Angiography
 

1. Computerized Analysis of Stenotic Lesions   We developed an iterative deconvolution technique to determine the sizes of “blurred” vessels in angiographic images by taking into account the unsharpness of the imaging system. A selected vessel segment was tracked and the centerline was determined by fitting the tracking points with a polynomial curve. The non-uniform background in the vicinity of the vessel was estimated by a 2D surface and subtracted from the original image to yield a DSA-like image. The blurred image profile was then obtained from pixel values across the vessel in a direction perpendicular to the center line. This image profile was compared iteratively with calculated profiles for vessels of various sizes, which were obtained by convolving a cylindrical vessel model with the line spread function, until the root-mean-square difference between the two profiles was minimized. The size of a cylindrical vessel yielding the matched profile was considered the best estimate of the unknown vessel size. Studies with a blood vessel phantom indicated that vessels larger than 0.5 mm could be measured with an accuracy and precision of approximately 0.1 mm, which was about 1/3 of the pixel size of our system. (218, 220, 232, 264)
2. Automated Tracking of Vessels Angiographic images can be analyzed by means of computer-aided methods to provide diagnostically useful information, such as the degree of stenosis, the extent of cardiac wall motion, and blood flow parameters. In addition, three-dimensional representations of vascular trees can be generated from biplane angiograms or from stereoangiograms for use in treatment planning. The diagnostic information could be obtained more efficiently and reproducibly if these analyses were automated. For automation, however, a reliable method of computerized detection and identification of the vascular structures in angiographic images is required. We developed an automated vessel-tracking method based on the double-square-box region-of-search technique, for efficient tracking of the connected vascular tree in an image. Tracking points and branch vessels were located by searching of the perimeter of boxes, which were centered on previously determined tracking points. The tracking points were verified as connected by means of region growing. In relatively straight regions of vessels, a large box was employed for efficient tracking; in curved regions of vessels, a small box was employed to ensure accurate tracking. When tracking was completed, accurate vessel information (i.e., the vessel position, size, and contrast determined at each tracking point) was available for further quantitative analysis. (204, 224, 239, 283, 302, 689)
3. Determination of Instantaneous and Average Blood Flow Rates We developed a new method for quantitation of blood flow rates based on determination of the spatial shift of the distribution of the contrast material in opacified vessels in angiographic images that were acquired while a bolus of the contrast material proceeded through the vessel. The distance that the contrast material traveled during the time between two image acquisitions was determined by comparison of “distance-density” curves, which represent the distribution of the contrast material along the length of the vessel in the respective images. The flow rate between image acquisitions was calculated by multiplication of the traveled distance by the frame rate and the vessel cross-sectional area. Therefore, for high-frame-rate acquisitions, “instantaneous” blood flow rates can be determined. For vessel-phantom studies in which angiograms were obtained at 15 frames/sec, the instantaneous flow rates measured with our technique agreed with those measured with an electromagnetic method to within an average of 2.3 cc/sec for pulsatile flow conditions with peak flow rates of up to 20 cc/sec; average flow rates agreed to within an average of 11%. For pulsatile-flow conditions, the accuracy of this new technique surpassed that of conventional methods in which the time shift of the distribution (based on the so-called “time-density” curves) was analyzed. (254, 280, 312, 344)
4. Determination of 3D Vasculature from Biplane Views   We developed a method for determination of the 3D vasculature based on analysis of biplane images of the coronary arteries. After image acquisition, the centerlines of major coronary arteries were identified in the images. The centerlines were used to determine bifurcation points in each of the images. The vessel hierarchy was then employed to identify the corresponding bifurcation points in the biplane image pair. The corresponding image points were input to a modification of the Metz-Fencil technique by which the imaging geometry was determined. Note that this technique does not require a calibration object. After determination of the imaging geometry, corresponding points along the identified vessel centerlines in both views were defined by means of epipolar lines. From the corresponding points and the calculated imaging geometry the 3D vessel centerlines were generated and the vascular tree rendered. Simulation and phantom studies indicated that the geometry can be determined with an accuracy of approximately 1 degree, and the 3D positions can be determined to within 0.04, 0.07, and 3.0 mm along the x, y, and z directions, respectively. Rendered versions of the 3D vascular tree appeared to correspond well with both the projections used and those not used for the calculations. (258, 272, 405, 444, 452, 483, 497, 498, 561, 562, 661)
5. Automated Computerized Scheme for Detection of Unruptured Intracranial Aneurysms in Three-Dimensional MRA We developed a computerized scheme for automated detection of unruptured intracranial aneurysms in magnetic resonance angiography (MRA) based on the use of 3D selective enhancement filter for dots (aneurysms).  Twenty-nine cases with 36 unruptured aneurysms (diameter: 3 - 26 mm, mean of 6.6 mm) and 31 non-aneurysm cases were used in this study. The isotropic 3D MRA images with 400 x 400 x 128 voxels (a voxel size of 0.5 mm) were processed by use of the selective enhancement filter. The initial candidates were identified by use of a multiple gray-level thresholding technique on the dot-enhanced images and a region-growing technique with monitoring some image features. All candidates were classified into four types of candidates according to the size and local structures based on the effective diameter and the skeleton image of each candidate, i.e., large candidates and three types of small candidates including short-branch type, single-vessel type, and bifurcation type. In each group, a number of false positives were removed by use of different rules on localized image features related to gray levels and morphology. Linear discriminant analysis was employed for further removal of false positives. With our computer-aided diagnostic (CAD) scheme, all of 36 aneurysms were correctly detected with 2.4 false positives per patient based on a leave-one-out-by-patient test method. Our CAD system would be useful in assisting radiologists for detection of intracranial aneurysms in MRA. (901, 902)