PDF(Sci-hub)
Abstract：
BACKGROUND: Evaluation of paroxysmal nocturnal hemoglobinuria (PNH) clones by flow cytometry (FCM) is not standardized and is associated with consistent inter-laboratory variability.
METHODS: In order to rule out the influence of particular approach in generating final results, we analyzed the performance characteristics of individual consensus strategies for small to intermediate (1%-20%) and minor (<1%) PNH clones within the white blood cell (WBC) and red blood cell (RBC) compartments with sensitivity up to 0.1%.
RESULTS: Coefficient of variation (CV) for precision/reproducibility analysis ranged from 0.67%/1.49% to 2.56%/3.09% for granulocytes, from 0.93%/3.09% to 7.76%/12.06% for monocytes and from 0.41%/4.73% to 6.53%/5.1% for RBCs. Coefficient of determination (r2) for linear regression analysis ranged from 0.95 to 0.99, Wilcoxon ranks test showed no statistically significant differences (p>0.05), Bland-Altman analysis demonstrated performance agreement with mean bias ranging from -0.18 to 1.24.
CONCLUSIONS: Our results confirmed very good performance characteristics for precision and reproducibility analysis, excellent correlation and favorable agreement between strategies, suggesting that reported inter-laboratory variability is related mainly to incorrect performance and/or insufficient experience with PNH testing by flow cytometry, rather than to relevant limitations of any particular approach.
METHODS: In order to rule out the influence of particular approach in generating final results, we analyzed the performance characteristics of individual consensus strategies for small to intermediate (1%-20%) and minor (<1%) PNH clones within the white blood cell (WBC) and red blood cell (RBC) compartments with sensitivity up to 0.1%.
RESULTS: Coefficient of variation (CV) for precision/reproducibility analysis ranged from 0.67%/1.49% to 2.56%/3.09% for granulocytes, from 0.93%/3.09% to 7.76%/12.06% for monocytes and from 0.41%/4.73% to 6.53%/5.1% for RBCs. Coefficient of determination (r2) for linear regression analysis ranged from 0.95 to 0.99, Wilcoxon ranks test showed no statistically significant differences (p>0.05), Bland-Altman analysis demonstrated performance agreement with mean bias ranging from -0.18 to 1.24.
CONCLUSIONS: Our results confirmed very good performance characteristics for precision and reproducibility analysis, excellent correlation and favorable agreement between strategies, suggesting that reported inter-laboratory variability is related mainly to incorrect performance and/or insufficient experience with PNH testing by flow cytometry, rather than to relevant limitations of any particular approach.
摘要：
暂无翻译
