Detection of Lung Cancer via Registered Temporal Images
DESCRIPTION (provided by applicant): Phase I
Effective registration functions of temporal digital images for many radiographic applications have been long over due. In the first phase of this project, we propose to develop key segmentation methods for automatic delineation of anatomic structures in both chest radiography and thoracic CT. It is our strategy to launch this important project to develop clinical tools for temporal chest image viewing and diagnosis for both thoracic CT and chest radiography. Even though thoracic CT will be used as a primary imaging device, chest radiography will most likely also be taken for its low-cost and high-resolution in the longitudinal direction. Technically speaking, it is also important to gain experience in segmenting and registering objects in 2D space before cultivating registration of lung structures in the 3-D space, particularly from the implementation aspects.We hypothesize that by eliminating the unchanged lung structure and/or by comparing the differences between the temporal images with the computer-aided system, the radiologist can more effectively detect the cancer in the lung field. We have done some preliminary studies and have appreciated the potential of this approach. Initially, we will put our main effort in segmentation of ribs in chest radiography and segmentation of lung structures including large and medium size bronchi and blood vessels in thoracic CT. Our specific aims in the Phase I program include: (1) Automatic extraction of large anatomic structures in chest imaging; (2) Accurate delineation of posterior ribs and anterior ribs in chest radiography; (3) Automatic slice matching for two CT scans; (4) Segmentation of large and medium sized tree structures in the thoracic CT; and (5) Evaluation of the computer segmentation results using radiologists' drawings and confirmation.We will then move into the registration of the segmented chest structures for matching and alignment between the temporal pair. Subtraction and visual inspection of the final image presentation will be carefully designed for clinical use in the second phase of this project.
Small Business Information at Submission:
CAELUM RESEARCH CORPORATION
CAELUM RESEARCH CORPORATION 1700 RESEARCH BLVD, STE 100 ROCKVILLE, MD 20850
Number of Employees: