Abstract Background SF3B1 spliceosome mutations are among the most common genetic lesions in myelodysplastic neoplasms (MDS). The canonical hotspot K700E variant is generally associated with a favorable prognosis, whereas the impact of other recurrent variants is poorly defined. We investigated the clinical, transcriptomic, and functional consequences of distinct SF3B1 mutations in MDS patients and isogenic cell models. Methods Clinical data from 121 SF3B1 -mutated MDS patients were analyzed. RNA sequencing was performed on CD34⁺ patient bone marrow cells and CRISPR-engineered NALM-6 lines harboring individual SF3B1 variants. Comprehensive profiling of splicing, gene expression, and mitochondrial bioenergetic parameters was performed. Results Compared with K700E, the K666 variant was associated with shorter progression-free survival, distinct splicing abnormalities, and a higher frequency of retained introns, showing partial overlap with the H662 variant. Mitochondria-related genes were frequently mis-spliced and differentially expressed in K666-variant cells. In isogenic models, reduced complex IV activity and marked OXPHOS impairment were observed, with the most pronounced effects seen in K666N. Overall, SF3B1 variants confer heterogeneous biological and clinical effects. The K666N variant, in particular, is associated with adverse prognosis and profound bioenergetic dysfunction. Conclusions These results support a refined view of SF3B1 -mutated MDS as a biologically heterogeneous entity and suggest that variant-specific mitochondrial vulnerabilities may represent exploitable targets for precision-based therapeutic strategies.

Neocentromeres are newly formed chromosomal regions that can replace the function of traditional centromeres and are well documented in human clinical studies. However, their occurrence in neoplasia, including acute leukemia, appears to be rare. We analyzed complex karyotypes in bone marrow cells from 113 patients with acute myeloid leukemia and one patient with acute lymphoblastic leukemia using centromeric/multicentromeric fluorescence in situ hybridization and identified four cases (3.5%) with derivative chromosomes exhibiting newly formed constrictions. Three of these patients had secondary leukemia following preexisting hematological disorders, suggesting a potential role for neocentromeres in disease progression. In all four cases, neocentromeres were detected on derivative chromosome 11. To our knowledge, this is the first report of neocentromeres derived from this chromosome in acute leukemia. All four patients in our study died; however, all exhibited complex karyotypes, which are independently associated with poor prognosis and an aggressive disease course. Neocentromeres are a rare but potentially important source of genomic instability in malignant diseases. Generally, the formation of a new constriction allows mitotic rescue of acentric chromosomes, preventing their loss. An increase in genomic alterations in tumor cells predicts a more aggressive disease course and adverse outcomes. Due to limited data, the prognostic significance of neocentromeres remains unclear. Further rigorous investigation is required to deepen our understanding of the mechanisms underlying neocentromere formation and their implications in cancer.

Abstract Background Myelodysplastic neoplasms (MDS) are heterogeneous hematopoietic disorders characterized by ineffective hematopoiesis and genome instability. Mobilization of transposable elements (TEs) is an important source of genome instability leading to oncogenesis, whereas small PIWI-interacting RNAs (piRNAs) act as cellular suppressors of TEs. However, the roles of TEs and piRNAs in MDS remain unclear. Methods In this study, we examined TE and piRNA expression through parallel RNA and small RNA sequencing of CD34+ hematopoietic stem cells from MDS patients. Results Comparative analysis of TE and piRNA expression between MDS and control samples revealed several significantly dysregulated molecules. However, significant differences were observed between lower-risk MDS (LR-MDS) and higher-risk MDS (HR-MDS) samples. In HR-MDS, we found an inverse correlation between decreased TE levels and increased piRNA expression and these TE and piRNA levels were significantly associated with patient outcomes. Importantly, the upregulation of PIWIL2, which encodes a key factor in the piRNA pathway, independently predicted poor prognosis in MDS patients, underscoring its potential as a valuable disease marker. Furthermore, pathway analysis of RNA sequencing data revealed that dysregulation of the TE‒piRNA axis is linked to the suppression of processes related to energy metabolism, the cell cycle, and the immune response, suggesting that these disruptions significantly affect cellular activity. Conclusions Our findings demonstrate the parallel dysregulation of TEs and piRNAs in HR-MDS patients, highlighting their potential role in MDS progression and indicating that the PIWIL2 level is a promising molecular marker for prognosis. Graphical Abstract

Introduction: Limited treatment options are available for patients with red blood cell (RBC) transfusion-dependent (TD) lower-risk myelodysplastic syndromes (LR-MDS). Imetelstat (IME), a first-in-class, direct, and competitive inhibitor of telomerase activity, was approved by the United States Food and Drug Administration in June 2024 for the treatment of RBC-TD LR-MDS in patients who were relapsed or refractory to or ineligible for erythropoiesis-stimulating agents (ESA) based on the results of the pivotal IMerge trial (NCT02598661). IMerge demonstrated significant and durable efficacy of IME (n=118) versus placebo (n=60) for ≥8-week, ≥24-week, and ≥1-year RBC-transfusion independence (TI), with a generally manageable safety profile in this patient population (Platzbecker U, Santini V, et al. Lancet. 2024). This analysis pooled data from the 3 parts of the IMerge trial (phase 2, phase 3 and QTc study) to investigate the effect of prior therapies on the clinical activity of IME. Methods: IMerge was a phase 2/3 trial with a clinical QTc study that enrolled patients with heavily RBC-TD LR-MDS who were ESA-ineligible or relapsed/ refractory. IME was administered intravenously every 4 weeks at 7.1 mg/kg (equivalent to 7.5 mg/kg IME sodium). Prior lenalidomide (LEN) and prior hypomethylating agent (HMA) use were exclusion criteria in phase 3 only. In this analysis, IME-treated patients (N=226) were pooled from phase 2, phase 3, and the QTc study of IMerge and analyzed on the basis of prior treatment as follows: ± ESA, luspatercept (LUSP), LEN, and HMA. Prior treatment was not exclusive; patients may have received >1 prior therapy. Outcomes included ≥8-week and ≥24-week RBC-TI, rates of hematologic improvement-erythroid (HI-E) based on the revised International Working Group (IWG) 2018 criteria, transfusion reduction of ≥4 U/8 weeks, and a hemoglobin (Hb) rise of ≥1.5 g/dL for ≥8 weeks. Results: As previously presented, data from the IMerge phase 3 pivotal analysis demonstrated that patients treated with IME (n=118) had ≥8-week and ≥24-week RBC-TI rates of 40% and 28%, respectively, 42% met HI-E (IWG 2018) criteria, 60% had transfusion reductions of ≥4 U/8 weeks, and 34% had a Hb rise of ≥1.5 g/dL for ≥8 weeks (Platzbecker U, Santini V, et al. Lancet. 2024). In the current analysis of all IME-treated patients pooled in IMerge (N=226), 204 had prior treatment with an ESA and 22 were ineligible for ESAs; 35 had prior LUSP, 26 had prior LEN, and 22 had prior HMA treatment. Of IME-treated patients with prior ESA therapy, 40% and 28% achieved ≥8-week and ≥24-week RBC-TI, respectively, 43% met HI-E, 64% had a transfusion reduction of ≥4 U/8 weeks, and 33% had a Hb rise of ≥1.5 g/dL for ≥8 weeks. In IME-treated patients ineligible for ESA therapy, 36% and 14% achieved ≥8-week and ≥24-week RBC-TI, respectively, 41% met HI-E, 64% had a transfusion reduction of ≥4 U/8 weeks, and 2% had a Hb rise of ≥1.5 g/dL for ≥8 weeks. Of IME-treated patients who had prior treatment with LUSP, 29% and 20% achieved ≥8-week and ≥24-week RBC-TI, respectively, 26% met HI-E, 69% had a transfusion reduction of ≥4 U/8 weeks, and 29% had a Hb rise ≥1.5 g/dL for ≥8 weeks. Of IME-treated patients with prior LEN treatment, 23% and 12% achieved ≥8-week and ≥24-week RBC-TI, respectively, 31% met HI-E, 54% had a transfusion reduction of ≥4 U/8 weeks, and 19% had a Hb rise of ≥1.5 g/dL for ≥8 weeks. Of the 22 IME-treated patients who had prior treatment with HMA, 14% and 9% achieved ≥8-week and ≥24-week RBC-TI, respectively, 18% met HI-E, 50% had a transfusion reduction of ≥4 U/8 weeks, and 14% had a Hb rise of ≥1.5 g/dL for ≥8 weeks. Conclusions: Patients who were ESA-ineligible or who had prior treatment with LUSP, LEN, or HMA in IMerge experienced clinical benefit from IME treatment, though the number of patients was small. Given the evolving therapeutic landscape for LR-MDS and the limited data available on outcomes in later lines of treatment, these results have important clinical implications, suggesting that IME demonstrates clinical activity regardless of prior therapies.

ABSTRACT:Higher-risk myelodysplastic syndrome (HR-MDS) with RARA gene overexpression is a subset of patients (pts) with an actionable target for tamibarotene, an oral and a selective retinoic acid receptor-α (RAR-α) agonist. Tamibarotene with azacitidine (AZA) showed complete remission (CR) rates in myeloid leukemia. SELECT-MDS-1 was a phase 3 study comparing the activity of tamibarotene + AZA to placebo + AZA in these pts with newly diagnosed HR-MDS with RARA overexpression. Eligible pts had confirmed RARA overexpression, untreated MDS with higher-risk features by revised International Prognostic Scoring System (IPSS-R), and marrow blast count >5%. Pts were randomized 2:1 to receive tamibarotene + AZA or placebo + AZA, respectively. A total of 246 participants were randomized with 164 and 82 in the tamibarotene + AZA and placebo + AZA groups, respectively. Baseline characteristics included: 69.9% male; median age 75 years (range, 38-93); primary MDS, 89.8%; MDS-excess blasts-1, 48% and MDS-excess blasts-2, 52%; and IPSS-R risk category intermediate (25.5%), high (35.7%), and very high (38.9%). The study did not meet the primary end point of CR, with a P value of .2084 for the treatment effect in the tamibarotene + AZA group. The CR rates were 23.81% and 18.75% in the tamibarotene + AZA and placebo + AZA groups, respectively. The use of tamibarotene-based therapy to target RAR-α as a novel approach in pts with HR-MDS with RARA gene overexpression is not a paradigm, which can augment response rates beyond AZA monotherapy. Further explorations of alternative approaches, including those with a biomarker, to alter the natural history of this disease are warranted. This trial was registered at www.clinicaltrials.gov as #NCT04797780.

BACKGROUND:Venetoclax (VEN) combined with azacitidine (AZA) has become the standard treatment for patients (pts) with newly diagnosed (ND) acute myeloid leukemia (AML) who are unfit for intensive chemotherapy. This study reports on a real-world cohort of 163 patients with ND AML treated with VEN/AZA across 7 hospitals in the Czech Republic. PATIENTS AND METHODS:The median age at diagnoses was 72 (39-98) years. Standard VEN/AZA (7 + 28; ie, AZA 75mg/m2 for 7 days + VEN for 28 days) was used in 89% of patients (145/163), with 73% (119/163) receiving a reduced dose of venetoclax (100 mg or 200 mg) due to concomitant azole antifungal prophylaxis. Treatment was considered palliative or nonintensive (including potential allogeneic transplant candidates) based on the clinician's decision. RESULTS:Among 150 evaluable patients, 71% (107/150) achieved composite complete remission (CCR = CR + CRi), with 28% (32/115) achieving MRD negativity. Median overall survival (OS) was 5 months for the palliative cohort and 21 months for the nonintensive cohort. Day 30 and day 60 mortality rate were 7% and 13%, respectively. Allogeneic transplantation (HSCT) was performed in 15% (25/163) of pts, with the median OS not reached in this group. Mutations in NPM1, and IDH1/2 were associated with improved survival, while FLT3, ASXL1, TP53, deletion/monosomy 7, and complex karyotype were linked to poorer outcomes. CONCLUSION:Data suggest that TP53 patients benefit from treatment if they achieve CR within the first 2 cycles. Discontinuing VEN/AZA in low-risk MRD-negative patients does not appear to worsen overall survival, though long-term outcomes remain to be seen.

PURPOSE:Bone marrow failure (BMF) in idiopathic aplastic anemia (AA) and hypoplastic myelodysplastic neoplasms (MDS-h) results from the destruction of hematopoietic progenitors by autoreactive T cells; however, the molecular events driving the pathogenesis of these disorders remain unclear. We therefore applied whole-exome sequencing (WES) in AA and MDS-h patients to identify acquired and inherited gene variants presumed to have functional consequences for BMF. We also used transcriptome profiling to investigate the molecular mechanisms underlying the aberrant T cell response. METHODS:WES was performed on DNA from 42 patients at diagnosis. Transcriptome profiling of CD3⁺ cells was conducted in 21 patients and 10 healthy donors. Peripheral blood cell populations were analyzed by flow cytometry. RESULTS:Pathogenic/likely pathogenic (P/LP) somatic gene variants were detected in 79% of patients and were functionally associated with BMF-relevant processes such as antigen processing/presentation, T cell-mediated immunity, and DNA repair. P/LP germline gene variants were found in all patients, almost half of whom harbored variants associated with inborn errors of immunity. Patient T cells displayed expression signatures of increased inflammation, apoptosis, hypoxia response, and decreased oxidative phosphorylation. Dysregulated long noncoding RNAs were predicted to primarily regulate the differentiation of T helper 17 cells. Patients also showed significantly lower frequencies of immature progenitors and natural killer cells compared with controls. CONCLUSION:Patients with idiopathic AA and MDS-h carried multiple germline immune-related gene variants that may increase susceptibility to immune-mediated BMF. Furthermore, patient T cells exhibited altered energy metabolism, which may represent a therapeutic target for modulating immune responses in autoimmune diseases.

Introduction: This analysis evaluates the impact of early administration of the growth factor G-CSF during the first therapy cycle in 117 patients with acute myeloid leukemia (AML) treated with venetoclax-based combination therapy. Not only were the clinical characteristics of patients across three groups (de novo AML, N=28; secondary AML [sAML], N=39; salvage treatment [previously treated with HMA for HR-MDS], N=50) taken into account, but also the influence of genetic mutations on hematological parameters, treatment response, and survival. Methods: We retrospectively analyzed AML patients receiving venetoclax-based therapy, stratified by treatment cohort. Associations between G-CSF use, infection rates, and clinical outcomes were assessed using chi-squared tests and Kaplan-Meier survival analysis. Multivariate Cox regression evaluated the impact of G-CSF, infections, and neutrophil recovery on OS. Separately, the influence of recurrent gene mutations on hematologic parameters, complete remission (CR), and OS was analyzed using linear regression, penalized logistic regression, and multivariate Cox models. Statistical significance was set at p < 0.05. Analyses were conducted in R (v4.4.3). Results: Use of G-CSF showed a statistically significant difference between the subgroups (χ² = 11.89; p = 0.0026). G-CSF was most frequently administered in patients in the sAML (89.7%) and salvage (78.0%) groups, while in de novo AML it was administered in only 53.6% of cases. The incidence of severe grade 4 neutropenia also differed significantly between groups (χ² = 8.56; p = 0.0138). The highest relative incidence was observed in patients with de novo AML (28.6%) and sAML (25.6%), while in the salvage group, Gr 4 neutropenia was present in only 6.0% of patients. The incidence of infections was higher in de novo and sAML patients and lower in the salvage group (p = 0.0099). The presence of infection during the first therapy cycle was associated with worse overall survival, although this did not reach statistical significance (p = 0.04). We did not observe statistically significant differences in response among subgroups (p = 0.2615), although the de novo group achieved the highest CR rate (64 %). Survival according to Kaplan-Meier analysis suggests a poorer prognosis for the salvage group. However, G-CSF administration was not associated with a decrease in overall survival (OS). Analysis of the effect of genetic mutations found in 97 patients showed that the presence of certain aberrations can significantly affect both the chance of achieving complete remission (CR) and selected hematological parameters. The most significant positive effect on achieving CR was observed for the IDH2 mutation, which was associated with a statistically significant higher probability of treatment response (OR 6.32; p = 0.0285). The NPM1 mutation also increased the chance of achieving CR (OR 4.62; p = 0.0500) and was associated with a trend toward improved overall survival (HR ~0.65). In contrast, EZH2 mutation indicated a negative impact on treatment response (OR 0.00), and although this effect was not statistically significant for survival. The mutations were further analyzed for their relationship to laboratory parameters such as baseline hemoglobin level (Hb), platelets (Plt), absolute neutrophil count (ANC), and percentage of blasts in peripheral blood. Mutations in EZH2 was associated with a significant decrease in hemoglobin levels (−30.13 g/L; p < 0.001), ASXL1 led to a significant reduction in ANC. Conclusions Based on the available data, we conclude that early administration of G-CSF during first-line treatment of AML is safe even in genetically at-risk patients. These results point to different patterns of toxicity and supportive care between clinical subgroups of AML, with the use of G-CSF not fully corresponding to the incidence of baseline severe neutropenia. The use of G-CSF may promote hematological regeneration in selected patients without fear of accelerating the progression of AML. Administration has no negative effect on survival. IDH2 and NPM1 mutations appear to be favorable in terms of treatment response, while EZH2 and ASXL1 mutations are associated with adverse changes in blood parameters. These results underscore the importance of the patient's genetic profile in predicting clinical course and response to therapy.

Introduction: IME is a first-in-class, direct, and competitive telomerase inhibitor approved for the treatment (tx) of certain adult pts with LR-MDS with red blood cell (RBC) transfusion-dependent anemia who are relapsed or refractory to/ineligible for erythropoiesis-stimulating agents. In the pivotal Phase 3 IMerge trial (NCT02598661), a significantly higher proportion of pts on IME vs placebo achieved ≥8-wk (40% vs 15%; P<.001) and ≥24-wk (28% vs 3%; P<.001) RBC transfusion independence (TI). The most common grade 3/4 tx-emergent adverse events with IME were neutropenia (68%) and thrombocytopenia (62%), typically occurring within the first 3 tx cycles and generally manageable and reversible. This post hoc analysis explored the possible association between IME-related cytopenias and Hb increase, a measure linked to RBC-TI achievement. Methods: Pooled pts from the 3 parts of IMerge (Phase 2/3 and QTc substudy) who received 7.1 mg/kg IME active dose (equivalent to 7.5 mg/kg IME sodium) were included. The relationship between percent reduction in platelet (PLT)/neutrophil (NEUT) counts occurring within the first 2 cycles of IME tx and subsequent outcomes was analyzed, including maximum (max) Hb increase from baseline (pre-tx Hb), hematologic improvement-erythroid (HI-E) per IWG 2006 criteria (defined as Hb rise ≥1.5 g/dL lasting 8 wk), and rates of ≥8-wk and ≥24-wk RBC-TI. Cytopenia levels were dichotomized at ≥50% reduction for PLTs and ≥75% reduction for NEUTs to assess categorical correlations with HI-E and RBC-TI responses (yes/no) using Fisher exact test and t test for group differences in Hb increase. Univariate linear regression was conducted to evaluate the association between continuous percent reduction in PLT/NEUT counts and max Hb increase. Multivariate analyses were conducted via stepwise linear regression (for max Hb increase) or logistic regression (for HI-E/TI response) with dichotomized percent reduction in PLT/NEUT counts in the presence of other baseline prognostic factors. An unadjusted nominal P value from each test was reported. Results: Data cutoff dates were 10/13/2023 (Phase 2/3) and 10/13/2024 (QTc substudy); 226 pts were included. Median values at baseline were 71 y of age, pre-tx Hb 7.8 g/dL, and PLT and NEUT counts 236×109/L and 2.6×109/L, respectively. Median IME tx duration was 34 wk. Pts with ≥50% max PLT reduction within the first 2 cycles of IME tx had a significantly greater mean max Hb increase from pre-tx (2.07 g/dL [n=169] vs 1.17 g/dL [n=39]; P=.003), HI-E rate (34.2% vs 11.9%; P=.005) and ≥24-wk RBC-TI rates (32.6% vs 7.1%; P<.001), and numerically greater ≥8-wk RBC-TI rates (42.4% vs 26.2%; P=.056) vs pts with <50% PLT reduction. As a continuous variable, PLT reduction remained significantly associated with Hb increase (P=.0095; linear regression). Similarly, pts with ≥75% max NEUT reduction within the first 2 cycles of IME tx had a significantly greater mean max Hb increase from pre-tx (2.53 g/dL [n=70] vs 1.58 g/dL [n=138]; P=.011), numerically greater HI-E rate (34.6% vs 27.7%), and comparable ≥8-wk (41.0% vs 38.5%) and ≥24-wk RBC-TI rates (30.8% vs 26.4%) vs pts with <75% NEUT reduction. As a continuous variable, NEUT reduction was also significantly associated with max Hb increase (P=.012). Multivariate analyses showed that the likelihood of achieving ≥8-wk and ≥24-wk RBC-TI was significantly correlated with the max Hb increase from pre-tx, g/dL (odds ratio [OR; 95% CI], 6.47 [3.73-11.24; P<.001] and 3.87 [2.66-5.64; P<.001], respectively). Multivariate analyses further confirmed that max Hb increase, g/dL, significantly correlated with ≥75% NEUT reduction (coefficient estimate [95% CI], 0.94 [0.32-1.57; P=.003]), and the likelihood of achieving HI-E significantly correlated with ≥50% PLT reduction (OR [95% CI], 4.16 [1.54-11.23; P=.005]). Conclusions: In this post hoc analysis, pts with ≥75% NEUT or ≥50% PLT reductions in the first 2 cycles of IME tx were more likely to have greater Hb increases from pre-tx or to achieve an HI-E response. The greater Hb increase from pre-tx emerged as a main driver for achieving ≥8-wk and ≥24-wk RBC-TI responses. Collectively, these data suggest that tx-emergent cytopenias with IME may be associated with potential for clinical benefit, similar to lenalidomide in del5q MDS. Further research is needed to confirm this on-target effect of IME resulting from its activity on clonal progenitor cells and subsequent recovery in blood cell production.

IntroductionProgressing myelodysplastic syndrome (MDS) into acute myeloid leukemia (AML) is an indication for hypomethylating therapy (HMA, 5-Azacytidine (AZA)) and a BCL2 inhibitor (Venetoclax, VEN) for intensive chemotherapy ineligible patients. Mouse models that engraft primary AML samples may further advance VEN + AZA resistance research.MethodsWe generated a set of transplantable murine PDX models from MDS/AML patients who developed resistance to VEN + AZA and compared the differences in hematopoiesis of the PDX models with primary bone marrow samples at the genetic level. PDX were created in NSGS mice via intraosseal injection of luciferase-encoding Lentivirus-infected MDS/AML primary cells from patient bone marrow. We validated the resistance of PDX-leukemia to VEN and AZA and further tested candidate agents that inhibit the growth of VEN/AZA-resistant AML.Results and discussionTransplantable PDX models for MDS/AML arise with 31 % frequency. The lower frequency of transplantable PDX models is not related to peritransplant lethality of the graft, but rather to the loss of the ability of short-term proliferation of leukemic progenitors after 10 weeks of engraftment. There exist subtle genetic and cytological changes between primary and PDX-AML samples however, the PDX models retain therapy resistance observed in patients. Based on in vitro testing and in vivo validation in PDX models, Panobinostat and Dinaciclib are very promising candidate agents that overcome dual VEN + AZA resistance.

Introduction: IME, a first-in-class telomerase inhibitor, was approved in the US and EU for the treatment (tx) of certain adult patients (pts) with LR-MDS and red blood cell (RBC) transfusion-dependent (TD) anemia who relapsed or are refractory (R/R) to, or are ineligible for erythropoiesis-stimulating agents (ESAs) based on the results of the Phase 3 IMerge trial (NCT02598661). In the primary analysis of IMerge (median follow-up, 18.5 mo) IME demonstrated statistically significant RBC transfusion independence (TI) vs PBO for ≥8-wk RBC-TI (primary endpoint; 40% IME vs 15% PBO; P<.001), ≥24-wk RBC-TI (key secondary endpoint; 28% IME vs 3% PBO; P<.001), and ≥1-y RBC-TI (post hoc endpoint; 18% vs 2%; P=.0023). At the time of the primary analysis, data were immature to assess overall survival (OS). Here, we report on secondary endpoints, including OS, progression-free survival (PFS), progression to acute myeloid leukemia (AML), safety, and long-term outcomes by subgroups of interest in IMerge, as well as ad hoc outcomes, including OS by response. Methods: In IMerge, pts with RBC-TD LR-MDS who were R/R to/ineligible for ESAs and lenalidomide- and hypomethylating agent–naive were randomized 2:1 to IME (7.1 mg/kg active dose, equivalent to 7.5 mg/kg IME sodium) or PBO. Kaplan-Meier and stratified Cox proportional hazard model were used; the study was not powered to detect statistical significance. Results: Of 178 pts enrolled in IMerge, 118 received IME and 60 received PBO (median age, 72 y; 62% of pts were male). At data cutoff (05/10/2025), median follow-up was 45 mo. In the intent-to-treat population, median (95% CI) OS was 47.8 mo (38.3-not estimable [NE]) for IME and 44.8 mo (37.4-NE) for PBO; pts treated with IME had 18% lower risk of death (hazard ratio [HR], 0.82 [95% CI, 0.48-1.38]). The landmark OS analysis (≥42 mo) showed a death rate of 11.4% (5/44) for IME and 31.6% (6/19]) for PBO (HR 0.33 [95% CI, 0.09, 1.20], P=.077). Median (95% CI) PFS in all pts was 47.6 mo (29.2-NE) for IME and 42.2 mo (16.7-NE) for PBO (HR, 0.82 [0.43-1.54], favoring IME). Progression to AML remained low in both tx arms (1.7% IME vs 3.3% PBO; HR, 0.45 [0.06-3.23], favoring IME). Among pts who achieved ≥8-wk RBC-TI with IME (n=47), median (95% CI) OS was 47.8 mo (35.3-NE) vs NE (30.7-NE) in nonresponders (n=71; HR, 0.93 [0.51-1.70]). Among pts who achieved ≥24-wk RBC-TI with IME (n=33), median (95% CI) OS was NE (40.4-NE) vs 45.7 mo (29.2-NE) in nonresponders (n=85; HR, 0.72 [0.37-1.38]). Among pts who achieved ≥1-y RBC-TI with IME (n=21), median (95% CI) OS was NE (39.1-NE) vs 47.8 mo (31.8-NE) in nonresponders (n=97; HR, 0.73 [0.35-1.53]). Among pts who achieved a 1.5-g/dL central hemoglobin (Hb) rise from pre-tx lasting ≥8 wk (International Working Group [IWG] 2006 hematologic improvement-erythroid criteria; n=40), median (95% CI) OS was NE (39.1-NE) vs 47.8 mo (29.2-NE) for nonresponders (n=78; HR, 0.74 [0.40-1.37]). Subgroup analyses of median OS generally favored IME over PBO. For pts with RS-positive disease, median OS was 47.6 mo for IME vs 47.2 mo for PBO (HR, 0.81 [95% CI, 0.43-1.52]). For pts with RS-negative disease, median OS was 47.8 mo for IME vs 44.8 mo for PBO (HR, 0.92 [95% CI, 0.34-2.50]). For pts with low-risk disease per International Prognostic Scoring System, median OS was NE for IME vs 47.2 mo for PBO (HR, 0.99 [95% CI, 0.53-1.86]). For pts with intermediate-1–risk disease, median OS was 45.7 mo for IME and 44.1 mo for PBO (HR, 0.52 [0.20-1.37]). For pts with serum erythropoietin (sEPO) ≤500 mU/mL, median OS was 47.6 mo for IME and 52.0 mo for PBO (HR, 0.97 [95% CI, 0.50-1.90]). For pts with sEPO >500 mU/mL, median OS was NE for IME vs 44.1 mo for PBO (HR, 0.54 [95% CI, 0.18-1.60]). For pts with high TB per IWG 2018 criteria, median OS was 44.2 mo for IME vs 44.1 mo for PBO (HR, 0.95 [95% CI, 0.52-1.71]). For pts with low TB, median OS was NE for IME vs 52.0 mo for PBO (HR, 0.23 [95% CI, 0.05-1.13]). Pts treated with IME with SF3B1-mutated disease had superior OS vs those without (HR 0.32 [95% CI, 0.16-0.64]; P<.001). No new safety signals emerged. Conclusions: IME resulted in a favorable trend in OS, PFS, and time to progression to AML vs PBO in the overall population; OS favored IME vs PBO in most subgroups. The OS results correlated with RBC-TI and Hb rise in IME-treated pts. Although not statistically powered to detect significance, these analyses support the clinical benefit of IME in pts with RBC-TD LR-MDS.

Disruption of redox metabolism is a hallmark of drug-resistant cancer cells, representing a major obstacle to the effective treatment of acute myeloid leukemia (AML). While recent studies have highlighted the importance of redox balance in AML therapy, the specific contribution of protein redox signaling to resistance remains poorly understood. Defining these mechanisms could uncover therapeutic vulnerabilities of resistant AML cells and guide the development of novel combination strategies. Here, we performed comprehensive mass spectrometry-based redox and quantitative proteomic profiling of AML cell lines and patient samples sensitive or resistant to the hypomethylating agent azacitidine (AZA). We demonstrate that AZA disrupts redox homeostasis, which inactivates the glyoxalase system and DNA damage response, and thereby induces cell death. In contrast, AZA resistance is associated with a redox reset characterized by elevated glutathione levels and diminished protein S-glutathionylation. Importantly, AZA failed to induce oxidation of proteins in these pathways in resistant cells and patient-derived AML samples. Pharmacological inhibition of glutathione synthesis restored protein S-glutathionylation and resensitized resistant AML cells to AZA.