BACKGROUND: Minimal residual disease (MRD) is a powerful prognostic indicator in childhood acute lymphoblastic leukemia (ALL). Multiparametric flow cytometry (FC) is a rapid and sensitive methodology for detection of MRD, applicable for most patients and is being incorporated in multicenter treatment protocols. The influence of different techniques and of individual interpretation of data on the interlaboratory variability in FC-MRD determinations has not been described. METHODS: We compared FC-MRD of identical bone marrow samples processed as either Ficoll separated mononuclear cells or lyse and wash nucleated cells (NC) in two central laboratories of a national multicenter childhood ALL study. A total of 290 samples at diagnosis and 494 follow-up samples (Day-15 n = 261; Day-33 n = 233) were analyzed. A group of 52 paired list mode data (LMD) of D-15 and D-33 samples was blindly reanalyzed by both laboratories. RESULTS: Pearson correlations for all samples of D-15 (n = 261) and D-33 (n = 233) were 0.875 and 0.82, respectively (P < 0.001), being lower for T-ALL 0.716 and 0.719, respectively. Quantitative concordance defined as less than 0.5 log difference in MRD measured by the two methodologies was 80.8% at D-15 but only in 57.9% at D-33. Reanalysis of LMD revealed that data interpretation explained half of the discordance. CONCLUSIONS: FC-MRD analysis of childhood ALL is a robust method during the earliest phases of induction therapy in a multicentric setting. Standardization of data analysis could improve about half of the discordance between different technical approaches.
BACKGROUND: Minimal residual disease (MRD) is a powerful prognostic indicator in childhood acute lymphoblastic leukemia (ALL). Multiparametric flow cytometry (FC) is a rapid and sensitive methodology for detection of MRD, applicable for most patients and is being incorporated in multicenter treatment protocols. The influence of different techniques and of individual interpretation of data on the interlaboratory variability in FC-MRD determinations has not been described. METHODS: We compared FC-MRD of identical bone marrow samples processed as either Ficoll separated mononuclear cells or lyse and wash nucleated cells (NC) in two central laboratories of a national multicenter childhood ALL study. A total of 290 samples at diagnosis and 494 follow-up samples (Day-15 n = 261; Day-33 n = 233) were analyzed. A group of 52 paired list mode data (LMD) of D-15 and D-33 samples was blindly reanalyzed by both laboratories. RESULTS: Pearson correlations for all samples of D-15 (n = 261) and D-33 (n = 233) were 0.875 and 0.82, respectively (P < 0.001), being lower for T-ALL 0.716 and 0.719, respectively. Quantitative concordance defined as less than 0.5 log difference in MRD measured by the two methodologies was 80.8% at D-15 but only in 57.9% at D-33. Reanalysis of LMD revealed that data interpretation explained half of the discordance. CONCLUSIONS: FC-MRD analysis of childhood ALL is a robust method during the earliest phases of induction therapy in a multicentric setting. Standardization of data analysis could improve about half of the discordance between different technical approaches.