Toward Better Utilization of Medical Images

An area that this research will investigate is to improve the detection and classification of breast tumours. Currently, mammography is the widely used tool to examine and screen breast tumours. However, mammography has several disadvantages. For example, it uses X-ray radiation, which may result in harmful effects on breast tissues over prolonged use. Moreover, mammograms sensitivity is very low when used to screen young women or women with dense breasts. On the contrary, ultrasound imaging has no known side effects (neither short-term nor long-term),. In addition, ultrasound imaging is less expensive than other screening systems. This research will investigate several approaches and techniques to combine various ultrasound image data (e.g., B-Mode and Doppler). The objective is to identify and differentiate between benign and malignant tumours.

Another area of investigation is to improve the detection of carotid artery plaque in three-dimensional images. It is believed that this type of plaque is caused by Atherosclerosis. Currently, most plaque detection is done using two-dimensional ultrasound images. Yet, three-dimensional ultrasound images, which can provide vessel and plaque volumes, are available, and are just as cost effective as their two-dimensional cousins. Since three-dimensional ultrasound will surely be the future of clinical trials, three-dimensional automatic segmentation will give an accurate measurement of plaque levels.

This research will develop several customized image-processing techniques including: image enhancement techniques (to reduce the effect of noise in images), image registration techniques (to combine the multi-modal data), and image segmentation/classification techniques (to reduce the radiologist's interaction time with the CAD system).