Pediatric AML before transplant: read the uncertainty behind the MRD signal

Six MRD-positive children can produce a striking association and an uncertain estimate. A practical appraisal of what a retrospective transplant study supports, and what it cannot settle.

Original CFCM appraisal graphic: single-centre retrospective cohort of 80 children, six MRD-positive before transplant, sparse subgroup and wide uncertainty; prognostic association does not establish treatment benefit. No patient-level data. — Original CFCM conceptual graphic. No patient-level data or publisher artwork reproduced. Educational appraisal, not a clinical algorithm.

What the study actually compared

Deng and colleagues report a single-centre retrospective cohort of 80 children with AML transplanted in first complete remission. Only six were flow-MRD positive immediately before transplantation. Pre-transplant positivity was associated with poorer survival outcomes, but the estimates were imprecise: Table 2 reports a disease-free-survival hazard ratio of 14.304 with a 95% confidence interval of 1.892–108.155. The authors call the multivariable analysis exploratory because of the limited events and small positive subgroup. This comparison does not establish the benefit of a treatment change.

CFCM appraisal: start with the question

When discussing an MRD paper, write the proposed inference in one sentence before reviewing its most striking result. Predicting outcome, transferring a laboratory method and demonstrating that an intervention improves outcome are different questions. A result can motivate the next study without answering all three.

For a journal-club record, keep the population, sampling time, measurement procedure and endpoint on the same page. Add a separate line for the intervention claim, if any. That small separation makes it easier to notice when a discussion has moved beyond the comparison the investigators performed.

Keep uncertainty visible

A practical appraisal sheet should place the estimate beside its interval and the number of informative patients. Avoid a slide that carries only the largest number. Ask which uncertainties remain even if the reported association is accepted: the magnitude in another cohort, performance in another laboratory and the consequences of acting on the finding.

This is an editorial reading discipline, not an instruction to ignore a positive result. The point is to preserve the uncertainty when translating a publication into a local discussion. A population-level association should not become a patient-specific numerical forecast without the evidence required for that use.

Separate technical transfer from clinical interpretation

Before considering any published method locally, compare its specimen, analysis definitions and sampling context with the intended procedure. Record differences as questions requiring verification. Reproducing a marker list or a stated threshold is not, by itself, an evaluation of the complete workflow.

For clinical use, follow the laboratory's validated method and the treating team's established decision process. This article proposes questions for appraisal; it does not recommend delaying transplantation, adding treatment or changing an MRD interpretation for an individual child.

Daily Learning Pill

Carry the interval and the subgroup size into the discussion. A strong association can be useful evidence while its magnitude and clinical use remain unsettled.

Primary source and review scope

Shaoyang Deng et al. Frontiers in Oncology (2026-07-17). Publisher full text; DOI 10.3389/fonc.2026.1864716.

Actual source scope: Publisher full HTML: Methods including patient selection, monitoring and statistical definitions; Results including Tables 1 and 2 and figure captions; Discussion and limitations. Supplementary source data were not reanalysed. Noted prose/table inconsistency in haplo-HSCT counts; those counts are not used here.

Scientific and editorial review: ChatGPT/Codex preparing editor, 21 October 2026. No independent, human or clinical validation is asserted. Study observations and CFCM editorial appraisal are distinguished above. Original CFCM artwork; no third-party figure reproduced.