Fractal dimension of canine mammary gland epithelial tumors on cytologic smears

Abstract

Fractal geometry is a tool that can be used for describing, modeling, analyzing, and processing irregular and complex figures. Past investigations in medicine have revealed that fractal analysis could also be applied in tumor pathology to characterize irregular boundaries of the nuclei of tumor cells. The aim of this study was to define whether the fractal dimension parameter could be used on cytologic specimens to differentiate benign from malignant canine mammary gland epithelial tumors. The fractal dimension of nuclear surface was determined by computer-assisted morphometry on Hemacolor-stained cytologic smears obtained by fine needle aspiration of normal canine mammary gland epithelial cells, and cells in mammary adenomas, tubulopapillary carcinomas, solid carcinomas, and anaplastic carcinomas. Data were analyzed by Mann-Whitney U test. Significant differences (P <.001) were observed in mean fractal dimension among all tumor types and in comparison with normal canine mammary gland epithelial cells (except for the fractal dimension between solid carcinomas and anaplastic carcinomas). The morphometric parameter, fractal dimension, could help in the diagnostic discrimination between benign and malignant canine mammary gland epithelial tumors on cytologic specimens.