Font Size: a A A

Computer Aided Diagnosis for Cone Beam Breast CT

Posted on:2012-05-28Degree:Ph.DType:Thesis
University:University of RochesterCandidate:Zhang, XiaohuaFull Text:PDF
GTID:2464390011968364Subject:Engineering
Abstract/Summary:
Cone Beam Breast Computed Tomography (CBBCT) has been developed as a technology dedicated to breast imaging. The system has relatively low radiation dose, high contrast resolution, high contrast-noise ratio, and isotropic high resolution. The CBBCT imaging field is large enough to cover the whole breast in one scan. Without tissue overlap, the breast's anatomic structure is clearly revealed in CBBCT reconstruction images with precise three-dimensional (3D) spatial distribution and morphological characteristics. Compared with mammograms and breast MRI, CBBCT images are able to provide more information to the radiologists. The accuracy of breast lesion diagnosis can potentially be improved with CBBCT images.;With high resolution 3D visualization, CBBCT images usually have a huge volume of data. Reviewing the data, locating all possible lesions, and making a correct diagnosis with CBBCT images may be a time consuming task for radiologists. Radiologists' accuracy, consistency, and efficiency may decrease when continuously reviewing the CBBCT volume data. In this thesis, a CBBCT-based Computer Aided Diagnosis (CAD) system is proposed and developed to serve as a "second reader" of the CBBCT volume data. With algorithms designed to take advantage of the CBBCT characteristics and reveal possible lesions with supplementary information, CBBCT CAD can help radiologists to further increase their diagnostic accuracy and efficiency. In this thesis five major components of the CBBCT CAD system have been designed based on the diagnostic requirements: 1. A 3D noise reduction scheme to improve the visualization of the 3D CBBCT image volume; 2. A 3D auto-detection scheme to detect calcifications in CBBCT images; 3. A 3D auto-detection scheme to detect suspicious masses in CBBCT images; 4. A mastectomy specimen experiment to validate the feasibility of CBBCT ductography; and 5. A 3D auto-segmentation scheme to evaluate the breast density based on CBBCT 3D images.;The experimental results of CBBCT CAD are presented and future work is planned.
Keywords/Search Tags:CBBCT, Breast, Diagnosis
Related items