Calculation of Breast Glandularity from Digital Mammography Using Spectrum Reconstruction and Standard Tissue Data

• Main Authors: LIN, YOU-LIN, 林侑霖
• Other Authors: PENG, QUAN
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/ymt4fy

碩士 === 東海大學 === 工業工程與經營資訊學系 === 106 === In the case of global breast cancer continues to rise, early detection and early treatment are the most effective measures to prevent breast cancer. At present, the most important breast cancer screening and diagnosis tool is digital mammography (DM), however, due to the overlapping of the tissues in the breast, the lesion is easily covered by the tissue and the detection efficiency is lowered. On the other hand, women with high breast density are more likely to develop cancer, and at the same time, the detection efficiency of breast cancer is greatly reduced. In this study, an imaginary breast prosthesis was made by using a printing material with similar attenuation characteristics as glandular tissue and lard as adipose tissue to simulate breast structure. Printed as a physical phantom through laminated manufacturing. The X-ray energy spectrum information of the mammography machine is obtained by using the Estimation Maximization (EM) and the transmission measurement information, and the breast ratio is calculated by the energy spectrum information. The percent error of the spectrum reconstructed is 0.84% for the Mo/Mo, 0.74% for the Rh/Rh. The percent error of the quantitative is 9.37% for Mo/Mo, 9.2% for Rh/Rh.