Automated identification of vessel contours in coronary arteriograms by an adaptive tracking algorithm

A viewpoint determination system for stenosis diagnosis and quantification in coronary angiographic image acquisition

This paper describes the usefulness of computer assistance in the acquisition of "good" images for stenosis diagnosis and quantification in coronary angiography. The system recommends the optimal viewpoints from which stenotic lesions can be observed clearly based on images obtained from initial viewpoints. First, the viewpoint dependency of the apparent severity of a stenotic lesion is experimentally analyzed using software phantoms in order to show the seriousness of the problem. The implementation of the viewpoint determination system is then described. The system provides good user-interactive tools for the semiautomated estimation of the orientation and diameter of stenotic segments and the three-dimensional (3-D) reconstruction of vessel structures. Using these tools, viewpoints that will not give rise to foreshortening and vessel overlap can be efficiently determined. Experiments using real coronary angiograms show the system to be capable of the reliable diagnosis and quantification of stenosis.

An algorithm for real-time vessel enhancement and detection

In this paper we present an algorithm for the real-time enhancement and detection of blood vessels in medical images. The algorithm is based on a set of linear filters sensitive to vessels of different orientation and thickness. Such filters are obtained as linear combinations of properly shifted Gaussian kernels. The output of multiple oriented vessel-enhancing filters can be integrated to obtain images in which vessels are highly enhanced independently of their direction and thickness. To avoid spurious responses in the presence of step edges or to enhance the skeletons of vessels, the output of directional filters can be validated before integration. Skeleton detection and vessel segmentation can be performed via thresholding with hysteresis. Experimental results on synthetic images and real coronary arteriograms are reported.

Shape recovery and volume calculation from biplane angiography in the stereotactic radiosurgical treatment of arteriovenous malformations

PURPOSE

A model for calculating the three-dimensional volume of arteriovenous malformations from biplane angiography.

METHODS AND MATERIAL

Three-dimensional (3D) volume reconstruction is easily feasible with axial, coronal, or sagittal computer tomography (CT) and nuclear magnetic resonance (NMR) scans. On the other hand, radiosurgical treatment of arteriovenous malformations (AVM) is exclusively based on two orthogonal stereotactic projections, obtained with angiographic procedures. Most commonly, AVM volumes have been calculated by assimilating the nidus volume to a prolate ellipsoid. We present an algorithm dedicated to 3D structure reconstruction starting from two orthogonal stereotactic projections. This has been achieved using a heuristic approach, which has been widely adopted in the artificial intelligence domain.

RESULTS

Tests on phantom of different complexity have shown excellent results.

CONCLUSION

The importance of the algorithm is considerable. As a matter of fact: (a) it allows calculations of complex structures far away from regular ellipsoid; (b) it permits shape recovery; (c) it provides AVM visualization on axial planes.

Extraction of line properties based on direction fields

The authors present a new set of algorithms for segmenting lines, mainly blood vessels in X-ray images, and extracting properties such as their intensities, diameters, and center lines. The authors developed a tracking algorithm that checks rules taking the properties of vessels into account. The tools even detect veins, arteries, or catheters of two pixels in diameter and with poor contrast. Compared with other algorithms, such as the Canny line detector or anisotropic diffusion, the authors extract a smoother and connected vessel tree without artifacts in the image background. As the tools depend on common intermediate results, they are very fast when used together. The authors' results will support the 3-D reconstruction of the vessel tree from stereoscopic projections. Moreover, the authors make use of their line intensity measure for enhancing and improving the visibility of vessels in 3-D X-ray images. The processed images are intended to support radiologists in diagnosis, radiation therapy planning, and surgical planning. Radiologists verified the improved quality of the processed images and the enhanced visibility of relevant details, particularly fine blood vessels.

Three-dimensional reconstruction of biliary tract from biplane projections

Over the past few years, very little has been done in the area of 3-D reconstruction of the bile duct. Since the system in use for 3-D visualization of the biliary tree is built by surgical or autopsy materials, it generally cannot be applied to clinical diagnosis. In this paper, an algorithm for three-dimensional reconstruction of the biliary tree from two mutually orthogonal is presented to provide accurate and reproducible 3-D information of the biliary tree structure. It has been proven to be useful in diagnosis prior to operation or non-surgical treatment, particularly, obstructive stones can be visualized by using a transparency technique. As experiments demonstrated, the proposed method can be used as a useful tool for the visualization of 3-D structure of biliary tree in clinical applications.

A fully automated identification of coronary borders from the tree structure of coronary angiograms

The accurate assessment of variations in coronary arterial dimensions plays an important role in the evaluation of ischemic heart disease and the effectiveness of treatment. Although there exist a variety of edge detection algorithms in the literature, most of them are human interactive and may provide a poor estimate on coronary lesion. In this paper, we present a new method for automatic identification of arterial borders. The proposed algorithm makes use of mathematical morphology to segment blood vessels which follow a tree structure, based on a priori knowledge of coronary anatomy. Finally, an adaptive tracking strategy is applied to automatically identify 2-D arterial borders along both sides of the vessels. This is accomplished by using an edge detection model at a branching point, matched filters, and the tree structure of the coronary artery. Experimental results show that our approach not only is insensitive to the intensity variations of background and noise, but also can extract the boundary of the coronary artery accurately.

Directional low-pass filtering for improved accuracy and reproducibility of stenosis quantification in coronary arteriograms

Considers the quantification of percent diameter stenosis in digital coronary arteriograms of low spatial resolution. To improve accuracy and reproducibility an edge-preserving smoothing method, called the directional low-pass filter (DLF), was developed to suppress quantum noise by averaging image intensity in a direction parallel to the vessel border. Accuracy of stenosis quantification was assessed by using stenosis phantoms. The standard error of the estimate (SEE) was 0.76 pixel-length (p) without spatial filtering and further reduced to 0.50 p by DLF; the average deviation as a measure of the regularity of border definition was also reduced by DLF from 1.00 to 0.68 p (n=50, P<0.001). It was shown that the DLF outperformed the conventional moving average filter and median filter. Reproducibility in terms of intraframe variability was assessed by using coronary arteriograms obtained from 10 patients. Intraframe variability of the percent stenosis measurements was reduced from 3.5% to 2.9% by DLF (n=10, P<0.005). An analysis of variance showed, however, that the interframe variability cannot be reduced by any of the spatial filters under investigation. The result of this study has provided a guideline for angiographically based quantification of percent stenosis under limited imaging resolution and suggests a new method for improving accuracy and reproducibility by directional low-pass filtering.

A nonsmoothing approach to the estimation of vessel contours in angiograms

Accurate and fully automatic assessment of vessel (stenoses) dimensions in angiographic images has been sought as a diagnostic tool, in particular for coronary heart disease. Here, the authors propose a new technique to estimate vessel borders in angiographic images, a necessary first step of any automatic analysis system. Unlike in previous approaches, the obtained edge estimates are not artificially smoothed; this is extremely important since quantitative analysis is the goal. Another important feature of the proposed technique is that no constant background is assumed, this making it well suited for nonsubtracted angiograms. The key aspect of the authors' approach is that continuity/smoothness constraints are not used to modify the estimates directly derived from the image (which would introduce distortion) but rather to elect (without modifying) candidate estimates. Robustness against unknown background is provided by the use a morphological edge operator, instead of some linear operator (such as a matched filter) which has to assume known background and known vessel shape.

A new method for automatic identification of coronary arteries in standard biplane angiograms

A new computerized method for analyzing biplane coronary arteriograms was proposed. The method extracted the vessels in an image, detected branching points, and identified vessel segments according to their anatomical names. The method was applied to a pair of orthogonal projection images acquired in the right anterior oblique and left anterior oblique views. It recognized the corresponding vessel segments in the images and labeled them correctly by their anatomical names. For diameter measurement, a combined first and second derivative filter process was applied to the images along the skeleton lines of the extracted vessel segments. We also designed a method of representation in which vessel diameters measured in the two projection images could be displayed in a single image.

A gray-level thinning method for delineation and representation of arteries

The quantification of three dimensional (3D) properties of coronary arteries is of significant importance. The performance of the 3D analysis is critically based on low-level representation of the arterial tree for different projections. A skeletal representation of arteries can provide appropriate data structure for registration of multiple angiographic projections and it can be further utilized for 3D reconstruction of the arterial tree. This paper presents an automated method for extracting the skeletal points of an arterial tree directly from the gray-level information without determining the edges a priori. It offers the advantage of improved reliability compared to methods based on detecting dual edges of the arteries. Novel application of filtering techniques provide accurate estimates of the statistics of the background. A recursive search scheme is used to aggregate the skeletal representation at multiple resolutions. Results on a set of Digitally Subtracted Angiograms (DSA) have been presented.

Lumen centerline detection in complex coronary angiograms

We have developed a method for lumen centerline detection in individual coronary segments that is based on simultaneous detection of the approximate positions of the left and right coronary borders. This approach emulates that of a clinician who visually identifies the lumen centerline as the midline between the simultaneously-determined left and right borders of the vessel segment of interest. Our lumen centerline detection algorithm and two conventional centerline detection methods were compared to carefully-defined observer-identified centerlines in 89 complex coronary images. Computer-detected and observer-defined centerlines were objectively compared using five indices of centerline position and orientation. The quality of centerlines obtained with the new simultaneous border identification approach and the two conventional centerline detection methods was also subjectively assessed by an experienced cardiologist who was unaware of the analysis method. Our centerline detection method yielded accurate centerlines in the 89 complex images. Moreover, our method outperformed the two conventional methods as judged by all five objective parameters (p < 0.001 for each parameter) and by the subjective assessment of centerline quality (p < 0.001). Automated detection of lumen centerlines based on simultaneous detection of both coronary borders provides improved accuracy in complex coronary arteriograms.

Algorithm effects on computerized vessel analysis from digitized cine film and a new method of generating the centerline of a vessel

This study concerns methods of accurately measuring vessel diameters using coronary cineangiograms. Various edge detection filters are compared and a new method for determining the centerline of a vessel is developed. Vessel phantoms of different sizes are used to evaluate the precision and accuracy of an entropy filter, a combined first and second derivative filter, a Laplace-Gaussian filter, a first derivative filter and a second derivative filter. The entropy filter is found to yield the most reliable, precise and accurate data. For use in a clinical case with an irregular arterial wall, a new method of determining the centerline is delineated in which the centerline of a vessel (which we define as a set of points having equal distance to each of any two opposite edge lines) is calculated with images called distance maps. We find that this method of detecting the centerline provides an acceptable assessment of the centerline on visual evaluation.

Recursive tracking of vascular networks in angiograms based on the detection-deletion scheme

A computer algorithm was developed for automated identification of 2-D vascular networks in X-ray angiograms. This was accomplished by using an adaptive tracking algorithm in a three-stage recursive procedure. First, given a starting position and direction, a segment in the vascular network was identified. Second, by filling it with the surrounding background pixel values, the detected segment was deleted from the angiogram. The detection-deletion scheme was employed to prevent the problem of tracking-path reentry in those areas where vessels overlap. Third, all branch points were detected by use of matched filtering along both edges of the vessel. The detected branch points were used as the starting points in the next recursion. The recursive procedure terminated when no new branch point was found. The algorithm showed a good performance when it was applied to angiograms of coronary and radial arteries. To provide a quantitative evaluation, vascular networks identified by the algorithm were compared to those identified by a human. The algorithm made some false-negative errors, but very few false-positive errors.

Matched filter estimation of serial blood vessel diameters from video images

A method for making a contiguous series of blood vessel diameter estimates from digitized images is proposed. It makes use of a vessel intensity profile model based on the vessel geometry and the physics of the imaging process, providing estimates of far greater accuracy than previously obtained. A variety of techniques are used to reduce the computational demand. The method includes the generation of measurement estimation error, which is important in determining total vessel patency as well as providing a basic measure of diameter estimate accuracy.

Two approaches to motion analysis of the ultrasound image sequence of carotid atheromatous plaque

It has been observed, with a sequence of ultrasonic images, that a significant number of carotid atheromatous plaques moved under the influence of cardiovascular forces. The aim of this project was to analyse the movement of the heterogeneous and homogeneous plaques using both the discrete approach and the continuous approach. In order to verify that the movement of plaque is not due to the movement of the transducer or the respiration of the patient, global movement estimation was also carried out. Experimentation on an artificial image sequence showed that both approaches can track the target accurately. These approaches were then applied on clinical image sequences and the results showed that carotid plaque was locally moved as a result of the momentum transfer rather than the global movement of the surrounding normal tissue.

Three-dimensional quantitative coronary angiography

A method for reconstructing the three-dimensional coronary arterial tree structure from biplane two-dimensional angiographic images is presented. This method exploits the geometrical mathematics of X-ray imaging and the tracking of leading edges of injected contrast material into each vessel for identification of corresponding points on two images taken from orthogonal views. Accurate spatial position and dimension of each vessel in three-dimensional space can be obtained by this reconstruction procedure. The reconstructed arterial configuration is displayed as a shaded surface model, which can be viewed from various angles. Such three-dimensional vascular information provides accurate and reproducible measurements of vascular morphology and function. Flow measurements are obtained by tracking the leading edge of contrast material down the three-dimensional arterial tree. A quantitative analysis of coronary stenosis based on transverse area narrowing and regional blood flow, including the effect of vasoactive drugs, is described. Reconstruction experiments on actual angiographic images of the human coronary artery yield encouraging results toward a realization of computer-assisted three-dimensional quantitative angiography.

Segmentation of coronary arteriograms by iterative ternary classification

A segmentation algorithm for extracting arterial structures in coronary angiograms is presented. The algorithm mimics the process of interactive interpretation in human vision by iteratively implementing a ternary classification and learning process. Two gray-scale thresholds are computed to define three pixel classes: artery, background, and undecided. Then, two new thresholds for undecided pixels are computed using statistics conditioned upon the current classification. The threshold adaptation is governed by a learning algorithm based on the line and consistency measurements around each pixel. The process converges and results in a binary image. The performance of this algorithm on human coronary arteriograms was compared qualitatively to that of a relaxation algorithm and of a scattering based algorithm. Quantitative comparison was also made possible with computer generated images, which were obtained with the help of a model of the imaging chain and a process of interactive visualization of the modeled data. The iterative ternary classifier showed the best performance over a broad range of image quality. The study also demonstrated the use of visualization and user interaction in model building and algorithm development.

A new approach for the automated definition of path lines in digitized coronary angiograms

For the quantitative analysis of a coronary segment from a coronary (cine)angiogram, an initial path line is required which functions as a model for the subsequent automated contour detection. For on-line applications, a new method for the automated definition of arterial path lines has been developed. Required user-interaction consists of the manual definition of a beginning and an endpoint of the arterial segment to be analyzed. The method is based on a combination of a beam tracer and a box technique. A validation study was performed on 47 non obstructed arteries of various lengths and diameters, and on 56 arterial segments with obstructions (up to 86 percent diameter stenosis). In 89% of the cases an acceptable path line was found after the first iteration; the success score increased to 99%, if a simple manual correction was allowed (2 iterations). The method is extremely fast: the overall average search time for the first iteration was 266 ms, for the second iteration 211 ms. Therefore, it may be concluded that this new technique for the automated definition of arterial path lines is extremely suitable for on-line applications.