Características Morfométricas de Núcleos Celulares para Diagnóstico e Prognóstico de Carcinomas Mamários em Cadelas

Abstract

For classification and prognostic purposes, histopathological evaluation is the method of choice for canine mammary carcinomas. However, it is not always clear and objective. To increase accuracy, molecular markers or a more detailed computational analysis of the cell nucleus can be used. The aim of this study was to identify which nuclear features are most relevant in determining prognosis and differentiating the main types of canine mammary carcinomas. A total of 74 cases of eight different types of mammary carcinomas in female dogs with known evolution were used. Using computer image analysis of slides stained with hematoxylin/eosin, several features related to size, shape, pleomorphism, and chromatin texture of the nucleus were evaluated. Canonical discriminant analyses were performed for prognosis and differential diagnosis, and the canonical functions obtained for prognosis in a single tumor type had a high hit rate. In general, features related to nucleus shape were the most relevant for prognosis, followed by features of size, pleomorphism, and chromatin texture. For the differential diagnosis of the main types of mammary carcinomas in bitches, features related to nucleus shape are the most relevant, while those related to nucleus size, pleomorphism and chromatin texture have similar relevance.