blasts and in paired marrow blasts from the same infant on the same day
also called compartment-resolved RAS genotyping in infant KMT2A-r ALL · the paired CSF/marrow experiment
Every genetic study of this leukaemia has sampled the bone marrow. This experiment asks whether the cells carrying the RAS mutation are actually somewhere else - in the spinal fluid - by genotyping both compartments from the same child at the same moment.
The same thing, for a scientist
A direct test of `hyp-ras-subclone-is-a-dissemination-modifier` and, at the same time, a test of whether the clonality literature in this disease has a sampling artefact at its centre. **Design.** Prospective or banked-sample study in infants with newly diagnosed KMT2A-rearranged B-ALL. At diagnostic lumbar puncture (performed routinely as part of CNS staging and intrathecal therapy) and at the same-day marrow aspirate, collect paired specimens. In every patient with detectable CSF blasts, perform **targeted duplex or error-corrected deep sequencing** of KRAS, NRAS, PTPN11, FLT3, NF1 and PIK3CA on both compartments, alongside a KMT2A-fusion-specific quantitative assay in both to normalise for leukaemic cell content. **The primary comparison is not the mutation's presence but its fraction OF LEUKAEMIC CELLS**, which is why the fusion assay is mandatory: a CSF specimen is a low-cellularity sample with variable blood contamination, and a raw variant allele fraction there is uninterpretable without a denominator that counts leukaemic genomes. **Readout.** For each patient, the ratio (RAS variant allele fraction / fusion-normalised leukaemic fraction) in CSF divided by the same ratio in marrow. The hypothesis predicts a ratio greater than 1 in most patients; the rival - that subclonal RAS lesions are prognostically inert noise - predicts a distribution centred on 1. **Why it is feasible.** Diagnostic lumbar puncture is already universal in infant ALL protocols. CNS involvement is common in this disease. Error-corrected sequencing on low-input material is established and has already been applied to this exact disease by the Lund group, on 27 patients [E3]. Nothing in the design requires a new procedure on a sick infant.
What makes this worth doing before anything more elaborate
It tests an assumption shared by every published clonality claim in this disease. The subclonal fraction, the 'lost at relapse' finding, the MAF clusters, the 6.25% relapse-founding clone - all of it comes from marrow or blood 1. If the leukaemia's compartments differ in composition, those claims are claims about one compartment.
It uses material that already exists. Interfant and COG banks hold diagnostic CSF cytospins and cell pellets alongside marrow. The experiment could in principle be run retrospectively on banked material, at the cost of DNA quality from cytospin slides.
It has an informative negative. If CSF and marrow allele fractions match, the dissemination hypothesis is falsified cleanly, the marrow-based literature is validated, and the graph gains a bounded reason to stop asking.
And it is one of the very few experiments in this beat that does not require a new drug, a new model or a new trial.
Controls and the ways this could produce a false positive
Blood contamination of CSF is the dominant artefact. A traumatic tap introduces peripheral blasts, which would make CSF look like blood rather than like a distinct compartment - biasing toward the null, not toward the hypothesis, but degrading power. Record red-cell count and exclude or model traumatic taps.
Low leukaemic input in CSF inflates variance in allele fraction estimation. Duplex sequencing controls the error rate but not the sampling noise: a CSF sample with 200 leukaemic genomes cannot estimate a 10% allele fraction precisely. Report confidence intervals per sample and pre-specify a minimum leukaemic-genome input.
Selection by CNS status. Patients with CSF blasts are not a random sample of the disease; they have higher-risk features. The comparison is within-patient, which protects the primary readout, but the generalisation to all RAS-mutant patients does not follow and should not be claimed.
The positive control that makes the result interpretable. Genotype NG2 expression or a NG2-associated marker in the same specimens. NG2 is highly upregulated in blasts infiltrating extramedullary sites and CNS in this disease, with a migratory gene signature, and specific NG2 blockade almost abolishes engraftment 4. If the CSF compartment shows NG2 enrichment and no RAS enrichment, the compartment biology is real and RAS is not part of it - which is the most likely informative negative.
| design | Within-patient paired comparison of cooperating-lesion allele fraction between CSF and bone marrow, in infants with newly diagnosed KMT2A-rearranged B-ALL and detectable CSF blasts. Targeted error-corrected (duplex) sequencing of KRAS, NRAS, PTPN11, NF1, FLT3 and PIK3CA on both compartments, with a KMT2A-fusion-specific quantitative assay in both to normalise for leukaemic genome content. Positive control: NG2 expression or an NG2-associated marker measured on the same specimens. |
|---|---|
| readout | Per patient, the ratio of (variant allele fraction / fusion-normalised leukaemic fraction) in CSF to the same quantity in marrow. Primary analysis: the distribution of that ratio across patients, with per-sample confidence intervals reflecting leukaemic-genome input. |
| controls | Red-cell count on every CSF specimen with traumatic taps pre-specified for exclusion or modelling; pre-specified minimum leukaemic-genome input per sample; germline sample to exclude inherited variants; NG2 as a positive control for compartment biology existing at all; marrow-versus-peripheral-blood pairs in the same patients as a negative control for compartment effects that are really just specimen effects. |
| cost tier | med |
| timeline | 12-24 months prospectively within an existing infant ALL protocol's sample collection; potentially faster on banked CSF pellets at the cost of DNA quality |
| status | proposed |
| feasibility notes | Diagnostic lumbar puncture is already universal in infant ALL protocols; no additional procedure on the patient is required. The rate-limiting step is CSF cellularity, which is the reason a pilot establishing minimum input must precede the main study. |
| informative negative | Equal allele fractions in the two compartments falsifies hyp-ras-subclone-is-a-dissemination-modifier cleanly and validates the marrow-based clonality literature in this disease. |
- A positive result would make CNS-directed MEK inhibition the implied therapy, and the available MEK inhibitors have poor CNS penetration - so the experiment could identify a target in a compartment the drugs do not reach.
- Genotyping CSF blasts creates a new stratification variable in a disease whose protocols are already toxic, and the immediate clinical temptation would be to intensify CNS therapy in RAS-mutant infants on the basis of a compartment allele fraction rather than an outcome.
- If the result is positive, every existing clonality claim in this disease becomes a claim about the marrow compartment, which would require revising language in several entries of this graph and in the literature it summarises.
- The experiment cannot distinguish a genuinely compartment-restricted clone from a clone that is merely better at entering CSF transiently, and the therapeutic implications of those two are different.
- What is the minimum leukaemic-genome input from a diagnostic infant CSF specimen, in practice? This determines whether the experiment is feasible at all and no published figure was found.
- Do Interfant or COG banks hold diagnostic CSF cell pellets, or only cytospin slides? The answer decides between a retrospective and a prospective design.
- Should the same design be applied to testis and other sanctuary sites, which are sampled in this disease only at relapse if ever?
- If RAS allele fraction is equal across compartments but NG2 is not, what does that say about whether any genetic lesion in this disease has a compartment identity?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
All clonality and relapse-evolution measurements in the index genomic study of infant MLL-rearranged ALL were made on diagnostic leukaemia cells and matched remission bone marrow or peripheral blood, with mutant allele frequencies clustered and purity-adjusted from those compartments.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, Wang J, et al. · Nature Genetics 2015What this evidence cannot show
Read from the methods and results of the full text via efetch db=pmc (PMC4553269) on 2026-09-20. The paper does not state that CSF was excluded; it states which materials were used, and no CSF analysis appears anywhere in the text. This is the basis for saying the compartment question is untested rather than answered negatively.
Evidence that argues against this (1)
- Marrow and peripheral blood are the standard specimens for leukaemia genomics because they contain enough material; the absence of CSF genomics reflects a practical constraint and is not evidence that anyone considered and rejected the question. Andersson AK et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Activated KRAS(G12V) in MLL-AF4-expressing human cord-blood CD34+ HSPCs enhanced extramedullary engraftment, leukocytosis and CNS infiltration with a cell-migration transcriptional signature, while failing to initiate leukaemia - the experimental basis for expecting a compartment difference in RAS allele fraction.
Activated KRAS Cooperates with MLL-AF4 to Promote Extramedullary Engraftment and Migration of Cord Blood CD34+ HSPC But Is Insufficient to Initiate Leukemia. · Prieto C, Stam RW, Agraz-Doblas A, et al. · Cancer Research 2016What this evidence cannot show
The engrafted cells in that model are MLL-AF4-expressing HSPCs, not leukaemic blasts, so the model shows KRAS altering where transduced progenitors go rather than where a leukaemia goes. Immunodeficient mice have altered blood-brain-barrier properties and no adaptive immunity. The design proposed here is the patient-level test the model motivates and cannot substitute for.
Evidence that argues against this (1)
- NG2 provides an alternative and better-evidenced route to CNS infiltration in this disease, and NG2 upregulation in CNS-infiltrating blasts was shown directly in patient-derived xenografts rather than in a transduced-progenitor model. Prieto C et al. 2018
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Targeted duplex sequencing has been applied to infant and childhood KMT2A-rearranged ALL on 27 patients including 5 longitudinal and 6 paired relapse sample sets, detecting cancer-associated variants at 0.06-0.34% cellular prevalence, which establishes the assay's feasibility and sensitivity in this disease.
Duplex Sequencing Uncovers Recurrent Low-frequency Cancer-associated Mutations in Infant and Childhood KMT2A-rearranged Acute Leukemia. · Pilheden M, Ahlgren L, Hyrenius-Wittsten A, et al. · HemaSphere 2022What this evidence cannot show
Feasibility was demonstrated on marrow or blood, which are high-input materials. CSF is low-input and low-cellularity, and duplex sequencing consumes input molecules to build consensus, so the sensitivity achieved on marrow should not be assumed for CSF. A pilot establishing minimum leukaemic-genome input from CSF is a prerequisite, not an optional step. Abstract read.
Evidence that argues against this (1)
- Duplex sequencing's error suppression does not address sampling noise, so a low-input CSF specimen can return a precise-looking allele fraction with a very wide true confidence interval - the failure mode most likely to produce a spurious compartment difference. Pilheden M et al. 2022
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In infant MLL-rearranged B-ALL, NG2 expression was highly upregulated in blasts infiltrating extramedullary haematopoietic sites and the CNS, gene expression profiling of primary blasts and primografts revealed a migratory signature of NG2-positive blasts, specific NG2 blockade resulted in almost complete loss of engraftment, and high NG2 expression in the Interfant cohort associated with lower event-free survival, more circulating blasts and more frequent CNS disease or relapse.
NG2 antigen is involved in leukemia invasiveness and central nervous system infiltration in MLL-rearranged infant B-ALL. · Prieto C, Lopez-Millan B, Roca-Ho H, et al. · Leukemia 2018What this evidence cannot show
The same paper reports that NG2 is a MALLEABLE marker that does NOT enrich for leukaemia-initiating or CNS-infiltrating cells on serial xenotransplantation - so NG2 marks a state rather than a clone, which is exactly the distinction this experiment needs to keep separate from a genetic clone. Abstract read; full text not retrieved for this entry.
Evidence that argues against this (1)
- Because NG2 is explicitly reported as a malleable state marker rather than a clonal one, a compartment difference in NG2 does NOT predict a compartment difference in a genetic lesion, and the two could dissociate completely - which is the most likely outcome of this experiment and the reason to run the NG2 control alongside it. Prieto C et al. 2018
the compartment prediction, testable in material infant protocols already collect
whether the clonality literature in this disease has a sampling-compartment artefact at its centre