Analysis
Clinical Flow
Ph-positive B-ALL MRD: when three methods disagree
A small pediatric study compared different residual-disease targets; samples, patients, specimens and detection limits must not be conflated.
At the end of treatment induction, a laboratory may have three different views of residual leukemia: an immunophenotype by flow cytometry, BCR::ABL1 transcripts by reverse-transcription PCR, and an immunoglobulin rearrangement tracked by sequencing. A disagreement is not automatically a laboratory error. The methods follow different biological targets and their practical detection limits depend on the assay, material and sampling time.

Chang and colleagues examined this problem in children with Philadelphia chromosome-positive B-cell acute lymphoblastic leukemia. The study is an informative opportunity to ask how different evidence should be reconciled. Its small selected cohort and incomplete description of the flow comparator prevent it from establishing a universal ranking of clinical assays.
Start with the denominator
The investigators enrolled 55 children. Six did not contribute to the post-induction evaluation for the reasons described in the paper; among the remaining 49, 48 had a trackable immunoglobulin rearrangement and were included in subsequent analysis. Seventeen of these 48 experienced relapse during follow-up. The reported median follow-up was 50 months.
One hundred post-treatment samples were assessed by immunoglobulin sequencing, flow cytometry and RT-PCR. These are samples, not 100 independent children. The molecular method was applied to frozen marrow or peripheral blood; the report describes the conventional comparators as marrow-based. The possible effect of specimen choice, sampling time and repeated sampling must remain visible when the methods are compared.
What the discordance shows
Twenty-five of the 100 samples gave different results by flow and immunoglobulin sequencing: three were flow positive and sequencing negative, while 22 were flow negative and sequencing positive. The direction of discordance is relevant, but the table does not by itself establish which assay is clinically correct in every case. Differences in tracked biology, cell or DNA input, analytic threshold and the timing of sample collection can all affect interpretation.
The immunoglobulin assay tracks an identified diagnostic rearrangement. The authors describe an analytical sensitivity of up to one in a million, dependent on DNA input. That ceiling is not a guarantee for every specimen. The paper does not provide enough detail about the flow assay's panel, event acquisition or per-sample detection limits to infer that all flow results had the same sensitivity, or to generalize the relative performance to other laboratories.
A small warning signal, not a treatment trial
The authors describe seven relapsing children with a measurement within the previous year. Five were tested within 150 days before relapse: all five samples were positive by immunoglobulin sequencing, three by RT-PCR and one by flow. This is a selected subgroup of five samples, not a reliable estimate of average lead time or evidence that changing therapy earlier would improve survival.
The sequencing platform examined in the study was supplied by NeoImmune, and several authors were employees of that company. That relationship should be stated when interpreting the results. Independent replication with matched sample types, documented assay performance and prospectively defined endpoints would strengthen the evidence.
What changes in the laboratory discussion?
Keep each result linked to its target, specimen, time point, input and validated reporting limit. Check that the tracked rearrangement remains informative, review gating and sample adequacy, and seek the clinical and molecular context rather than collapsing discordant results into a single unsupported answer. Neither a negative flow result nor a sequencing signal alone explains every possible residual clone.
Daily Learning Pill: A comparison of methods on 100 samples is not a comparison on 100 independent patients, and different molecular and cellular targets do not have interchangeable denominators.
Primary source: Chang L et al., Frontiers in Immunology 16:1677013, published 2 February 2026, https://doi.org/10.3389/fimmu.2025.1677013. The article is available under CC BY; no publisher figure is reused in this draft.