Konstanze Döhner

The 2022 ELN recommendations reclassified fms‐like tyrosine kinase‐3 ( FLT3 )‐ITD mutated acute myeloid leukemia (AML) as intermediate‐risk, irrespective of allelic ratio (AR), and designated capillary electrophoresis (CE) as the preferred detection method. However, CE requires substantial expertise, particularly to identify FLT3 ‐ITD microclones ( FLT3 ‐ITD ^m ; AR < 0.05), whose clinical significance remains uncertain. Herein, we investigated diagnostic aspects and the prognostic impact of FLT3 ‐ITD ^m . An initial methodological study performed on 48 FLT3 ‐ITD+ samples, with AR 0.001–6.00, showed overall concordance between CE and next‐generation sequencing (NGS) results ( p < 0.001), with only three microclones (AR: 0.002, 0.001, and 0.008) identified by CE and not by NGS. We next analyzed 658 AML patients diagnosed with a FLT3 ‐ITD+ AML ( n = 446 with AR ≥ 0.05, and n = 212 with a microclone) between 2017 and 2022. Patients with FLT3 ‐ITD ^m , as compared to patients with AR ≥ 0.05, were older, more often had secondary‐type AML, and displayed a higher prevalence of myelodysplasia‐related gene mutations, while NPM1 ‐mutation frequency was similar in the two groups. Complete remission rates were lower in FLT3 ‐ITD ^m (74%) than in patients with AR ≥ 0.05 (88%) treated with intensive chemotherapy without FLT3‐inhibitors (FLT3i). In the intensively treated subgroup, multivariate analyses demonstrated that age, NPM1 mutation, and allogeneic stem cell transplantation, rather than FLT3 ‐ITD AR, were independent predictors of disease‐free survival. At relapse, 15 of 33 (45%) evaluable patients remained FLT3 ‐ITD+, with 10 expanding the original microclone. In conclusion, FLT3 ‐ITD ^m are associated with older age, secondary AML, and with expansion of the FLT3 ‐ITD clone upon relapse, thereby providing a rationale for the evaluation of FLT3i in randomized clinical trials.

Ruxolitinib (RUX) is approved as second-line treatment for HU-intolerant/-resistant patients with polycythemia vera (PV). Here, we assessed RUX vs. best available therapy (BAT) as first-line treatment in PV in the randomized multicenter RuxoBEAT clinical trial (NCT02577926). Crossover from BAT to RUX was possible after 6 months. Baseline characteristics in the RUX ( n = 44) and BAT ( n = 34) arms were comparable. The primary endpoint, the clinico-hematologic complete response (CR) rate at month 6, was not met (2.3% vs. 2.9%, p = 1.000). Analyses of secondary and exploratory endpoints showed that RUX was significantly more effective than BAT in improving hematocrit, splenomegaly, and patient-reported pruritus and fatigue, but less effective in reducing platelet counts. Compared to baseline, JAK2V617F variant allele frequency decreased significantly in both RUX and BAT groups at 6 months. The time on treatment ( p < 0.001) and duration of treatment ( p = 0.022) were significantly superior for RUX vs. BAT during the 24-month study period. There were no new safety signals. These findings demonstrate that RUX did not lead to higher CR rates than BAT and suggest that RUX may induce clinically meaningful additional disease improvements in first-line therapy of PV patients.

Allogeneic hematopoietic stem cell transplantation (allo‐HSCT) is a curative treatment option for a significant proportion of patients with acute myeloid leukemia (AML), and it is generally recommended when the relapse risk without allo‐HSCT outweighs the estimated non‐relapse mortality significantly. While current recommendations for allo‐HSCT are based on risk groups, there is considerable heterogeneity within these individual categories. We analyzed 2550 intensively treated AML patients aged 18–70 with cytogenetic and next‐generation sequencing data from the HARMONY Alliance database, who did not receive allo‐HSCT in first complete remission (CR1). A non‐parametric machine learning (ML) model based on Bayesian Additive Regression Trees (BART) integrated clinical variables and genomic aberrations to provide individualized outcome estimations. External validation was performed in a cohort of 714 patients enrolled in UK‐NCRI trials. The predictive performance of the HARMONY ML model, measured by the area under the time‐dependent receiver operating curve (AUC(t)), was superior to European LeukemiaNet (ELN)2022 risk classification in estimating 5‐year overall survival (0.741 vs. 0.700), 5‐year relapse‐free survival (0.752 vs. 0.705), and 5‐year cumulative incidence of relapse (0.742 vs. 0.708), which was confirmed in the external validation cohort. Notably, the model revealed substantial heterogeneity within ELN2022 risk groups, identifying a significant proportion of favorable‐risk patients with a predicted 5‐year CIR > 40%, who could potentially benefit from allo‐HSCT in CR1. The HARMONY ML model provides individualized risk prediction in intensively treated adult AML patients and supports more tailored therapeutic decisions regarding allo‐HSCT in CR1, which should be further advanced by integrating measurable residual disease in the future.

Acute myeloid leukemia (AML) is characterized by recurrent chromosomal abnormalities that form the basis of the European LeukemiaNet (ELN) risk classification and serve as essential determinants of prognosis and therapeutic decision-making. Conventional metaphase karyotyping remains the diagnostic gold standard for detecting these abnormalities; however, its utility is limited by longer turnaround times, often delaying critical clinical management. Here, we present a long-read sequencing-based (LRS) low-coverage whole genome sequencing (lcWGS) approach using Oxford Nanopore Technology as a rapid and scalable alternative for cytogenetic profiling. A total of 100 diagnostic AML samples were analyzed, comprising 50 retrospectively selected cases with known adverse-risk cytogenetics and 50 prospectively enrolled patients with clinically defined de novo AML. LcWGS demonstrated robust analytical performance, identifying chromosomal aberrations with 93% sensitivity, specificity, and overall accuracy, respectively. Complex karyotypes were reliably detected, with an area under the curve (AUC) of 0.971. Reproducibility was validated through replicate sequencing at two independent laboratories (R=0.99). LcWGS-derived estimates of clone size showed moderate correlation with conventional cytogenetic assessments (R=0.54). Patients with complex karyotypes identified by lcWGS exhibited significantly shorter overall and relapse-free survival, closely mirroring outcomes defined by conventional karyotyping and underscoring the value of lcWGS for risk stratification. Median turnaround time from sample receipt to bioinformatics interpretation was approximately 34 hours, enabling delivery of actionable karyotype results within 72 hours. These findings establish lcWGS as a rapid, reproducible, and accurate platform for detecting clinically relevant chromosomal abnormalities, addressing a critical need for timely risk stratification and treatment initiation in AML.

The World Health Organization fifth edition and International Consensus Classification for myeloid neoplasms both incorporate empirical numerical thresholds to morphologic and molecular features defining certain disease entities. However, the clinical implications of these thresholds remain unclear. We analyzed a large cohort (N = 6,976) of patients with myeloid neoplasms to evaluate the impact of proposed yet different numerical thresholds for variant allele frequency of genetic mutations or hematologic parameters set forth by the World Health Organization fifth edition and International Consensus Classification for classification of SF3B1-mutated myelodysplastic neoplasms, NPM1-mutated acute myeloid leukemia (AML), and oligomonocytic chronic myelomonocytic leukemia. Our analysis demonstrated that the clonal burden of SF3B1 mutation in myelodysplastic neoplasms informs classification and prognosis. Our findings support the notion that NPM1 mutation should be AML-defining regardless of blast percentage and highlight the adverse prognostic impact of the cumulative number of myelodysplasia-related mutations in NPM1-mutated AML. Finally, we provide evidence that integrating specific molecular signatures could improve the accuracy of oligomonocytic chronic myelomonocytic leukemia classification.Significance:Using comprehensive clinical and molecular profiling, this study provides a data-driven approach for evaluating numerical thresholds of variant allele frequency or hematologic parameters (i.e., blast percentage and absolute monocyte count) included in current classification schemas across a spectrum of myeloid malignancies, enabling refinement of disease classification and prognostication.

Comparison of clinical and molecular demographics and survival in patients with NPM1-mutated myeloid neoplasms with <20% blasts from the UK NCRI AML17/19 trials. (A) Density plots demonstrating no significant difference in baseline age or hematologic parameters between patients with <10% vs. 10% to 19% blasts. (B) Similarly, no significant difference was observed with respect to frequently identified mutations between patients with mutated NPM1 and <10% vs. 10% to 19% blasts. (C and D) Event-free and overall-survival was similar between patients with mutated NPM1 and <10% vs. 10% to 19% blasts. Reported p-values utilized the cox model Wald test.

Biplot of quantitative variables included in factor analysis of mixed data (FAMD) of myelodysplastic CMML cases (MD-CMML) including platelets, bone marrow blast (BM blasts) percentage, absolute neutrophil count (ANC), absolute monocyte count (AMC), relative monocytes, hemoglobin, and patient age at diagnosis. Arrow length and direction with respect to the circle indicates relative contribution to each quadrant. Closer proximity of an arrowhead to the circle indicates a stronger contribution of a variable to the quadrant.

WHO 2016 classification of included patients with NPM1-mutated myeloid neoplasms (MN)

Density plot depicting Bradley-Terry modeling of the estimated order of mutation acquisition in AML samples from the UKNCRI and AMLSG cohorts. Mutations in genes classified as MR mutations (SRSF2, ASXL1, RUNX1, U2AF1, ZRSR2, EZH2, STAG2, BCOR, SF3B1) tended to occur as earlier events relative to mutations in NPM1. VAF: Variant allele frequency