Clinicians usually assess the three-dimensional (3-D) characteristics of the cardiac anatomy and function by mentally reassembling data acquired from different modalities using the two-dimensional (2-D) display format such as coronary angiography, echocardiography, computed tomography (CT), magnetic resonance (MR) imaging, or radionuclide tomography. Our major goal in the proposed research is to develop computer-based techniques for 3-D reconstruction and integration using complementary imaging modalities to facilitate diagnostic and interventional procedures. The initial effort will be concentrated on the 3-D reconstruction and correlation of coronary vascularture of human heart based on x-ray angiography and intravascular ultrasonography (IVUS). Specifically, the spatial 3-D geometry of coronary arteries and the morphology of coronary disease will be generated from respective routine angiograms and IVUS cross-sectional images to facilitate diagnostic and therapeutic cardiac catheterization by (i) suggesting optimal views with minimal vessel overlap and foreshortening, (ii) comprehending the composition, morphology, and distribution of coronary atherosclerotic plaques, and (iii) optimizing input for quantitative coronary analysis and interventional device selection. In a single patient examination, multiple coronary angiograms will be acquired during several cardiac cycles either by using a single-plane or biplane imaging system. The initial 3-D coronary arterial tree will be reconstructed based on an initial pair of angiograms. With the reconstructed 3-D patient-specific coronary arterial tree model, optimal views that minimize vessel overlap and foreshortening with respect to a selected arterial segment or a bifurcation will be predicted and employed for subsequent image acquisition to continue the diagnostic and therapeutic procedure. The initially reconstructed 3-D coronary arterial tree model can then be refined by using the subsequently acquired angiograms. The calculated optimal views will be utilized to facilitate IVUS examination for advancing the IVUS catheter. The 3-D vascular lumen and plaques will be reconstructed from a sequence of IVUS cross-sectional images acquired by means of a systematic, timed pull-back of the ultrasound catheter through the respective arterial segments. The reconstructed 3-D coronary arterial tree and vascular plaque will then be correlated and integrated. The cardiologist can then incorporate the prediction of optimal views and the well-defined morphology and composition of coronary disease to continue the study during the fluoroscopic or ultrasound based interventional procedure. With the 3-D coronary processing techniques, we expect that the prospective clinical study can accomplish: a) reduction in radiation exposure and contrast volume, b) reduction of interventional procedure time by initially optimizing visualization and quantification of stenosis, and c) accurate assessment of coronary vascular pathology and histology for guidance of interventional therapy to improve and provide safer health care.
