A computer aided diagnosis (CAD) method is proposed to assist radiologists in detecting pulmonary cancer in thoracic computed tomography (CT). Specifically, Phase I's aim is to assist radiologists (1) to compare current and previous thoracic CT scans to identify new findings or to assess the effects of treatments on lung cancer; (2) to interpret the same CT image containing two anatomic structures (lung and mediastinum) rapidly and efficiently; (3) to identify the malignancy of nodules by determining and removing vessels and bronchi connecting with nodule suspects; and (4) to provide other CAD functions to enhance the display capability for diagnosis of CT images. In this project, we will (1) collect numerous CT images, (2) develop sequence matching method between two CT scans, (3) develop dual window-level function to facilitate viewing, (4) develop 3-D tree construction method to identify the malignancy of lesions, and (5) develop additional functions to enhance the CAD-specific display capabilities Our R&D work will extend our existing CAD technology for lung cancer detection from chest x-ray images to CT images and introduce CAD-specific display system in the detection and diagnosis of lung cancer. The success of this project will eventually lead to an efficient, cost-effective, robust, and user-friendly system improving the accuracy and speed of lung cancer detection.
A computer aided diagnosis method will be developed in order to assist radiologists efficiently in the detection of lung cancer. This method will enhance the current patient care system. The proposed R & D will also provide an effective tool to improve the diagnosis of other diseases by using CT imaging.
