Quantificação de fibras em imagens de cortes histológicos de tecido mamário de gatas

Abstract

Mammary carcinoma is the third that most affects the feline species, and represents a large part of mammary neoplasms in cats. The risk factors for breast cancer in women and mammary carcinoma in cats are, and, like cats, a great model for comparative studies with women. However, studies of tumors in feline mammary tissues using digital image processing are scarce. One of the factors that alter breast tissue cells and, consequently, contribute to the increase in breast cancer is the increase in breast stroma density. This is due to the uncontrolled deposition of collagen fibers in these regions. Among them, there are type I and type III collagen fibers. One way to differentiate as red fibers is through the method of a feline histological section is stained and, through a polarized tissue, as type I fibers seen red and, as type III, red tons of green. An image bank of histological sections was then analyzed, classified by histological grade (control, Grade I, Grade II and Grade III) and by location (center or periphery). Using digital image processing techniques, this reliable work quantified the proportions of red and green in the images, by determining a differentiation method, validated by an expert, through three different estimates. It was possible to quantify the green and used regions of the bank images, which may contribute to the comparative models between breast cancer and breast cancer, and, consequently, to the diagnosis of these tumors.