Cytology of Feline Nodal Lymphoma: Low Interobserver Agreement and Variable Accuracy in Immunophenotype Prediction

J Comp Pathol. 2021 Apr:184:1-6. doi: 10.1016/j.jcpa.2021.01.007. Epub 2021 Feb 22.

Abstract

Nodal lymphomas are less common in cats than in dogs and, consequently, no specific studies have been published. Cytology is the first step in the diagnosis of nodal lymphoma but is highly subjective. Morphological features have been introduced for the cytological classification of canine lymphomas but not for cats. Therefore, the aim of this study was to evaluate interobserver agreement on various cytological features of feline nodal lymphomas and to investigate the accuracy in predicting B or T immunophenotypes. Four veterinary cytologists examined 25 feline nodal and mediastinal lymphoma cytological samples by adapting the criteria used for the evaluation of canine lymphomas and setting histopathology and immunohistochemistry as the gold standard. High interobserver variability was found in the evaluation of most features except for the presence or absence of cytoplasmic vacuoles, which were more common in B-cell lymphomas. Cytology training centre was the major factor influencing the extent of agreement among evaluators. Diagnostic accuracy in predicting lymphoma immunophenotype varied from 35% to 75% and did not appear to be correlated with the experience of the evaluators. We conclude that cytological criteria, commonly used to describe canine lymphomas, are not adaptable to the counterpart feline neoplasms. Cytology-based immunophenotyping of feline lymphomas from different laboratories, and different cytologists within the same laboratory, differ substantially and should not be considered reliable. Specific cytological criteria are needed to describe feline lymphoma.

Keywords: accuracy; cytology; feline nodal lymphoma; interobserver agreement.

MeSH terms

  • Animals
  • Cat Diseases*
  • Cats
  • Cytodiagnosis / veterinary
  • Immunophenotyping / veterinary
  • Lymphoma* / veterinary
  • Observer Variation