Do we agree on counting mitotic figures in cytology of canine lymphoma? Three field selection methodologies by three...

Mitotic count is a key parameter of the cytological evaluation of canine lymphomas (CLs), but consensual guidelines for counting are missing. Variations in cell preservation are common in cytological smears and this can compromise the reproducibility of mitotic counting. We compared three methodologies for selecting fields for mitotic counting in CL cytological smears and interobserver agreement was assessed. Mitotic counting was performed in 19 CLs by three observers in five high-power fields (HPFs), selected according to the following methodologies: (A) randomly and with well-preserved cells; (B) adjacent fields within the best-preserved monolayer area; and (C) in the area with the highest mitotic activity (hot spots). Counts were analysed both as the total number in five HPFs and as number per mm²; using published cut-off values, mitotic counts were classified as low, intermediate or high. Counts by methods A and B were comparable, while consistently higher values were obtained with method C. Overall, the mitotic count classification shifted in 21% of the cases, with a higher assignment of cases to the high category using method C. Interobserver agreement was poor between observers using microscopes with different field of view diameters. When counts were normalized to number per mm², the agreement improved. The methodology of counting in adjacent fields showed better interobserver agreement compared to random fields. However, there was a high variability in the recorded numbers by each observer in highly proliferative CLs. The data highlights that mitotic counts in cytological smears of CL are highly dependent on methodology for selection of the fields and on the microscope.