Kristina Meichner1, Tracy Stokol2, Jaime Tarigo1, Anne Avery3, Mary J Burkhard4, Stefano Comazzi5, Jonathan Fogle6, Devorah Marks Stowe6, Barbara Rütgen7, Davis Seelig8, Adi Wasserkrug-Naor9, William Vernau10, Dorothee Bienzle11. 1. Department of Pathology, College of Veterinary Medicine, University of Georgia, Athens, GA, USA. 2. Department of Population Medicine and Diagnostic Sciences, Cornell University, Ithaca, NY, USA. 3. Department of Microbiology, Immunology & Pathology, University of Colorado, Fort Collins, CO, USA. 4. Department of Veterinary Biosciences, The Ohio State University, Columbus, OH, USA. 5. Department of Veterinary Medicine, University of Milan, Milan, Italy. 6. Department of Population Health and Pathobiology, North Carolina State University, Raleigh, NC, USA. 7. University of Vienna, Vienna, Austria. 8. Department of Veterinary Clinical Sciences, University of Minnesota, St Paul, MN, USA. 9. Department of Diagnostic Medicine/Pathobiology, Kansas State University, Manhattan, KS, USA. 10. Department of Pathology, Microbiology & Immunology, University of California, Davis, Davis, CA, USA. 11. Department of Pathobiology, University of Guelph, Guelph, ON, Canada.
Abstract
BACKGROUND: Flow cytometry (FC) is used increasingly in veterinary medicine for further characterization of hematolymphoid cells. Guidelines for optimizing assay performance and interpretation of results are limited, and concordance of results across laboratories is unknown. OBJECTIVES: This study aimed to determine inter-investigator agreement on the interpretation of FC results from split samples analyzed in different laboratories using various protocols, cytometers, and software; and on the interpretation of archived FC standard (FCS) data files contributed by the different investigators. METHODS: This was a multicenter observational cross-sectional study. Anticoagulated blood or lymph node aspirate samples from nine client-owned dogs were aliquoted and shipped to participating laboratories. Samples were analyzed with individual laboratory-developed protocols. In addition, FCS files from a set of separate samples from 11 client-owned dogs were analyzed by participating investigators. A person not associated with the study tabulated the results and interpretations. Agreement of interpretations was assessed with Fleiss' kappa statistic. RESULTS: Prolonged transit times affected sample quality for some laboratories. Overall agreement among investigators regarding the FC sample interpretation was strong (κ = 0.86 ± 0.19, P < .001), and for specific categories, ranged from moderate to perfect. Agreement of the lymphoproliferation or other leukocyte sample category from the analysis of the FCS files was weak (κ = 0.58 ± 0.05, P < .001). CONCLUSIONS: Lymphoproliferations were readily identified by FC, but identification of the categories of hematolymphoid neoplasia in fresh samples or archived files was variable. There is a need for a more standardized approach to maximize the enormous potential of FC in veterinary medicine.
BACKGROUND: Flow cytometry (FC) is used increasingly in veterinary medicine for further characterization of hematolymphoid cells. Guidelines for optimizing assay performance and interpretation of results are limited, and concordance of results across laboratories is unknown. OBJECTIVES: This study aimed to determine inter-investigator agreement on the interpretation of FC results from split samples analyzed in different laboratories using various protocols, cytometers, and software; and on the interpretation of archived FC standard (FCS) data files contributed by the different investigators. METHODS: This was a multicenter observational cross-sectional study. Anticoagulated blood or lymph node aspirate samples from nine client-owned dogs were aliquoted and shipped to participating laboratories. Samples were analyzed with individual laboratory-developed protocols. In addition, FCS files from a set of separate samples from 11 client-owned dogs were analyzed by participating investigators. A person not associated with the study tabulated the results and interpretations. Agreement of interpretations was assessed with Fleiss' kappa statistic. RESULTS: Prolonged transit times affected sample quality for some laboratories. Overall agreement among investigators regarding the FC sample interpretation was strong (κ = 0.86 ± 0.19, P < .001), and for specific categories, ranged from moderate to perfect. Agreement of the lymphoproliferation or other leukocyte sample category from the analysis of the FCS files was weak (κ = 0.58 ± 0.05, P < .001). CONCLUSIONS: Lymphoproliferations were readily identified by FC, but identification of the categories of hematolymphoid neoplasia in fresh samples or archived files was variable. There is a need for a more standardized approach to maximize the enormous potential of FC in veterinary medicine.