External quality assessment for Flow cytometry

Transcription

1 ISSN IPH J. Wytsmanstreet 14 B-1050 Brussels MINISTRY OF PUBLIC HEALTH CLINICAL BIOLOGY COMMISSION CLINICAL BIOLOGY SECTION COMMITTEE OF EXPERTS External quality assessment for Flow cytometry Annual report 2008 IPH 2008/Flow 31 This report may not be reproduced, published or distributed without permission of the IPH.

2 Ad hoc committee of experts Tel. Fax. IPH (sec.) 02/ / Coordinator Dr. Van Blerk 02/ Dr. Bossuyt X. 016/ / Dr. Chatelain B. 081/ / Dr. Demanet C. 02/ / Dr. De Schouwer P. 03/ / Dr. Hougardy N. 063/ / Dr. Kestens L. 03/ / Dr. Pradier O. 02/ / Dr. Schaaf-Lafontaine N. 04/ / Dr. Van Bockstaele D. 015/ /

3 I. Surveys A voluntary triannual external quality assessment scheme for lymphocyte immunophenotyping is operational in Belgium since Each survey, participating laboratories are sent 3 fresh K2EDTA anticoagulated whole blood samples by overnight mail. The laboratories are surveyed for methodology and are asked to report white blood cell count (WBC), percentage of lymphocytes, percentages and absolute numbers of T (CD3 + ), B (CD19 + ) and NK cells, and of the CD4 + and CD8 + T cell subsets as well as the percentages of κ and λ chain expressing B cells and the κ/λ ratio. In 2008, surveys were conducted in January (FC 8265, FC 8266, FC 8267), April (FC 8498, FC 8499, FC 8500) and October (FC 8988, FC 8989, FC 8990). 1 Canadian, 1 Latvian and 49 Belgian laboratories participated in these surveys. II. Methodologies used (survey 2008/3) The majority of laboratories (76.0%) used dual platform analysis. Following tables provide an overview of the haematology analysers and flow cytometers used: Haematology analyser Number of participants Abbott Cell-Dyn Abbott Cell-Dyn Abbott Cell-Dyn Abbott Sapphire 6 Bayer H1/H2/H3 1 Bayer Advia 120/ Beckman Coulter LH Sysmex XE 2100/XE-alpha 16 Sysmex XT 2000i/XT 1800i 3 Flow cytometry, annual report Printing date: 8/06/2009 1/48

4 Flow cytometer Number of participants Becton Dickinson FACSCalibur 19 Becton Dickinson FACSCanto/FACSCanto II 13 Coulter Cytomics FC Coulter Epics XL/XL-MCL 6 Becton Dickinson FACSCan 2 Becton Dickinson FACSCalibur + FACSCanto 1 Abbott Sapphire 1 All laboratories applied the whole blood lysis technique. The majority of the laboratories (59.6%) used a lyse no wash procedure. Most laboratories used 3-, 4-, 5- or 6-colour combinations. Number of participants CD3 + CD4 + CD8 + CD colour colours colours colours colours colours colours Belgian consensus recommendations have been issued for the performance of many flow cytometric-based tests (Acta Clinica Belgica 1999; 54:88-98). 82.0% of the laboratories used the recommended monoclonal antibody panels for performing CD3, CD4 and CD8 determinations (two colour systems: CD3/CD4 and CD3/CD8; three colour systems: CD3/CD4/CD45 and CD3/CD8/CD45; four colour systems: CD3/CD4/CD8/CD45). Five laboratories still measured total CD4 + and CD8 + cells independently of CD % of the participants used the combination of CD16 and CD56 to identify NK cells. The other participants used either CD16 (n=1) or CD56 (n=2). 84.0% of the laboratories utilized CD45 as gating agent. Flow cytometry, annual report Printing date: 8/06/2009 2/48

5 Following table displays the sample quality control assessment procedures used by the participating laboratories: Sample quality control assessment Number 100% CD45 positive cells 1 + lymphosum 2 + CD3 consistency check % CD45 positive cells 1 + lymphosum 2 9 Lymphosum 2 + CD3 consistency check 3 8 Lymphosum % CD45 positive cells 1 4 CD3 consistency check % CD45 positive cells 1 + CD3 consistency check 3 1 Other 4 None 1 1 CD45 Gating for routine flow cytometric analysis of human bone marrow specimens. Stelzer GT, Shults KE, Loken MR. Annals of the New York Academy of Sciences 1993; 677: Use of CD45 fluorescence and side-scatter characteristics for gating lymphocytes when using the whole blood lysis procedure and flow cytometry. Nicholson JK, Hubbard M, Jones BM. Cytometry 1996; 26: Sum of CD3 + % plus CD19 + % plus CD3 - CD16 + and/or CD56 + % (lymphosum) should equal the purity of lymphocytes in the gate ± 5%, with a maximum variability of 10% 3 Replicate results within a panel (e.g. CD3 + %) for the same sample should be within 5% of each other for FSC/SSC gating or within 3% for CD45/SSC gating III. Results 68.1% (survey 2008/3) to 91.8% (survey 2008/2) of the participants indicating the day of sample testing performed the analyses on day 1 of the survey and 4.1% (survey 2008/2) to 25.5% (survey 2008/3) on day 2 (day 0: send-out of blood samples). The WBC count, the percentage of lymphocytes by haematology analyser as well as the absolute counts for the different lymphocyte subsets were not evaluated for the participants who performed the analyses on day 3 or later, given the instability of the control material. Flow cytometry, annual report Printing date: 8/06/2009 3/48

6 The following tables show the medians and coefficients of variation obtained for the different parameters on the samples sent in 2008: WBC 10 9 /l Median CV.% N FC FC FC FC FC FC FC FC FC Lymphocytes % Haematology analyser Median CV.% N FC FC FC FC FC FC FC FC FC Lymphocytes % Flow cytometer Median CV.% N FC FC FC FC FC FC FC FC FC Flow cytometry, annual report Printing date: 8/06/2009 4/48

7 CD3 % Median CV.% N FC FC FC FC FC FC FC FC FC CD3 Median CV.% N FC FC FC FC FC FC FC FC FC CD4 % Median CV.% N FC FC FC FC FC FC FC FC FC Flow cytometry, annual report Printing date: 8/06/2009 5/48

8 CD4 Median CV.% N FC FC FC FC FC FC FC FC FC CD8 % Median CV.% N FC FC FC FC FC FC FC FC FC CD8 Median CV.% N FC FC FC FC FC FC FC FC FC Flow cytometry, annual report Printing date: 8/06/2009 6/48

9 CD19 % Median CV.% N FC FC FC FC FC FC FC FC FC CD19 Median CV.% N FC FC FC FC FC FC FC FC FC NK % Median CV.% N FC FC FC FC FC FC FC FC FC Flow cytometry, annual report Printing date: 8/06/2009 7/48

10 NK Median CV.% N FC FC FC FC FC FC FC FC FC κ % B lymphocytes Median CV.% N FC FC FC FC FC FC FC FC FC λ % B lymphocytes Median CV.% N FC FC FC FC FC FC FC FC FC Flow cytometry, annual report Printing date: 8/06/2009 8/48

11 κ/λ ratio Median CV.% N FC FC FC FC FC FC FC FC FC Lymphosum Median CV.% N FC FC FC FC FC FC FC FC FC The CVs for the WBC count ranged between 3.7 and 6.2%. The CVs for the % lymphocytes ranged from 3.6 to 5.7% for the % lymphocytes obtained with haematology analysers and ranged from 4.0 to 11.5% for the % lymphocytes obtained with flow cytometers. Three participants performed the differential lymphocyte count manually. For the different lymphocyte subsets, the average between-laboratory variability was 2.7, 4.5, 6.5, 10.6, and 16.6% for the % of CD3 +, CD4 +, CD8 +, CD19 +, and NK cells, respectively. The average CVs of the absolute values were higher and amounted to 7.8, 9.3, 10.6, 14.4 and 20.6% for CD3 +, CD4 +, CD8 +, CD19 +, and NK cells, respectively. The overall CVs were larger for small subsets. For the percentages of κ and λ chain expressing B cells and the κ/λ ratio, the average CVs were 9.0, 12.7 and 12.4%, respectively. Flow cytometry, annual report Printing date: 8/06/2009 9/48

12 The following graphs show for the different parameters the evolution of the interlaboratory variability over the years. The black lines show the mean CV per survey. The grey lines are a smoothed representation of the black lines and depict the evolution of the mean CV over time. For the determination of the percentages and absolute values of CD3, CD4, CD8, and CD19 positive cells, interlaboratory variability has slightly decreased over the years. CD3 % CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

13 CD3 CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

14 CD4 % CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

15 CD4 CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

16 CD8 % CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

17 CD8 CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

18 CD19 % CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

19 CD19 CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

20 NK % CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

21 NK CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

22 Kappa CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

23 Lambda CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

24 Kappa/lambda ratio CV (%) Flow cytometry, annual report Printing date: 8/06/ /48

25 IV. Long-term analytical performance As the last 6 years (2000-2, , , , , ), a linear regression model based on the least-squares method (Clinical Chemistry 2002; 48: ) was applied to evaluate the long-term analytical performance of the laboratories using the data of the external quality assessment programme over the period (9 surveys). Only those laboratories with at least nine test results (3 surveys) were included in the evaluation In this model, the consensus value defined as the median value of all test results (x) is used as the independent and the laboratory result as the dependent value (y). The slope (b) and the variability (standard error, Sy x) of the regression line characterize the curve. The long-term analytical performance is reflected by the analytical coefficient of variation LCVa, which is based on the variability of the regression line and the mean value of all consensus S y x b values: LCVa =.100% X To allow comparison of LCVa values with those from other laboratories the participants are provided with the percentiles of the absolute LCV a values of all participating laboratories (see table below) for each of the parameters. This approach allows the participants to evaluate themselves within the group of participating laboratories. Example LCVa value of 1.9 for the percentage of CD4 1.9 corresponds to p10 (see table below), meaning that only 10% of the laboratories show a better performance for the percentage of CD4. LCV a value of 7.0 for the percentage of CD4 7.0 corresponds to p80 (see table below), meaning that 80% of the laboratories show a better performance for the percentage of CD4. A LCV a value equal to or higher than the 95 th percentile is considered unacceptable. If they are interested, participants who reported an outlying result for one or more parameters can contact the members of the expert committee to examine their data in order to find a possible explanation for the erroneous result. Flow cytometry, annual report Printing date: 8/06/ /48

26 LCV a Percentiles WBC Ly Ly CD3 CD3 CD4 CD4 CD8 CD8 HA FC % % % Percentiles CD19 CD19 NK NK κ λ κ/λ % % Flow cytometry, annual report Printing date: 8/06/ /48

27 The evolution of the LCVa values over time is summarized in the following tables: LCV a WBC Percentiles LCV a Lymphocytes Haematology analyser Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

28 LCV a Lymphocytes Flow cytometer Percentiles LCV a CD3 % Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

29 LCV a CD3 Percentiles LCV a CD4 % Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

30 LCV a CD4 Percentiles LCV a CD8 % Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

31 LCV a CD8 Percentiles LCV a CD19 % Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

32 LCV a CD19 Percentiles LCV a NK % Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

33 LCV a NK Percentiles LCV a κ % B lymphocytes Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

34 LCV a λ % B lymphocytes Percentiles LCV a κ/λ ratio Percentiles Flow cytometry, annual report Printing date: 8/06/ /48

35 The following graphs show the evolution of the 5 th, 25 th, 50 th, 75 th and 95 th percentiles of the log(lcva) values of the participants over the years. Each percentile line depicts the evolution of different segments of the laboratories. The 25 th, 50 th and 75 th percentiles depict the evolution of the majority of the laboratories. The 5 th percentile depicts the evolution of the best performing laboratories. The 95 th percentile line depicts the evolution of the worst performers. The different lines may evolve in independent ways. For WBC count, % lymphocytes obtained by haematology analyser and the percentages and absolute values of CD3, CD4, CD8, and CD19 positive cells, laboratory performance has significantly improved for the majority of the laboratories (simple linear regression analysis). WBC log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

36 Lympho HA log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

37 Lympho FC log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

38 CD3 % log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

39 CD3 log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

40 CD4 % log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

41 CD4 log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

42 CD8 % log(lcva) P 95 P 75 P 50 P 25 P 5 Flow cytometry, annual report Printing date: 8/06/ /48

43 CD8 log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

44 CD19 % log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

45 CD19 log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

46 NK % log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

47 NK log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

48 Kappa log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

49 Lambda log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48

50 Kappa/lambda ratio log(lcva) P 95 P 75 P 50 P 25 P Flow cytometry, annual report Printing date: 8/06/ /48