Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/pubmed-central/performance-of-cellavision-dm96-in-leukocyte-classification-aTVm0sY0Ek
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- Pubmed Central
- Copyright
- Copyright: © 2013 Lee LH.
- ISSN
- 2229-5089
- eISSN
- 2153-3539
- DOI
- 10.4103/2153-3539.114205
- Publisher site
- See Article on Publisher Site
Abstract
Background: Leukocyte differentials are an important component of clinical care. Morphologic assessment of peripheral blood smears (PBS) may be required to accurately classify leukocytes. However, manual microscopy is labor intensive. The CellaVision DM96 is an automated system that acquires digital images of leukocytes on PBS, pre‑classifies the cell type, and displays them on screen for a Technologist or Pathologist to approve or reclassify. Our study compares the results of the DM96 with manual microscopy. Methods: Three hundred and fifty‑nine PBS were selected and assessed by manual microscopy with a 200 leukocyte cell count. They were then reassessed using the CellaVision DM96 with a 115 leukocyte cell count including reclassification when necessary. Correlation between the manual microscopy results Access this article online and the CellaVision DM96 results was calculated for each cell type. Results: The Website: correlation coefficients (r ) range from a high of 0.99 for blasts to a low of 0.72 for www.jpathinformatics.org metamyelocytes. Conclusions: The correlation between the CellaVision DM96 and DOI: 10.4103/2153-3539.114205 manual microscopy was as good or better than the previously published data. The Quick Response Code: accuracy of leukocyte classification depended on the cell type, and in general, there was lower correlation for rare cell types. However, the correlation is similar to previous studies on the correlation of manual microscopy with an established reference result. Therefore, the CellaVision DM96 is appropriate for clinical implementation. Key words: Digital pathology, hematopathology, peripheral blood smear BACKGROUND classification of abnormal and immature cells, and do not provide morphological information. Current procedures Complete blood counts and differentials have been indicate that when certain criteria are met during an important and integral component of the clinical the blood count, the blood sample will be flagged for [1] management of patients, since its introduction a peripheral blood smear (PBS). These criteria are set by [2] over 100 years go by Ehrlich. Today, automated the machine manufacturer and laboratory using the rules blood cell counters based on laser‑light scatter and such as those published by the International Consensus [3] flow cytometry principles, such as the Coulter counter Group for Hematology Review. Reasons for PBS have become standard for most blood counts and microscopy include evaluation of immature and abnormal differentials and are able to provide a five to six part white cell, review of red cell morphology in hemolytic leukocyte differential. However, they are unreliable in the conditions and platelet morphology in thrombocytopenia, J Pathol Inform 2013, 1:14 http://www.jpathinformatics.org/content/4/1/14 and confirmation of abnormal or unexpected blood lymphocytes, blasts, and left shifted blood counts. For the slides whole venous blood samples were collected counts. in K EDTA vaccutainer tubes (BD diagnostics Franklin Currently at most institutions, PBS is assessed manually Lakes, NJ USA). Samples were processed utilizing an by laboratory technologists and pathologists. However, this automated hematology analyzer‑LH750 (Beckman‑coulter, practice is labor intensive and time consuming. Moreover, Brea, California). These instruments were equipped with there is substantial inter‑ and intra‑observer variability in automated slide maker and stainer. PBS were stained with [4,5] this process, which negatively impact efficiency. since the Wright‑Giemsa stain. there is increasing demand on the hematopathology services to provide improved turnaround times and 24 h Manual differential counts of 400 white blood cells were performed at various participating sites by service for clinicians in addition to increasing volumes, two (200 cell count each) experienced laboratory there is an eminent need for improved systems with enhanced productivity. technologists trained in PBS morphology. The PBS slides were then analyzed by the DM96 slide scanning unit with As a result, there has been research and development CellaVision Blood Differential Software (CellaVision, into automation of morphological PBS assessment. AB, Lund, Sweden). This system automatically selects The first automated morphological assessment system 115 leukocytes for image analysis. The actual white blood introduced with the Cydac Scanning Microscope System cells counted may have been lower when non‑leukocytes [6] (Cydac, Uppsala, Sweden) in 1966. Early systems such as were mistakenly counted by the machine. The same this were not adopted as they were slow and had limited technologists who performed the manual differential [1,7,8] automation and poor accuracy. In recent years, there also performed the CellaVision reclassification; however has been significant improvement in technology with new to minimize the influence of prior exposure of PBS machines such as the CellaVision Diffmaster Octavia, the slide morphology on DM96 analysis; the slides were CellaVision DM96, and the next slide digital review network. randomly selected, and analysis was performed at Many of these systems, including the CellaVision DM96, different time points. Statistics were performed with operate by scanning barcode‑labeled Wright‑Giemsa Microsoft Excel 2010 (Microsoft, Redmond, WA) and stained slides at low power to locate white blood cells. SPSS Statistics 19 (IBM, Armonk, NY). The white The system then takes an image of each white blood cell blood cell differential percentage for each cell type was at high power. These images are analyzed by an artificial compared between manual counts by the technologists neural network based on a large database of cells to and automated counts by the DM96 by calculation of pre‑classify the type of leukocytes into subtypes including: Pearson’s correlation coefficient and represented visually by scatter plots. A P ≤ 0.05 was selected as the level of Band neutrophils, segmented neutrophils, lymphocytes, significance. eosinophils, monocytes, promyelocytes, myelocytes, metamyelocytes, and blasts. The cells are then presented on a computer display for someone to review and confirm RESULTS or reclassify if incorrectly pre‑classified. In addition, Three hundred and fifty nine blood smears were included the systems can review red blood cell morphology and in the study. The percentage cell type breakdown by estimate platelet counts. manual count and as counted by the DM96 is presented The objective of our study is to assess the ability of in Table 1. Figure 1 shows scatterplots for each leukocyte the CellaVision DM96 (DM96) system and software to cell type. The correlation coefficients (r ) range from a classify leukocytes by comparing it with the manual PBS high of 0.99 for blasts to a low of 0.72 for metamyelocytes. examination. Due to low cell counts, metamyelocytes, myelocytes, and promyelocytes were grouped together as “immature METHODS granulocytes” for analysis. The study was performed at five sites at Calgary CONCLUSIONS Laboratory Services. There were three academic adult hospitals including one Tertiary Care Academic Centre, Examination of PBS is a labor intensive and time one academic children’s hospital, and one community consuming, but clinically necessary, activity in today’s laboratory servicing the metropolitan area. hematopathology laboratory. We evaluated the Three hundred and fifty nine PBS slides were included performance of the DM96 automated morphologic PBS in the study. They were selected from PBS performed analysis system in classifying the leukocytes in 359 cases by Calgary Laboratory Services as part of routine and compared the results to a manual assessment of the clinical service. Smears were selected to include PBS. Our data showed excellent correlation (r > 0.90) examples of leukocyte abnormalities including, atypical between the DM96 and manual microscopy for segmented J Pathol Inform 2013, 1:14 http://www.jpathinformatics.org/content/4/1/14 Figure 1: Comparison between manual differential counts and DM96 differential counts for each leukocyte cell type Table 1: Characteristics of peripheral blood slides grouped together as immature granulocytes there was included in this study excellent correlation. Cell type N Manual DM96 Our results are consistent with several previous studies of [9‑12] the DM96, and a study of the next slide digital review Average SD Average SD [13] network [Table 2]. Notably, our eosinophil correlation % % % % 2 [9] was higher compared to previous findings (r = 0.50 Segmented neutrophils 359 48.3 24.7 48.4 25.2 [12] to 0.85 ). The reasons for this are not clear but it is Lymphocytes 359 29.0 23.2 28.3 23.3 important to note that we had generally higher numbers Monocytes 359 7.7 7.2 8.5 7.7 of eosinophils in our selected samples compared to the Eosinophils 359 2.3 4.2 2.2 4.4 previous studies. Basophils 359 0.7 1.8 0.7 1.9 Bands 359 4.0 7.0 3.5 6.4 Our results for myelocytes and promyelocytes were Metamyelocytes 359 1.2 2.4 1.5 2.8 significantly better than the only previous study that Myelocytes 359 1.1 2.7 1.2 3.0 assessed these cells (myelocytes: r = 0.88 for our study, 2 [9] 2 Promyelocytes 359 0.1 0.6 0.2 1.0 vs. r = 0.37 in Briggs et al.; promyelocytes: r = 0.74 for 2 [9] Immature granulocytes 359 2.4 5.1 3.0 5.9 our study, vs. r = 0.42 in Briggs et al. ). Their correlation Blasts 359 5.6 18.0 5.4 17.6 for metamyelocytes was higher than ours (r = 0.72 for 2 [9] SD: Standard deviation our study, vs. r = 0.93 in Briggs et al. ). These results may be due to the low numbers of immature cells in neutrophils, lymphocytes, and blasts. Lower correlations their study as they included a large number of normal were seen with eosinophils, monocytes, basophils, and PBS (45/136) while our data does not include normal bands. Metamyelocytes, myelocytes, and promyelocytes smears. The aggregation of metamyelocytes, myelocytes, also showed lower correlations. However, when these were and promyelocytes provided an improved correlation J Pathol Inform 2013, 1:14 http://www.jpathinformatics.org/content/4/1/14 Table 2: Correlation coefficients between DM96 and manual microscopy in the classification of leukocytes. Correlation for the nextslide digital review network and correlation between technologists and an expert reference are included for comparison Cell type This Briggs Kratz Cornet Ceelie Yu Koepke [9] [10] [11] [12] [13] [4] study et al.* et al. et al. et al. et al.** et al.*** Neutrophils (total) 0.9859 0.9536 0.9134 Lymphocytes 0.9547 0.9591 0.9393 0.9405 0.901 0.73 Monocytes 0.8316 0.805 0.6658 0.7004 0.8176 0.41 Eosinophils 0.8821 0.672 0.73 0.846 0.7671 0.83 Basophils 0.7637 0.0534 0.5592 0.32 Segmented neutrophils 0.9611 0.8771 0.9528 0.87 Bands 0.874 0.6852 0.7961 0.8868 Metamyelocytes 0.717 0.9331 Myelocytes 0.8806 0.3709 Promyelocytes 0.7357 0.4175 Blasts 0.9861 0.9953 0.9 0.984 0.9769 Immature granulocytes 0.9064 0.9514 0.9285 (meta‑, myelo‑, and promyelocytes) Atypical lymphocytes 0.9326 *Cells per liter used for correlation coefficient calculation rather than percentage of cell type, **Correlation between nextslide digital review network and manual microscopy, ***Correlation between 73 technologists and expert reference compared to the individual cell subtypes in our study of smears referred for pathologist’s review is yet to be [9] 2 2 as well as in Briggs et al. (r = 0.91 and r = 0.95 seen through case controlled studies in the future. respectively). This is likely due to the ability of the There are several limitations to our study. First, DM96 and technologists to easily identify immature our selection of PBS is not random. This may have granulocytes (of which the majority would be of produced a biased result. However, this may also neutrophilic lineage), but difficulty due to subjectivity in represent a useful aspect of our study as one goal of subclassifying their maturity. automation is for the DM96 to classify abnormal Basophil identification also had a significantly better smears. The percentage of abnormal smears is low correlation in our study compared to the only previously in routine hematopathology practice and a random published result published (r = 0.76 for our study, vs. sample would assess very few abnormal smears. Despite 2 [9] r = 0.05 in Briggs et al. ). Again, this may be due to the the non‑random selection of cases, basophils, and low number of basophils in the older study. Our study immature granulocytes remain low in numbers for our potentially had more basophils as we did not include analysis. Another limitation is that the technologists normal PBS. It is interesting to note that in a study by participating at each laboratory involved in our study [4] Koepke et al. where the correlation in the classification only analyzed the slides from their own hospital of PBS cells between 73 technologists and expert sites. Within each laboratory, a small group of 2 or 3 references showed an r of only 0.32 for basophils. This technologists manually read the PBS and reclassified may be due to improved technologist training in recent the results of the DM96. As a result, there is a large years. It may also be related to the fact that in our study, probability that the same technologist manually read basophils are being identified by the same small group the slide and also reclassified the DM96 results. This of technologists in both the manual microscopy and in may artificially improve the correlation between the the reclassification of CellaVision cells, while in Koepke manual and automated classification methods. Finally, [4] et al., the reference cell identification was performed by our study, similar to most previous studies, focused on a separate group of individuals. the reclassified cell results rather than the automated pre‑classified results. Although, this is how the DM96 Overall, the correlation between the DM96 and will be used in clinical practice, the resultant studies manual microscopy in our study is similar to, and in ultimately assess the ability of technologists to identify some cases, better than the range of variance between cells on a computer screen rather than test the ability individual technologists. This is demonstrated in Table 2, of the DM96 algorithms to classify cells. As a result, comparing our correlation coefficients to those of the the studies, including our study are unable to test the [4] study by Koepke et al. Previous precision studies have possibility of full automation of PBS analysis. [9,12] also confirmed this. The impact of this enhanced efficiency and inter‑observer correlation on the number Our study shows that the DM96 is a useful tool in the J Pathol Inform 2013, 1:14 http://www.jpathinformatics.org/content/4/1/14 visual differential leukocyte counting method. Blood Cells 1985;11:173‑86. examination of PBS morphology with performance 5. Rümke CL. Statistical reflections on finding atypical cells. Blood Cells similar to that of manual microscopy. The DM96 will 1985;11:141‑4. only be useful if it is cost effective, particularly as 6. Prewitt JM, Mendelsohn ML. The analysis of cell images. Ann N Y Acad Sci cost control has become essential in the current era of 1966;128:1035‑53. 7. Riley RS, Ben‑Ezra JM, Massey D, Cousar J. The virtual blood film. Clin Lab economic uncertainty. It has previously been shown that Med 2002;22:317‑45. the DM96 is faster than manual smear examination an 8. Beksaç M, Beksaç MS, Tipi VB, Duru HA, Karakás MU, Cakar AN. An artificial aspect confirmed in our internal departmental validation intelligent diagnostic system on differential recognition of hematopoietic [9,10,12,14,15] studies. This is of particular importance as cells from microscopic images. Cytometry 1997;30:145‑50. labor is one of the major expenditures in the laboratory. 9. Briggs C, Longair I, Slavik M, Thwaite K, Mills R, Thavaraja V, et al. Can automated blood film analysis replace the manual differential? An evaluation In addition, digitized images provide many advantages of the CellaVision DM96 automated image analysis system. Int J Lab compared to manual slide microscopy. Images of Hematol 2009;31:48‑60. individual leukocytes can be stored for educational 10. Kratz A, Bengtsson HI, Casey JE, Keefe JM, Beatrice GH, Grzybek DY, et al. activities, quality control, and expert consultation. Images Performance evaluation of the CellaVision DM96 system: WBC differentials may be transmitted from remote locations or areas with a by automated digital image analysis supported by an artificial neural network. Am J Clin Pathol 2005;124:770‑81. lack of trained technologists to institutions with expertise. 11. Cornet E, Perol JP, Troussard X. Performance evaluation and relevance Finally, it has been shown that education using images of the CellaVision DM96 system in routine analysis and in patients with captured by systems such as the DM96 allows for quicker malignant hematological diseases. Int J Lab Hematol 2008;30:536‑42. [16] leukocyte recognition among new trainees. Future 12. Ceelie H, Dinkelaar RB, van Gelder W. Examination of peripheral blood films using automated microscopy; evaluation of Diffmaster Octavia and studies include the assessment of the ability of the DM96 Cellavision DM96. J Clin Pathol 2007;60:72‑9. to identify specific PBS morphologic diagnoses. 13. Yu H, Ok CY, Hesse A, Nordell P, Connor D, Sjostedt E, et al. Evaluation of an automated digital imaging system, Nextslide Digital Review Network, REFERENCES for examination of peripheral blood smears. Arch Pathol Lab Med 2012;136:660‑7. 14. Billard M, Lainey E, Armoogum P, Alberti C, Fenneteau O, Da Costa L. 1. Tatsumi N, Pierre RV. Automated image processing. Past, present, and future of blood cell morphology identification. Clin Lab Med 2002;22:299‑315, viii. Evaluation of the CellaVision DM automated microscope in pediatrics. Int J 2. Ehrlich P. Farbenanalytische Untersuchungen Zur Histologie Und Klinik Des Lab Hematol 2010;32:530‑8. 15. Rollins‑Raval MA, Raval JS, Contis L. Experience with CellaVision DM96 Blutes: Gesammelte Mittheilungene. Berlin, Germany: Hirschwald; 1891. for peripheral blood differentials in a large multi‑center academic hospital 3. Barnes PW, McFadden SL, Machin SJ, Simson E, international consensus group for hematology. The international consensus group for hematology system. J Pathol Inform 2012;3:29. review: Suggested criteria for action following automated CBC and WBC 16. Horiuchi Y, Tabe Y, Idei M, Bengtsson HI, Ishii K, Horii T, et al. The use of differential analysis. Lab Hematol 2005;11:83‑90. CellaVision competency software for external quality assessment and 4. Koepke JA, Dotson MA, Shifman MA. A critical evaluation of the manual/ continuing professional development. J Clin Pathol 2011;64:610‑7.
Journal
Journal of Pathology Informatics – Pubmed Central
Published: Jun 29, 2013