PF178 DEFINITION OF REPRESENTATIVE BONE MARROW SAMPLE BASED ON PARALLEL EVALUATION OF PERIPHERAL BLOOD DURING THE THERAPY OF ACUTE LYMPHOBLASTIC LEUKEMIA

Background: Acute lymphoblastic leukemia (ALL) treatment leads to elimination of the blasts and subsequent regeneration of non‐malignant populations. Amount of minimal residual disease (MRD) is the strongest prognostic marker and its value differs between bone marrow (BM) and peripheral blood (PB). Therefore for patients stratification is necessary to evaluate non‐hemodiluted BM sample. Our AIEOP BFM 2009 protocol considers a BM sample non‐hemodiluted if >2% of its cells are erythroid precursors (EP). EPs are defined as CD19neg(orCD7neg)CD45neg cells Aims: 1. To evaluate utility of the currently used marker of BM representativeness. 2. To define new, more exact approach for definition of representative BM. Methods: Patients (n = 528) diagnosed with ALL between 2007 and 2015 were included into the study. We analyzed proportion of EPs defined as CD19neg(orCD7neg)CD45neg in BM and in parallel also in PB at day 15 together with viability dye DAPI. We tested whether a new definition of EPs including an added CD71 marker (transferrin receptor expressed since early phases of erythropoiesis) improves its specificity. Then we evaluated other non‐malignant populations (n = 342) in BM and PB for the new definition of representative BM (lymphocytes, granulocytes, monocytes, erythroid precursors, mature B cells, T cells and NK cells in both B cell precursor and T ALL; CD4+, CD8+ and CD4negCD8neg T lymphocytes in T ALL only). Results: 1. We identified high proportion of apoptotic cells in EPs defined as CD19neg(orCD7neg)CD45neg (6.5–96%, median 55%) The population defined as CD19neg(orCD7neg)CD45negCD71pos was more viable (apoptotic cells 0–66%, median 9%) p < 0.0001. Then we evaluated population CD19neg(orCD7neg)CD45neg in PB. We found that as many as 29% of PB samples have >2% of EP (the value that is currently used as a marker of BM without hemodilution). 2. We defined a new assessment of BM representativeness based on comparison of non‐malignant populations between BM and PB. With increasing difference between the proportion of any population between BM and PB, the possibility of hemodilution lowers. According to this approach we can calculate maximal possible hemodilution and its influence on stratification of the patient. We evaluated 342 patients with available paired BM‐PB samples and identified 6 patients with highly hemodiluted BM samples who might have been misstratified. Summary/Conclusion: We defined a new, more accurate approach for assessment of BM representativeness by comparison of non‐malignant populations between PB and BM. By this approach we can exclude hemodilution in majority of the samples. The currently used marker of representativeness – proportion of EPs, defined as CD19neg(orCD7neg)CD45neg – is not sufficient for assessment of BM representativeness. If proportion of EPs is required, CD71 should be added for the definition. Supported by grants NV18–03–00343, NV18–07–00430 and LO1604.