Rapid bone marrow blast reduction-guided selinexor-based regimen therapy in relapsed/refractory and newly diagnosed acute myeloid leukemia

Abstract Background:Therapeutic options for relapsed/refractory acute myeloid leukemia (R/R AML) remain limited, with poor long-term survival. Our observation of rapid blast decline kinetics in R/R AML patients treated with the selinexor-based regimen predicted therapeutic efficacy. We hypothesized that these kinetics could guide therapy intensification in AML. This prospective study evaluates the XAB regimen, leveraging blast-reduction patterns observed in R/R AML to optimize new diagnosed AML (ND AML) therapy. Methods:The single-arm prospective study was conducted at Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology. Patients aged 18-60years diagnosed with R/R or ND AMLwere included. All received the XAB regimen: selinexor (35 mg/m² twice weekly), azacitidine (75 mg/m²/day days 1–5), and venetoclax (100mg day 1, 200mg day 2, and 400 mg daily days 3- 21). Bone marrow assessments at weeks 1 and 2 were combined to classify blast reduction kinetics: Type A (sensitive): complete remission with/without count recovery (CR/CRi) at both timepoints; Type B (moderately sensitive): partial remission (PR) at week 1 → PR (blast continuously to decline)/CR/CRi at week 2; Type C (low sensitivity): PR/CR/CRi at week 1 → NR at week 2; Type D (resistant): NR at both timepoints. The Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (TJ-IRB20230955, TJ-IRB202502063) approved these prospective studies. Results:Among 16 R/R AML patients, blast kinetics classified 50.0% (n=8) as Type A, 6.3% (n=1) as Type B, 25.0% (n=4) as Type C, and 18.8% (n=3) as Type D. Type A, five of these patients were soon bridged to allogenic hematopoietic stem cell transplantation (allo-HCST). However, Type B/C patients receiving XAB + cytarabine (Ara-C 50mg/day×3-5days, A50) showed Type B had PR and Type C all NR. All Type D patients discontinued therapy without response. At median follow-up of 8.2 months, 62.5% (10/16) survived—50% (5/10) of whom were Type A responders. The XAB regimen demonstrated significant efficacy within the initial two-week period, with no added benefit from the next two weeks of therapy. Consequently, the duration of treatment was restricted to two weeks, to inform subsequent therapeutic decisions. Applying this kinetic-guided algorithm to ND AML patients. Among 21 ND AML patients, 8 (38.1%) achieved CR/CRi at week 1. These early responders continued XAB for a second week, with all 8 (100%) maintaining CR/CRi; transplant-eligible patients will proceed to two cycles of consolidation (X + Ara-C + idarubicin/mitoxantrone) followed by allo-HCST The remaining 13 patients (61.9%) with PR/NR at week 1 received XAB + A50 during week 2. This approach resulted in a 92.3% (12/13) CR/CRi/PR response, indicating efficacy. Following this response, patients will be administered a third week of XAB + A50. If patients are eligible or willing to undergo allo-HCST, consolidation therapy will be administered with the aforementioned. One patient (7.7%) remained NR after two weeks and switched to alternative intensive chemotherapy.Conclusion: Bone marrow blast reduction kinetics (Types A-D) robustly predict therapeutic efficacy of the XAB regimen across AML disease states. In R/R AML, Type A kinetics identify patients benefiting from continued XAB (100% CR/CRi). In ND AML, this regimen achieves 100% CR/CRi in rapid responders. Early kinetic stratification enables real-time treatment adaptation—maintain in responders and prompt transition to intensive therapy in non-responders. This study validates bone marrow blast kinetics as a dynamic biomarker for personalized AML therapy optimization, thereby enabling timely adjustments to treatment regimens.