REPRODUCIBILITY AND FEASIBILITY OF CLASSIFICATION AND NATIONAL GUIDELINES FOR HISTOLOGICAL DIAGNOSIS OF CANINE MAMMARY GLAND LESIONS: A MULTI-INSTITUTIONAL RING STUDY.

A standardised histological diagnosis provides the basis for clinical treatment and management decisions but also contributes to monitoring cancer incidence, using cancer epidemiology to enlighten the causes of the disease, and framing the activity of cancer research. However, the accuracy of histological diagnosis is inadequately understood in veterinary oncology and often lacks standardisation and includes misleading terminology, causing variable/personal interpretation of histological criteria and resulting in non-uniform diagnosis and impossibility to compare studies and interchange data. The aim of the present work was to quantify interobserver agreement of canine mammary tumours (CMTs) among pathologists with different diagnostic experiences and belonging to different institutions when the same classification (including application of ICD-O codes), grading systems, and guidelines were applied. A group of 15 pathologists reviewed the recently published Davis-Thompson Foundation classification for mammary lesions, identified critical points and agreed on guidelines to overcome these points. The guidelines were applied in a blinded ring test on 36 CMTs. The interobserver agreement, feasibility, and reproducibility of guidelines, were statistically assessed by Kappa analysis. The overall concordance rate of diagnostic interpretations of participating pathologists expressed in terms of identification of Hyperplasia-Dysplasia vs Benign vs Malignant lesions showed a substantial agreement (average k ranging from 0.66 to 0.82, with a k-combined of 0.76). Instead, outcomes expressed in terms of ICD-O morphological code/diagnosis of histotype had only a moderate agreement (average k ranging from 0.44 and 0.64, with a k combined of 0.54). Results were encouraging, demonstrating that internationally established classifications and consensus panel guidelines can produce moderate to substantial agreement, however additional efforts are needed to further reduce interobserver variability in CMTs diagnosis, grading and classification.