DP089 | DEVELOPMENT OF A PERSONALIZED LEUKEMIC STEM CELLS MINIMAL RESIDUAL DISEASE IN ACUTE MYELOID LEUKEMIA PATIENTS: THE ROLE OF NEXT GENERATION FLOW AS A SPECIFIC TOOL FOR DEFINITION OF PROGNOSIS AND PROGRESSION

Introduction. Leukemic Stem Cells (LSCs) are responsible for high relapse rates in Acute Myeloid Leukemia (AML). Next Generation Flow (NGF) is a technology that permits the detection of these cells with high sensitivity, even when present at a low burden. In our study we want to identify and characterize the phenotype of AML-LCSs at single patient level, in order to create a personalized LSCs-Minimal Residual Disease (MRD) that could improve the prognostication of patients. Methods. Bone Marrow (BM) samples have been collected and AML-LSCs have been identified with a unique flow cytometry antigen panel comprising CD34, CD38, CD90, CD123, CD45RA, CLL-1 and TIM-3. A strict gating strategy has been developed to characterize pts specific LSCs that generally express CD34, are negative for CD38, lack CD90 while express CD45RA, CD123, CLL-1 and TIM-. In order to discriminate LSCs from Hematopoietic Stem Cells (HSCs), normal BM samples and BM suspected for Myelodysplastic Syndromes (MDS) have been collected and a platform of data has been created to compare them to AML. Molecular and cytogenetic analysis have been performed in order to classify AML pts into favourable, intermediate and adverse, according to European Leukemia Network (ELN) guidelines. Results. Our study comprises 48 AML patients: 31/48 (65%) of them have been enrolled from diagnosis, and a personalized AML-LSCs panel has been created to monitor these cells alongside treatment; 17/48 (35%) AML patients have been searched for LSCs while they were already undergoing consolidation therapy. We have identified AML-LSCs in all the tested samples, and differentiated them from normal Hematopoietic Stem Cells (HSCs) by comparing AML samples to 7 normal BM and 40 BM of MDS patients. In particular, LSCs were present with a higher burden in pts during relapse or with no response to induction therapy [6/48 (12,5%) patients]. Of note, just 13/48 (27%) patients had an MRD evaluable by flow, but by constructing a LSCs-MRD we have been able to monitor at individual level the behaviour of these cells, in order to predict relapse or no response to therapy. Summary/Conclusion. These data corroborate the possibility to detect AML-LSCs by NGF. In the next future the accrual of more data from a higher cohort of AML patients will allow us to explore the possible prognostic role of this patient-specific reservoir, and to correlate LSCs persistence with reduced response to therapy and a higher chance of relapse.