Development of techniques and methods for evaluation of quality of scanned image in mammography

Cancer is the second cause of death in the Brazilian female population and breast cancer is the most frequent neoplasm amongst women. Mammography is an essential tool for diagnosis and early detection of this disease. In order to be effective, the mammography must be of good quality. The Brazilian College of Radiology (CBR), the National Agency for Health Surveillance (ANVISA) and international bodies recommend standards of practice for mammography. Due to the risk of ionizing radiation, techniques that minimize dose and optimize image quality are essential to ensure that all women are submitted to mammography procedures of high quality for the detection of breast cancer. In this research were analyzed components of the image treatment via digital and developed methods and techniques of analysis aiming the detection of structures for medical diagnosis, decreasing variations due to subjectivity. It used free software Image J, to make the evaluations of the information contained in the scanned images. We use the scanned images of calibration of a simulated breast to calibrate the program Image J. Thus, it was able to correctly convert the values of the scale of shades of gray in optical density values of presenting the standard deviation for each measure held. Applying the test t-student noticed that the values obtained with the digital system to the level of contrast and spatial resolution are consistent with the results obtained so subjective, since there was no significant difference (p <0.05) for all comparisons evaluated. Since then, this methodology is recommended in routine evaluations of services of mammography.