DEFINITIVE GLOBAL REPORT FLOW CYTOMETRY: LYMPHOCYTE SUBSET ANALYSIS SURVEY 2016/3

Transcription

1 EXPERTISE, SERVICE PROVISION AND CUSTOMER RELATIONS QUALITY OF MEDICAL LABORATORIES CLINICAL BIOLOGY COMMISSION COMMITTEE OF EXPERTS EXTERNAL QUALITY ASSESSMENT IN CLINICAL BIOLOGY DEFINITIVE GLOBAL REPORT FLOW CYTOMETRY: LYMPHOCYTE SUBSET ANALYSIS SURVEY 2016/3 IPH/Flow cytometry/59 Expertise, service provision and customer relations Quality of medical laboratories J. Wytsmanstreet, Brussels Belgium

2 ISSN: COMMITTEE OF EXPERTS IPH (secretariat) TEL: 02/ /22 FAX: 02/ Scheme coordinator: Dr. Van Blerk M. Alternate coordinator: Dr. Vernelen K. TEL: TEL: 02/ / Experts Dr. Bossuyt X. TEL: 016/ FAX: 016/ Dr. Chatelain B. TEL: 081/ FAX: 081/ Dr. De Schouwer P. TEL: 03/ FAX: 03/ Dr. Gothot A. TEL: 04/ FAX: 04/ Dr. Hougardy N. TEL: 063/ FAX: 063/ Dr. Kestens L. TEL: 03/ FAX: 03/ Dr. Kornreich A. TEL: 071/ FAX: 071/ Dr. Mullier F. TEL: 081/ FAX: 081/ Dr. Pradier O. TEL: 02/ FAX: 02/ Dr. Van Bockstaele D. TEL: 015/ FAX: 015/ Dr. Vander Meeren S. TEL: 02/ FAX: 02/ Dr. Willemse J. TEL: 014/ FAX: 014/ Committee of experts: 17/01/2017 Authorisation to release report: by Marjan Van Blerk, scheme coordinator, on 27/01/2017. All the reports are also available on our webpage: FORM 43/124/E v7. 2/41

3 TABLE OF CONTENTS Interpretation of the individual report 4 Sample material 7 Participation 9 Results 10 Next survey 41 FORM 43/124/E v7. 3/41

4 INTERPRETATION OF THE INDIVIDUAL REPORT Besides this global report, an individual report is at your disposal via toolkit. Below you can find information to help you interpreting this report. The position of your quantitative results is presented on the one hand in comparison with the results from all the participants and on the other hand in comparison with the results of the laboratories using your method. Following information is provided: Your result (R) Your method Global median (M g ): central value of the results obtained by all laboratories (all methods together). Global standard deviation (SD g ): measure of the spread of the results obtained by all laboratories (all methods together). Global median of your method (Mm): central value of the results obtained by the laboratories using your method. Standard deviation of your method (SD m ): measure of the spread of the results obtained by the laboratories using your method. The coefficient of variation CV (expressed in %) for all laboratories and for the laboratories using your method: CV m = (SD m / M m ) * 100 (%) and CV g = (SD g / M g ) * 100 (%). Z score: difference between your result and the median of your method (expressed as a number of SD): Z m = (R - M m ) / SD m and Z g = (R - M g ) / SD g. The result is flagged when Z m > 3. U score: relative deviation of your result from the median of your method (expressed in %): U m = ((R - M m ) / M m ) * 100 (%) and U g = ((R - M g ) / M g ) * 100 (%). The result is flagged when U m > d, where d is a parameter-dependent fixed limit, namely the percentage maximal deviation from the method median. A graphical interpretation of the position of your result (R) towards the results of all the participants as well as the results of the participants using your method, based on the method of Tukey, for each parameter and for each analyzed sample. R : your result M m/g : median H m/g : percentiles 25 and 75 I m/g : internal limits (M ± 2.7 SD) O m/g : external limits (M ± 4.7 SD) FORM 43/124/E v7. 4/41

5 The global graph and the one of your method are presented on the same scale, which allows you to compare them. These graphs give you a rough estimation of the position of your result (R) with respect to the medians (M m/g ). More information can be found in 3 brochures available on our website (only in Dutch and French): (choose brochures in the menu) or directly on the following webpage (only in Dutch and French): Informatiebrochure over de externe kwaliteitsevaluatieprogramma's voor klinische laboratoria (Algemene informatiebrochure over de externe evaluatie)/ Brochure d information sur les programmes d évaluation externe de la qualité pour les laboratoires cliniques (Brochure d information générale sur l évaluation externe). 2. Statistische brochure (Algemene statistische berekeningsprocedure opgesteld door Professor Albert)/ Brochure statistique (Procédure générale de calcul statistique mis au point par le professeur Albert). 3. Verwerking van gecensureerde waarden (Statistische berekeningsprocedure toegepast op de gecensureerde waarden opgesteld door Professor Albert)/ Traitement des valeurs censurées (Procédure de calcul statistique appliquée aux valeurs censurées rédigée par le Professeur Albert). FORM 43/124/E v7. 5/41

6 Graphical representation Besides the tables with the results a Box and whisker plot is added. It contains the following elements for the methods with at least 6 participants: a rectangle ranging from percentile 25 (P 25 ) to percentile 75 (P 75 ) a central line representing the median of the results (P 50 ) a lower limit showing the smallest value x > P * (P 75 - P 25 ) an upper limit representing the largest value x < P * (P 75 - P 25 ) all points outside this interval are represented by a dot. O = P * (P 75-P 25) I = P * (P 75-P 25) x < P * (P 75 -P 25 ) LIMITS OF TUKEY H M H I = P * (P 75-P 25) P 50 P 75 P 25 x > P * (P 75 -P 25 ) < value < quantification limit O = P 25-3 * (P 75-P 25) O I H M H I O M - 4.7σ M - 2.7σ P 25 P 75 M + 2.7σ M + 4.7σ Corresponding limits in case of normal distribution FORM 43/124/E v7. 6/41

7 SAMPLE MATERIAL Three blood samples (FC 14822, FC 14823, and FC 14824) collected on K 2 EDTA were sent to the laboratories. Two of these samples were identical (FC 14822/FC 14823). The sample material was collected from 2 healthy and voluntary blood donors at the Dienst voor het Bloed and distributed into aliquots at the Institute of Public Health (IPH, WIV-ISP). The samples were sent by Taxipost 24h and the laboratories were informed by of the send-out of the control material (day 0). The samples tested negative for HIV 1 and 2, hepatitis B surface antigen, hepatitis C and syphilis. Homogeneity was confirmed by hematocrit determination. Control analysis on the day of collection and distribution yielded the following results (UZ Brussel): FC % 10 9 /L Leukocytes 4.5 Lymphocytes 42.6 CD3 + cells CD4 + CD3 + cells CD8 + CD3 + cells CD19 + cells NK cells κ % B lymphocytes 54.2 λ % B lymphocytes 45.8 κ/λ ratio 1.18 FC % 10 9 /L Leukocytes 4.4 Lymphocytes 41.6 CD3 + cells CD4 + CD3 + cells CD8 + CD3 + cells CD19 + cells NK cells κ % B lymphocytes 54.2 λ % B lymphocytes 45.8 κ/λ ratio 1.18 FORM 43/124/E v7. 7/41

8 FC % 10 9 /L Leukocytes 4.8 Lymphocytes 30.1 CD3 + cells CD4 + CD3 + cells CD8 + CD3 + cells CD19 + cells NK cells κ % B lymphocytes 62.6 λ % B lymphocytes 37.3 κ/λ ratio 1.68 FORM 43/124/E v7. 8/41

9 PARTICIPATION Forty-eight laboratories (1 Latvian, 1 Canadian, 1 clinical trial laboratory and 45 Belgian clinical laboratories) participated in the survey 2016/3 (send-out of blood samples: 21 st of November 2016 (day 0)). FORM 43/124/E v7. 9/41

10 RESULTS Due to a problem with the sample delivery to the laboratories by Taxipost 24h, almost all Flemish laboratories received the blood samples only on day 2 (n=22). All other Belgian participants received the blood samples on day 1 (n=23). 53.3% of the Belgian laboratories (n=24) performed the analyses on day 1, 40.0% on day 2 (n=18), 4.4% on day 3 (n=2) and 2.2% on day 4 (n=1). Since the samples are fresh and not stabilised, it is extremely important to perform sample testing as soon as possible upon receipt. Statistics for the evaluation are solely based on the results of the Belgian clinical laboratories (n=45). Statistics for the evaluation of the WBC count, the percentage of lymphocytes by haematology analyser as well as the absolute counts for the different lymphocyte subsets are solely based on the results of the Belgian clinical laboratories that performed the analyses on day 1 or 2 (n=42). Given the problem with the sample delivery, about half of the laboratories were able to analyse the samples only on day 2. To test for the influence of an eventual sample instability, the results obtained on day 1 and day 2 have been compared with each other. No significant difference in flagging rates (Z-scores >3 or <-3) were observed between the results obtained on day 1 compared to day 2. FORM 43/124/E v7. 10/41

11 The following tables shows the medians and coefficients of variation obtained by the Belgian clinical laboratories for the samples FC 14822/FC and FC 14824: FC 14822/FC Median SD CV,% N Leukocytes 10 9 /L Lymphocytes % Haematology analyser Lymphocytes % Flow cytometer CD3 % CD /L CD4 % CD /L CD8 % CD /L CD19 % CD /L NK cells % NK cells 10 9 /L κ % B lymphocytes λ % B lymphocytes κ/λ ratio κ+λ % B lymphocytes Lymphosum FORM 43/124/E v7. 11/41

12 FC Median SD CV,% N Leukocytes 10 9 /L Lymphocytes % Haematology analyser Lymphocytes % Flow cytometer CD3 % CD /L CD4 % CD /L CD8 % CD /L CD19 % CD /L NK cells % NK cells 10 9 /L κ % B lymphocytes λ % B lymphocytes κ/λ ratio κ+λ % B lymphocytes Lymphosum FORM 43/124/E v7. 12/41

13 WBC FORM 43/124/E v7. 13/41

14 Lymphocytes Haematology analyser FORM 43/124/E v7. 14/41

15 Lymphocytes Flow cytometer FORM 43/124/E v7. 15/41

16 WBC 10E9/L Lymphocytes % Haematology analyser Lymphocytes % Flow cytometer FORM 43/124/E v7. 16/41

17 CD3% FORM 43/124/E v7. 17/41

18 CD3 FORM 43/124/E v7. 18/41

19 CD3 % CD3 10E9/L FORM 43/124/E v7. 19/41

20 CD4% FORM 43/124/E v7. 20/41

21 CD4 FORM 43/124/E v7. 21/41

22 CD4 % CD4 10E9/L FORM 43/124/E v7. 22/41

23 CD8% FORM 43/124/E v7. 23/41

24 CD8 FORM 43/124/E v7. 24/41

25 CD8 % CD8 10E9/L FORM 43/124/E v7. 25/41

26 CD19% FORM 43/124/E v7. 26/41

27 CD19 FORM 43/124/E v7. 27/41

28 CD19 % CD19 10E9/L FORM 43/124/E v7. 28/41

29 NK% FORM 43/124/E v7. 29/41

30 NK FORM 43/124/E v7. 30/41

31 NK % NK 10E9/L FORM 43/124/E v7. 31/41

32 FORM 43/124/E v7. 32/41 κ

33 FORM 43/124/E v7. 33/41 λ

34 κ/λ ratio FORM 43/124/E v7. 34/41

35 κ % B lymphocytes λ % B lymphocytes κ/λ ratio v FORM 43/124/E v7. 35/41

36 Sum κ+λ FORM 43/124/E v7. 36/41

37 Lymphosum FORM 43/124/E v7. 37/41

38 Sum κ + λ % B lymphocytes Lymphosum For technical validation purposes it is worth noting that in non-pathological peripheral blood of adults the sum of kappa and lambda (expressed as a % of CD19+ B-cells) should be between 90 and 110. The lymphosum (sum of CD3 + % plus CD19 + % plus CD3 - CD16 + and/or CD56 + %) should equal the purity of the lymphocytes in the gate ± 5%, with a maximum variability of 10%. FORM 43/124/E v7. 38/41

39 Following figures represent the box-and-whisker plots of the differences between the duplicate samples FC and FC expressed towards their mean WBC CD3 CD4 CD8 CD19 NK Kappa/lambda % Results not represented on the graph Parameter Result, % CD CD CD CD CD CD CD CD CD NK NK 31.4 NK 35.5 FORM 43/124/E v7. 39/41

40 Ly HA Ly FC CD3 % CD4 % CD8 % CD19 % NK% Kappa Lambda Kappa+lambda Lymphosum % Results not represented on the graph Parameter Result, % Ly FC Ly FC Ly FC Ly FC 9.8 Ly FC 9.9 CD3 % CD3 % 6.0 CD4 % CD4 % CD4 % 7.5 CD8 % CD8 % NK% NK% 24.8 NK% 36.0 Lambda Kappa+lambda 5.4 Kappa+lambda -4.2 Kappa+lambda -3.6 Kappa+lambda -3.2 Kappa+lambda -2.3 Kappa+lambda -2.1 Kappa+lambda -1.7 Kappa+lambda 2.0 Kappa+lambda 2.6 Kappa+lambda 5.3 Lymphosum Lymphosum -1.8 Lymphosum -1.4 Lymphosum -1.1 Lymphosum 0.9 Lymphosum 1.0 Lymphosum 1.0 Lymphosum 1.5 Lymphosum 1.5 Lymphosum 5.2 FORM 43/124/E v7. 40/41

41 NEXT SURVEY The next survey is scheduled for the 13 th of February END Scientific Institute of Public Health, Brussels This report may not be reproduced, published or distributed without the consent of the WIV-ISP. FORM 43/124/E v7. 41/41