Coronary artery disease (CAD) continues to be the leading cause of death in America. Coronary arteriography is the standard for diagnosing the anatomical basis of CAD: however, this invasive test has a non-negligible procedural risk and significant cost. Moreover, ~30% of coronary arteriograms show no demonstrable coronary artery stenosis, and only 1/3 are performed in conjunction with an interventional procedure. Non-invasive evaluation of coronary anatomy with multislice spiral CT coronary arteriography (CTCA) could potentially replace these diagnostic angiograms, but has limitations including image artifacts related to coronary arterial calcification and coronary stenting. Non-invasive ECG-gated myocardial perfusion imaging (MPI) provides physiologic information and is the standard for the assessment of hypoperfusion resulting from CAD;however, it also has limitations related to its measurement of flow being relative rather than absolute. We hypothesize that explicit integration of the anatomic information from CTCA and the physiologic information from MPI can ameliorate the limitations of each, improve assessment of CAD, and eliminate many diagnostic catheterizations, reserving invasive angiography primarily for interventional use. Our long term objective is to develop and validate computer-based methods to fuse, render and quantify complementary CTCA+MPI cardiac imagery. Our primary hypothesis is that 3D fusion and rendering of cardiac information from these complementary imaging modalities is more accurate and efficient in diagnosing CAD than conventional approaches of independently viewing each study.
Our specific aims are to develop new methods for fusing and rendering CTCAs and MPIs that take advantage of all relevant information available in both images. The new approach segments the ventricular surfaces from the CTCA using their expected shape and intensity, and then incorporates a mechanical model to nonlinearly fuse the MPI to these surfaces. The arteries segmented from the CTCA will thus be in alignment with the transformed MPI data. Quantitative information about arterial anatomy and myocardial perfusion will be combined and displayed using surface rendered 3D graphics. This approach will be entirely automatic and much more accurate than existing techniques. All work will be validated with animal and patient studies to show both accuracy of fusion and the efficacy of reading the fused display vs. separate CTCA and MPI slices, and vs. standard angiograms.
Showing the most recent 10 out of 16 publications
