The transition from myelodysplastic neoplasm to secondary acute myeloid leukemia is revealed by molecular analysis, while functional drug screening demonstrates novel sensitivity patterns with clinical implication

Abstract Background Survival of patients with high or very high-risk myelodysplastic neoplasm (MDS) by IPSS-M is decreased due to transformation to secondary AML (sAML). There are only a limited number of agents that are FDA approved for treatment of MDS including BCL2 and IDH inhibitors, hypomethylating agents (HMAs), ESAs, luspatercept and imetelstat. For patients who relapse after treatment or transform to sAML, there are few effective treatment options, and prognosis is poor. To identify promising new agents and combinations, evaluate individual patient responses and create predictive models, we have developed a new multicolor flow cytometry based high throughput drug screen (MFC-HTS) using the iQue3 (Sartorius) for MDS. Methods We have collected longitudinal sequencing data from 22 patient samples from the City of Hope POSEIDON databases that includes mutation data for 523 genes. We also analyzed public clinical and multiomics datasets for MDS and sAML to understand the transition (Menssen AJ et al. Blood Cancer 2022). The difference between variant allele frequency (VAF) between the MDS and sAML samples was estimated using the Wilcoxon test. We developed a state-of-the-art MFC-HTS with a custom drug panel of approved and investigational agents, able to directly assess viability of the CD34+ blast population from patient sample mononuclear cell populations in the presence of components of the marrow immune microenvironment including lymphocytes and stromal cells. The assay includes 64 drugs, both FDA approved and investigational, that can be assayed at 5 concentrations in a single 384 well plate. We have tested primary MDS patient samples, as well as sAML samples that arose from MDS. We utilized Forecyt® software to generate drug sensitivity curves, determine IC50 and AUC and Breeze 2.0 to generate Drug sensitivity scores (DSSs), heatmaps, and waterfall plots of the DSSs ranking the top agents (highest likelihood of sensitivity) and resistant drugs. Similar methodology (Kytölä S et al., Blood 2025) has been able to prospectively demonstrate accurate prediction of clinical response for patients with de novo or relapsed AML. Ranksum test was used to compare the difference in DSS between the MDS and sAML samples. Results Mutation analysis of the 22 serial samples revealed different patterns of clonal evolution during the disease progression during MDS or the transition to sAML. Among them, one group showed acquisition of new mutations, such as FLT3 or NRAS, one group showed stable clones, and a third group showed clonal change during the time course with acquisition of multiple mutations. Comparing the VAF of individual variants from a published dataset (Menssen AJ et al., Blood Cancer Disc, 2022), we observed most of the variants show increased VAF at the AML compared to the MDS stage. There was a significant difference (p-value 0.018, Wilcoxon test) for the VAF between the paired MDS and AML samples for the AML driver mutations including NRAS, PTPN11, and FLT3, similar to what we observed in our cohort. We then sought to identify the differences in drug sensitivity for MDS vs. sAML. For MDS patients, we identified the most effective agents with the proportion of sensitive samples noted in parentheses: mitoxantrone (100%), olutasidenib (78%), venetoclax (67%), dinaciclib (67%), trametinib (67%), olaparib (56%), GSK3368715 (56%), pacritinib (44%), lenalidomide (44%), fludarabine (55%), tasquinimod (44%), veliparib (44%). For sAML patients, we identified lenalidomide, olutasidenib, GSK3368715 have high DSSs in all tested samples, while fludarabine, fedratinib, tasquinimod, trametinib, veliparib, mitoxantrone and olaparib have high DSSs in 75% of the AML samples. PARP inhibitors veliparib and olaparib, exhibit sensitivity in 5/9 MDS and 3/4 sAML patient samples, indicating their potential application in MDS and sAML. MDS with splicing mutations was previously shown to exhibit sensitivity to PARP inhibitors (Liu ZS et al. Cancer Research 2024), and our samples included SF3B1 and U2AF1 mutations. Summary This study of the transition of MDS to sAML highlights molecular changes and reveals new drug sensitivity with agents that could be utilized to optimize therapy for high/very high risk MDS patients or those in transition to sAML, to be determined in future clinical trials. We noted in vitro drug sensitivity to PARP, IDH, JAK2, BCL2 inhibitors and other agents with drug sensitivity scores that would predict clinical response.