| Summary: | 碩士 === 南台科技大學 === 電機工程系 === 100 === As ultrasound can clearly distinguish between breast cysts and tumors in real terms, it is also without radiation concerned, easy to be used and with lower cost advantage. Therefore, there are many doctors in Taiwan used the ultrasound as the diagnosis of breast lesions in the first line. Recently, the better resolution of the elastic image of breast, the elastic image of breast is a very active field for research. However, it still cannot integrate the conventional B-mode image with the elastic image in the same picture for commercial instrument. In this research, a computer assistant system (CAS) developed on MatLab environment was designed as a platform for semi-automatic loading of the elasticity imaging and the clinical B mode image of breast simultaneously. Then, the investigation of the quantitative parameters such as area ratio or compressed ratio used to distinguish the benign or the malignant tumors was performed within this system.
The development CAS was designed as an initial diagnosis system of breast lesion under the MatLAB software environment. First, a retrospective study from the Siemens ultrasound-based breast imaging system in Chi-Mei Medical center was used to investigate. The scope of this study only includes the analysis of B-mode and elasticity image. A total of 30 patients with 60 data were analyzed in this study. The areas belonged B-mode image and breast elasticity image both pre compression and post compression were automatic computed based on the manual selection of clinician. Then the quantitative parameters of area ratio were compared with the diagnosis results of biopsy and were analyzed with multi-variable static analysis by SPSS (v.13) in final.
This study shows that the use of the proposed quantitative parameters, especially the area ratio (pre-compression/post- compression) of breast tissue, indicated very good statistical significance (p<0.05).
|
|---|
