Quantificação de tecidos em imagens mamográficas por meio de histogramas

• Autor(a) principal: Souza, Rafael Toledo Fernandes de [UNESP]
• Data de Publicação: 2012
• Outros Autores: Pina, Diana Rodrigues de [UNESP], Alvarez, Matheus [UNESP], Velo, Alexandre França [UNESP], Oliveira, Marcela de [UNESP], Pavan, Ana Luiza Menegatti [UNESP], Miranda, José Ricardo de Arruda [UNESP]
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://www.rbfm.org.br/index.php/rbfm/article/view/214; http://hdl.handle.net/11449/140822
• topic: Mammography; Breast; Tissues; Algorithms; Mamografia; Mama; Tecidos; Algoritmos

Breast cancer is the main cause of death among women. The amount of fibroglandular tissue in breast is directly associated with the probability of breast cancer and also indicates the difficulty degree in in the detection of this disease. This study aims to quantify the breast tissues on mammographic planar images through an algorithm that uses the histogram of these images. We analyzed 10 breast exams (each exam containing an image on the Cranio-Caudal projection (CC) and the projection Middle Lateral Oblique (MLO)). For validation, three mammographic images were analyzed with the developed algorithm and compared with measurements carried out by segmenting images in the same breast Magnetic Resonance Images (MR). The variations found between the CC and MLO images were not significant by Bland and Altman test. For validation, the biggest difference between the results obtained by the algorithm and the segmentation of MR images for the fibroglandular tissue was 3.50%, showing a good agreement between methods, and thus showing that the algorithm developed satisfactorily quantifies the mammographic images.