Cytomorphological description and intra-observer agreement in whole slide imaging for canine lymphoma

• Published in: The veterinary journal (1997) Vol. 236; pp. 96 - 101
• Main Authors: Bonsembiante, F., Martini, V., Bonfanti, U., Casarin, G., Trez, D., Gelain, M.E.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-06-2018
• Subjects: Animals; Canine; Cytology; Digital pathology; Dog Diseases - diagnostic imaging; Dogs; Image Interpretation, Computer-Assisted; Lymph Nodes - cytology; Lymph Nodes - ultrastructure; Lymphoma; Lymphoma - diagnostic imaging; Lymphoma - ultrastructure; Lymphoma - veterinary; Microscopy - instrumentation; Microscopy - methods; Microscopy - veterinary; Observer Variation; Pathology, Clinical; Whole slide imaging; Digital pathology; Whole slide imaging; Lymphoma; Cytology; Canine
• ISSN: 1090-0233; 1532-2971
• DOI: 10.1016/j.tvjl.2018.04.020

•A whole slide imaging scanner is a robotic microscope able to digitise a glass slide to a composite digital image.•The cytomorphological capability of a whole slide imaging scanner was evaluated on canine lymphoma cytology samples.•There are some technical limitations that affected the diagnostic power of the instrument.
Whole slide imaging (WSI) uses robotic microscopes for computerising entire slides into digital images. The aim of this study was to assess the agreement between WSI and optical microscopy for evaluating canine lymphoma cytological samples. Forty-four slides were computerised using a WSI scanner and the digital and glass slides were examined by three observers with different levels of expertise. Morphology and grade of lymphoma were scored on the basis of the updated Kiel classification and intra-observer agreement was assessed. The accuracy of determining the grade of lymphoma with digital and glass slides based on the results of flow cytometry (FC) was established. The overall intra-observer agreement for cytomorphological features was fair to moderate (κ=0.34–0.52) for the three observers and moderate (κ=0.44–0.53) for the evaluation of grade of malignancy. The diagnostic agreement between FC and digital slides was slight (κ=0.16) for the inexperienced observer, fair (κ=0.32) for the mildly experienced observer and moderate (κ=0.50) for the very experienced observer. The diagnostic agreement between FC and glass slides was fair (κ=0.37) for the inexperienced observer, substantial (κ=0.63) for the mildly experienced observer and moderate (κ=0.50) for the very experienced observer. These findings underline the importance of observer experience in determining the grade of malignancy, especially if digital slides are used. The study also identifies some technical limitations of the WSI scanner used in this study, mainly linked to image quality, which might affect the morphological evaluation of neoplastic cells.