On-Line 3D Vasculature From Biplane Angiograms

DESCRIPTION (provided by applicant): The goal of the proposed research is to develop automated methods for accurate and reliable quantitative vascular measurements (vessel diameters, cross sections, and distances) in clinical settings. The developed technology will not only provide the cardiologist with immediate quantitative information as well as the means for longitudinal patient assessment, but may also be the basis for more extensive and cross-modality patient evaluations. Under the previously funded research, we developed a system for accurate and reproducible determination of 3D vascular tree centerlines from biplane image pairs within 5 minutes of initiation of analysis. In this continuation proposal, we will not only improve the centerline accuracy, but we will also develop new techniques (a) for fully automatic centerline determination, (b) for lumen reconstruction using breakthroughs developed in our lab, and (c) for alignment and comparison with IVUS data with speeds sufficient to provide clinically critical information during interventions. The studies will extend beyond evaluations of the accuracy and reliability of the measured quantities to include evaluations of impact of these data on clinical decisions during catheterization procedures in pig models and in patients. With the increasing complexity of cardiovascular interventions and the need for accurate characterization of the vessel lumen in clinical studies for cross modafity and longitudinal comparisons, we see this research as one of the critical next steps in vascular analysis. As a result, we are convinced that the proposed research will result in more accurate and more comprehensive vascular data, safer angiographic studies, better longitudinal tracking of patients, quantitative evaluations of progression and regression of stenosis, and subsequent outcome analysis.