Claudia Sargas

RAS mutations are recurrent in acute myeloid leukemia (AML); yet, their genomic and clinical landscape remain poorly characterized. We analyzed centralized molecular profiling data in 8,553 patients with newly diagnosed AML. Of these, 1,775 (21%) harbored RAS mutations: NRAS in 1,103 (13%), KRAS in 445 (5%), and both in 227 (3%). RAS mutations primarily clustered in P-loop hotspots (G12/G13, 70%), with NRAS showing higher Q61 mutation prevalence (24% vs. 9%) and KRAS showing additional K117/A146 hotspots. Multiple RAS mutations occurred in 23% of patients. Convergent evolution of RAS signaling mutations emerged as a key feature of RASmut AML. RTK/RAS pathway mutations were non-uniformly enriched: PTPN11, KIT and FLT3-TKD mutations were positively associated, CBL mutations showed no association, and FLT3-ITD mutations were negatively associated. The prognostic impact of RAS mutations was dependent on the genomic background and treatment. In multivariate analyses in intensively treated patients, NRAS mutations showed opposing associations with overall survival (OS): adverse in patients with AML, myelodysplasia-related (HR=1.32, p=0.036) and favorable in those with NPM1 (HR=0.61, p=0.036) or FLT3-ITD (HR=0.63, p=0.048) mutations, among others. KRAS mutations predicted inferior OS (HR=1.41; p=0.003). In patients treated with non-intensive, venetoclax-based regimens, RAS mutations were associated with markedly inferior OS in patients with NPM1 (HR=2.98, p=0.007) or IDH1/2 mutations (HR=4.06, p<0.001). These data inform risk stratification and development of RAS-targeting strategies in AML.

NPM1-mutated acute myeloid leukemia (AML) is genetically well-defined, but clinical outcomes remain heterogeneous, suggesting that quantitative clonal features may refine current risk stratification. We analyzed 688 intensively treated NPM1-mutated AML integrating variant allele frequency (VAF), mutation order, and clonal architecture inferred by PyClone and ClonEvol. Co-mutations were present in 97% of patients (median = 3 per case), dominated by DNMT3A (49%), FLT3-ITD (46%), TET2 (22%), and IDH2 (20%). Prognostic modelling of NPM1 VAF identified an optimal cut-off of 31.44%, defining NPM1high and NPM1low groups. NPM1low correlated with splicing-related alterations and independently predicted inferior overall (HR = 1.46; p = 0.037) and relapse-free survival (HR = 1.40; p = 0.036). Gene-specific VAF analyses revealed divergent effects across partners, high DNMT3A, FLT3-OTHER, KRAS, and PTPN11 burdens were adverse, whereas high IDH2 VAF was protective. Combined models showed that patients with NPM1high and favorable co-mutation VAFs had the best outcomes, while dual unfavorable burdens conferred the poorest survival. Mutation ordering inferred from VAFs positioned NPM1 after epigenetic and splicing lesions but before signaling and transcription-factor mutations. Non-canonical orders, such as early FLT3-OTHER/TKD or WT1 prior to NPM1, significantly stratified outcomes. Clonal reconstruction revealed predominantly linear evolutionary trajectories (84.3%), with increased mutational burden and clonal diversity associating with inferior survival. Notably, intra-clonal co-localization of NPM1 with IDH1 or TET2 was associated with improved outcomes, whereas co-localization with WT1 predicted dismal prognosis. These results demonstrate that quantitative and structural dimensions of clonality refine the biological and prognostic landscape of NPM1-mutated AML beyond mutational status alone.

Background/Objectives: The incidence of cardiac morbimortality in patients with relapsed/refractory (R/R) acute myeloid leukemia (AML) is unknown. Methods: We analyze the characteristics, incidence, risk factors, and outcomes of cardiac events in AML patients treated for second-line (2L) or third-line (3L) episodes. Results: Among 327 2L AML patients (median age 62 years old), 135 experienced cardiac events, with an incidence of 38.6% non-fatal and 1.3% fatal events at 6 months. The grade 1–2 incidence was 16.8%, and the grade 3–4 incidence was 23.5% at 6 months. Overall, 207 cardiac events occurred in the 2L cohort, the most frequent being hypertension (n = 45), bradycardia (n = 39), QTc prolongation (n = 35), heart failure (n = 33), syncope/presyncope (n = 22), arrhythmia (n = 18), and myocardial ischemia (n = 8). Median OS in the 2L cohort was 9.4 months, 21.4 months in patients with grade 1–2, 8.8 months in patients without a cardiac event, 7.6 months in grade 3–4 patients, and 2.1 months with in 5 patients (p = 0.0035). The multivariate analysis showed prior cardiologic antecedents (p = 0.013), intensive 2L chemotherapy (p = 0.01), and inclusion in a 2L clinical trial (p < 0.001) as independent risk factors for non-fatal cardiac events. Among 189 patients of the 3L cohort, the incidence of non-fatal and fatal cardiac events was 49.2% and 0% at 6 months, respectively. Non-fatal cardiac events were more frequent in patients with prior cardiac antecedents (p = 0.004). Conclusions: In summary, cardiotoxicity is a frequent and challenging complication in R/R AML patients. We identified the risk factors that could be relevant to implementing risk-adapted management guidelines, aiming to reduce morbi-mortality in this difficult-to-treat setting.

Background: In recent years, numerous recurrently mutated genes have been identified in acute myeloid leukemia (AML), some of which, such as FLT3 and IDH1/2, serve as therapeutic targets, offering new treatment options. Rapid mutational analysis is crucial for timely and optimal therapy selection. This study aims to develop and validate a rapid, cost-effective, and sensitive screening method for detecting IDH1, IDH2, and FLT3-TKD2 mutations using polymerase chain reaction (PCR) and high-resolution melting curve analysis (HRM). Methods: A PCR-HRM assay was developed to simultaneously detect mutations in IDH1, IDH2, and FLT3-TKD2. The method was applied to a cohort of 1363 AML patients, and its performance, including turnaround time, was evaluated through comparison with next-generation sequencing (NGS) results. Results: The PCR-HRM method demonstrated a positive percent agreement of 98%, 98%, and 92% for IDH1, IDH2, and FLT3-TKD2, respectively, and a negative percent agreement of 100% for all three genes compared to NGS. No false positives were observed, and false negatives were detected in less than 1% of cases, mostly in FLT3-TKD2, all occurring below the established limit of detection. The turnaround time and cost of PCR-HRM were significantly lower than those of NGS. Conclusions: This method offers a highly sensitive, specific, and time-efficient approach for the simultaneous detection of IDH1, IDH2, and FLT3-TKD2 mutations in AML patients. Its rapid turnaround time and cost-effectiveness make it a valuable tool for routine clinical screening, facilitating timely and targeted treatment decisions.

Background/Objectives: This PETHEMA PCR-LMA study aimed to evaluate whether mutations detected by NGS (VAF cut-off of ≥5%) correlate with NPM1, FLT3-ITD, FLT3-TKD, IDH1, and IDH2 mutations detected using conventional PCR (analytical sensitivity 3%) in a nationwide network of seven reference laboratories. Methods: Between 2019 and 2021, 1685 adult AML patients with at least one centralized sample (NGS or PCR) at primary diagnosis or relapse/refractory episode were included. Results: During this period, 1288 paired NGS/PCR samples (1094 at diagnosis, 103 at relapse and 88 at refractoriness) were analyzed. Considering PCR the gold-standard, for NPM1 NGS sensitivity was 98.5% and specificity 98.9%, for FLT3-ITD 73.8% and 99.6%, for FLT3-TKD 84.5% and 99.3%, for IDH1 98.7% and 98.7%, and for IDH2 99.1% and 97.7%, respectively. Overall concordance rate of positive results between NGS (and PCR was 95% (262/276) for NPM1, 72% (149/206) for FLT3-ITD, 74% (49/66) for FLT3-TKD, 87% (77/89) for IDH1 and 84% (107/127) for IDH2. Overall, median days from sample reception until report were 7 for PCR and 28 for NGS. Conclusions: This study shows high concordance between NPM1 and IDH results using PCR and NGS. However, sensible important discrepancies are observed for FLT3 mutations. In our context, rapid screening for these druggable mutations should be performed by conventional PCR.

SUMMARY: Myelodysplastic syndrome disease (MDS) has a variable risk for progression to AML. Mutations in CEBPA are associated with a high risk of disease progression, but whether this mutation is causative for AML development is unclear. To answer this question, we generated patient-derived, MDS-specific iPSCs recapitulating the patient disease phenotype upon differentiation to blood, with hematopoietic progenitor cells showing erythroid and myeloid-dysplasia. Introduction of a frameshift mutation affecting the C/EBPα bZIP domain led to disease progression, with a reduction in clonogenic potential, block in granulocyte development and increased self-renewal capacity of erythroid progenitors. ATAC-seq revealed that the acquisition of this mutation reshaped the chromatin landscape at distal cis-regulatory regions, promoting changes in clonal composition as observed by single cell RNAseq. Our work identifies mutant CEBPA as causative for MDS disease progression, providing a new isogenic MDS experimental model for drug screening to improve diagnostic and therapeutic strategies. Highlights: Development of isogenic iPSC model of clonal evolution of MDS Monoallelic disruption of CEBPA bZIP domain is causative for MDS disease progression Monoallelic disruption of CEBPA bZIP reshapes chromatin landscape Patient derived iPSCs recapitulate drug responsiveness

Background and Objective: The main objective in this paper is to validate a machine-learning model trained to predict the 90-day risk of complications for patients with Acute Myeloid Leukemia using variables available at diagnosis. This is a first fundamental step towards the development of a tool that could help physicians in their therapeutic decisions. Methods: 266 patients and 36 variables form the training dataset collected by Hospital 12 de Octubre (Madrid, Spain). The external test cohort provided by Instituto de Investigación Sanitaria La Fe (Valencia, Spain) contains 162 observations. An XGBoost model was trained with one dataset and validated with the other. Additionally, the features were ranked by permutation importance and compared with the ELN 2022 risk classification by genetics at initial diagnosis. Results: The model was evaluated with the training cohort using leave-one-out cross-validation, reaching a ROC-AUC of 0.85. By setting the functioning point that maximises Youden’s index, 3 out of 4 patients with complications and 84 out of 100 in remission are correctly classified. The model was validated with external data collected in a different hospital, achieving 0.7 ROC-AUC. At the best functioning point, almost 6 out of 10 patients with complications and 8 out of 10 patients in remission are correctly classified. Ranking the variables by descending importance, the top four are, in order: age, white-blood-cells count, Gender, and TP53. The list exhibits good coherence with the ELN 2022 risk classification. Conclusions: The model achieves performances that suggest it could be used as a therapeutical decision support tool. Important variables are coherent with ELN 2022 risk classification. Further work is needed to understand the reasons for the drop in test performance. The 90-day model should be supplemented by others that predict the risk of complications at six months or one year.

Next-Generation Sequencing is needed for the accurate genetic risk stratification of acute myeloid leukemia according to European LeukemiaNet (ELN) guidelines. We validated and compared the 2022 ELN risk classification in a real-life cohort of 546 intensively and 379 non-intensively treated patients. Among fit patients, those aged ≥65 years old showed worse OS than younger regardless risk classification. Compared with the 2017 classification, 14.5% of fit patients changed the risk with the 2022 classification, increasing the high-risk group from 44.3% to 51.8%. 3.7% and 0.9% FLT3-ITD mutated patients were removed from the favorable and adverse 2017 categories respectively to 2022 intermediate risk group. We suggest that midostaurin therapy could be a predictor for 3 years OS (85.2% with vs. 54.8% without midostaurin, P = 0.04). Forty-seven (8.6%) patients from the 2017 intermediate group were assigned to the 2022 adverse-risk group as they harbored myelodysplasia (MDS)-related mutations. Patients with one MDS-related mutation did not reach median OS, while patients with ≥2 mutations had 13.6 months median OS (P = 0.002). Patients with TP53 ± complex karyotype or inv(3) had a dismal prognosis (7.1 months median OS). We validate the prognostic utility of the 2022 ELN classification in a real-life setting providing supportive evidences to improve risk stratification guidelines.