Proposed Article Structure
This article is organized into the following sections: clinical background and unmet need; study design and methods; major results; interpretation and biological rationale; clinical implications for transplant decision-making and post-transplant management; strengths and limitations; and conclusion with future directions.
Highlights
In a large PETHEMA registry analysis of 717 adults with acute myeloid leukemia in first complete remission undergoing allogeneic hematopoietic stem cell transplantation, both ELN2017 and ELN2022 risk systems retained prognostic value in the transplant setting.
On multivariable analysis, DNMT3A, SF3B1, TP53, and WT1 mutations were associated with shorter relapse-free survival, whereas FLT3-ITD was associated with longer relapse-free survival after transplant.
The investigators integrated clinical and molecular variables into a prognostic score that included age, AML type, transplant timing, and genetic alterations, and then validated this score.
The findings support more granular pre-transplant and post-transplant risk stratification, with potential relevance for conditioning intensity, donor strategy, measurable residual disease surveillance, and maintenance therapy.
Background and Clinical Context
Acute myeloid leukemia (AML) remains a biologically diverse and clinically challenging hematologic malignancy. Although induction and consolidation regimens can achieve first complete remission in many patients, relapse remains the dominant cause of treatment failure. Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a central post-remission strategy for patients with intermediate- or adverse-risk disease, primarily because graft-versus-leukemia effects can lower relapse risk beyond what chemotherapy alone can achieve.
Yet allo-HSCT is not a biologic equalizer. Outcomes after transplant are still strongly shaped by disease genetics, remission depth, patient age and fitness, transplant timing, donor characteristics, and transplant-related toxicity. In routine care, European LeukemiaNet (ELN) risk classification guides much of the initial therapeutic decision-making, but it was developed mainly around diagnosis-level prognostication and not specifically to predict outcomes after allo-HSCT in first remission. As molecular profiling has expanded, a key question has become increasingly important: which mutations continue to matter after a patient has already reached remission and proceeds to transplant?
This issue is clinically relevant because transplant physicians increasingly need mutation-informed strategies before and after transplant. Examples include selecting conditioning intensity, deciding on measurable residual disease (MRD) testing platforms, prioritizing post-transplant maintenance, and planning surveillance in patients at especially high risk of relapse. The PETHEMA group study by Colmenares and colleagues directly addresses this need by evaluating how individual and co-mutational profiles influence outcomes in AML patients transplanted in first complete remission.
Study Design and Methods
This was a retrospective registry-based analysis using data from the PETHEMA registry. The investigators included 717 patients with AML who underwent allo-HSCT in first complete remission. The median age was 56.5 years, which is notable because it reflects contemporary transplant practice in an older adult AML population rather than a highly selected younger cohort. Most patients were classified as adverse risk by ELN2022 criteria.
The main outcomes evaluated were overall survival and relapse-free survival (RFS). Cox proportional hazards regression was used to identify variables independently associated with outcomes. Based on hazard ratios from the multivariable model, the authors developed a prognostic score that incorporated both clinical and molecular features. The score included age, AML type, transplant timing, and genetic or molecular alterations, and the authors report that it was validated.
Because this was a retrospective observational study, treatment allocation was not randomized and the analysis should be interpreted primarily as prognostic rather than causal. Even so, registry studies are often the most practical way to examine real-world transplant populations of this size, especially for molecular subgroup analyses that single centers are usually underpowered to assess.
Key Results
Validation of ELN Risk Classification in the Transplant Setting
An important initial finding was that both ELN2017 and ELN2022 classifications retained prognostic relevance in this cohort of patients already selected for allo-HSCT in first remission. This matters because there has been lingering uncertainty about whether transplant attenuates baseline genetic risk enough to make ELN categories less informative. The study suggests that ELN risk still carries meaningful signal even after remission and transplantation are achieved.
At the same time, the study also shows that ELN classification alone is not the full story. Patients within the same broad ELN category can still have materially different post-transplant outcomes depending on specific mutations. That observation provides the rationale for the investigators’ expanded molecular risk model.
Mutations Associated With Inferior Relapse-Free Survival
Four mutations emerged as being associated with shorter relapse-free survival on multivariable analysis: DNMT3A, SF3B1, TP53, and WT1. Each of these findings is biologically plausible, though they likely operate through different mechanisms.
TP53 is the most expected adverse marker. TP53-mutated AML is associated with genomic instability, poor response durability, and a particularly high risk of post-transplant relapse. Even when remission is achieved, resistant subclones may persist and eventually escape graft-versus-leukemia effects. The persistence of TP53-associated risk after transplant is consistent with prior literature showing that allo-HSCT improves outcomes in some patients but rarely fully overcomes the adverse biology.
WT1 mutations also correlated with shorter RFS. WT1 is relevant not only as a recurrently altered gene but also as a common MRD marker in AML. Mutated or overexpressed WT1 may reflect disease with enhanced leukemic stem cell fitness or subclinical residual disease persistence, either of which could contribute to relapse after transplant.
DNMT3A mutations were linked to inferior RFS as well. DNMT3A-altered AML often sits within clonal hematopoiesis-related architecture and may coexist with other high-risk lesions. One possible interpretation is that DNMT3A marks a disease substrate with greater clonal resilience and relapse potential even after cytoreduction and immunologic pressure from the graft.
SF3B1 is a spliceosome gene more commonly associated with myelodysplastic biology and secondary-type leukemogenesis. Its association with worse post-transplant RFS in this study is clinically coherent, especially if it identifies AML with antecedent myeloid neoplasm features, more complex clonal architecture, or less transplant-sensitive disease.
FLT3-ITD and the Unexpected Association With Prolonged RFS
One of the most intriguing findings was that FLT3-ITD mutations were associated with prolonged relapse-free survival. Historically, FLT3-ITD has been regarded as an adverse-risk feature because of its association with early relapse. Why might the transplant cohort show the opposite direction?
Several explanations are plausible. First, transplant may preferentially mitigate FLT3-ITD-associated relapse risk more effectively than it mitigates some other molecular subtypes, particularly when patients reach first remission rapidly. Second, contemporary management of FLT3-mutated AML increasingly includes FLT3 inhibitors before and after transplant, which may improve disease control. Third, this may reflect selection effects: only patients who attained remission and proceeded to transplant were analyzed, creating a biologically enriched subset that had already demonstrated treatment responsiveness. Without the full paper’s granular subgroup data, it is difficult to determine the relative contribution of each explanation, but the result is clinically important because it argues against treating all FLT3-ITD-positive transplanted patients as uniformly poor risk.
Development of a Prognostic Score
The investigators translated the multivariable findings into a practical prognostic score. The score incorporated age, AML type, transplant timing, and genetic or molecular alterations, weighted according to hazard ratio-derived effects. Although summary-level information does not provide the exact scoring algorithm here, the concept is clinically attractive: post-remission transplant counseling is more useful when it integrates disease biology with real-world transplant variables instead of relying on diagnosis-era genetics alone.
The authors also report validation of the score, which strengthens confidence that the model is not simply overfit to a single derivation dataset. If externally reproducible, such a tool could help identify patients with especially high relapse risk despite first-remission transplantation and guide intensified peri-transplant strategies.
Clinical Interpretation
The central message of this study is that molecular profiling remains highly informative even after the major therapeutic milestone of complete remission and even when allo-HSCT is planned. In other words, transplant is not a biologic reset. Some mutational programs continue to predict relapse despite the addition of graft-versus-leukemia effects.
This has several practical implications. First, pre-transplant counseling should move beyond broad ELN categories whenever next-generation sequencing data are available. A patient with adverse-risk AML by ELN and TP53 or WT1 mutation may warrant a different discussion than a patient whose risk is driven by a feature that appears more transplant-sensitive.
Second, transplant platform choices may eventually be refined by molecular risk. Although this study does not prove that any specific donor source, conditioning intensity, or graft manipulation strategy improves outcomes in any one mutational group, it identifies patient subsets most likely to benefit from such investigations. For example, patients with TP53-, WT1-, DNMT3A-, or SF3B1-mutated AML may be candidates for trials testing intensified conditioning, post-transplant maintenance, or preemptive MRD-triggered intervention.
Third, surveillance intensity after allo-HSCT may need to be mutation-informed. A patient carrying a mutation associated with shorter RFS may reasonably undergo closer molecular monitoring, earlier marrow assessment when blood counts change, and lower thresholds for therapeutic intervention. This is especially relevant as centers increasingly use MRD, donor chimerism, and mutation tracking together.
Fourth, the finding regarding FLT3-ITD supports a more nuanced view of post-transplant risk in the era of targeted therapy. It may also reinforce the importance of integrating FLT3 inhibitor strategies around transplantation rather than assuming that FLT3-ITD inevitably predicts poor outcomes despite transplant.
Biological Plausibility and Translational Relevance
The study fits well with current understanding of AML as a set of biologically distinct diseases rather than a single entity. Mutations such as TP53 likely mark intrinsic therapy resistance and genomic complexity. Spliceosome-related alterations such as SF3B1 may point toward secondary AML biology and persistent clonal fitness. DNMT3A may identify epigenetically dysregulated founding clones capable of surviving treatment bottlenecks. WT1 may reflect leukemia persistence that remains clinically occult but biologically active.
These mechanistic differences are not merely academic. They suggest that post-transplant relapse prevention may ultimately need to be mutation-specific. The field already has examples of this strategy in FLT3-mutated AML, where maintenance approaches using sorafenib or other FLT3 inhibitors have shown benefit in selected settings. Comparable strategies are still less developed for TP53-mutated or spliceosome-mutated disease, but this study underscores the need.
Strengths and Limitations
The study’s main strengths are its large sample size, focus on a clinically meaningful and relatively homogeneous disease state, and use of real-world transplant registry data. Restricting analysis to first complete remission is important because it avoids mixing outcomes from later transplants, where prognosis is strongly confounded by prior relapse biology. The validation of both ELN systems and the development of a composite prognostic score also add practical value.
Several limitations should temper interpretation. First, the retrospective design introduces potential selection bias and unmeasured confounding. Patients who reached first remission and proceeded to transplant are not representative of all newly diagnosed AML patients. Second, registry studies may contain heterogeneity in conditioning regimens, donor selection, supportive care, MRD assessment, and maintenance therapy, all of which can influence relapse risk. Third, the abstract does not provide effect sizes, confidence intervals, or exact model calibration statistics, which are necessary for judging the magnitude and precision of each association. Fourth, mutational analyses are only as robust as the completeness and standardization of sequencing across centers. Finally, external validation outside the PETHEMA network will be essential before widespread adoption of the score.
How This Study May Influence Practice
For clinicians, the most immediate takeaway is not that one should abandon current guideline-based transplant decision-making, but that mutation-level information should be incorporated more systematically into transplant planning and follow-up. In practical terms, this study supports several actions:
Obtain comprehensive molecular profiling at diagnosis and ensure those results remain visible during transplant discussions.
Recognize that high-risk lesions such as TP53 and WT1 continue to matter after transplant and should prompt consideration of clinical trials, enhanced surveillance, and relapse-prevention strategies.
Interpret FLT3-ITD in context rather than assuming uniformly poor transplant outcomes, particularly in patients who achieve first remission and may receive targeted therapy.
Use broader prognostic tools that combine clinical and molecular variables, especially when counseling patients about expected relapse risk after allo-HSCT.
As always, these steps should be integrated with MRD status, comorbidity assessment, donor availability, and patient preferences.
Conclusion
This PETHEMA registry study adds important evidence that molecular profiling meaningfully refines prognosis in AML patients undergoing allo-HSCT in first complete remission. While ELN2017 and ELN2022 remain valid, they do not fully capture post-transplant relapse risk. DNMT3A, SF3B1, TP53, and WT1 identified patients with shorter relapse-free survival, whereas FLT3-ITD was associated with longer relapse-free survival in this selected transplant population. By combining clinical and molecular variables into a validated prognostic score, the investigators provide a framework for more individualized transplant-era risk assessment.
The broader significance is clear: the future of AML transplantation lies not only in deciding who should undergo allo-HSCT, but also in determining how transplantation should be tailored according to disease biology before and after the procedure. Prospective validation and integration with MRD and maintenance strategies will be the next critical steps.
Funding and ClinicalTrials.gov
Funding information was not provided in the source material available here. No ClinicalTrials.gov registration number is applicable from the abstracted information for this retrospective registry study.
References
Colmenares R, Barragán E, Rodríguez-Veiga R, Torres-Miñana L, Sánchez-García J, Tormo M, Bernal T, Martínez-Sánchez P, Rodríguez-Arbolí E, Gil C, Soria-Saldise E, Serrano J, Colorado M, García-Fortes M, Bilbao C, López-Lorenzo JL, Larráyoz MJ, Pérez-Santaolalla E, Lavilla-Rubira E, Algarra L, García-Garay MC, Chillón C, Torres-Ochando M, Couto C, García-Boyero R, Almela Á, Noriega V, Callejas M, Barrios M, Casado S, Balerdi A, Cabello A, Labrador J, Mateos MC, Amigo ML, Pérez-Encinas M, García-Pérez MJ, Costilla L, Bergua J, Carreño-Tarragona G, Martínez-López J, Ayala R, Montesinos P, PETHEMA Group. Impact of molecular profiling in patients with acute myeloid leukemia undergoing allogeneic transplantation in first remission: a study by the PETHEMA group. Bone Marrow Transplantation. 2026-05-07. PMID: 42098381. Available at: https://pubmed.ncbi.nlm.nih.gov/42098381/
Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377.
Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447.
Schlenk RF, Kayser S, Bullinger L, et al. Differential impact of allelic ratio and insertion site in FLT3-ITD-positive AML with respect to allogeneic transplantation. Blood. 2014;124(23):3441-3449.
Xuan L, Wang Y, Huang F, et al. Sorafenib maintenance in patients with FLT3-ITD acute myeloid leukaemia undergoing allogeneic haematopoietic stem-cell transplantation: an open-label, multicentre, randomised phase 3 trial. Lancet Oncology. 2020;21(9):1201-1212.