IFEG   20353
INSTITUTO DE FISICA ENRIQUE GAVIOLA
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Implementation of several mathematical algorithms to breast tissue density classification
Autor/es:
C. QUINTANA; M. REDONDO; G. TIRAO
Lugar:
Rio de Janeiro
Reunión:
Congreso; 12th International Symposium on Radiation Physics – ISRP; 2012
Institución organizadora:
International Radiation Physics Society
Resumen:
The accuracy of mammographic abnormality detection methods is strongly dependent on the breast tissue characteristics, where a dense breast tissue can hide lesions causing cancer to be detected at later stages. In addition, breast tissue density is widely accepted to be an important risk indicator for the development of breast cancer. This paper present the implementation and the performance of different mathematical algorithms designed to standardizethe categorization of mammographic images, according to the American College of Radiology classifications. These mathematical techniques are based on intrinsic properties calculations and on comparison with an ideal homogeneous image (joint entropy, mutual information, normalized cross correlation and index Q) as categorization parameters. The algorithms evaluation was performed on 100 cases of the mammographic data sets provided by the Breast Cancer Prevention Program of the Health Ministry of Córdoba, Argentina. The obtained breast classifications were compared with the expert medical diagnostic, showing a good performance. The implemented algorithms revealed a high potentiality to breast density classification, which would be proven over an extended data set to obtain a more general and reliable validation, according to breast density classification performed by several medical and radiologist expert.