X-ray or gamma ray systems or devices – Specific application – Computerized tomography
Reexamination Certificate
2000-08-25
2003-03-25
Bruce, David V. (Department: 2882)
X-ray or gamma ray systems or devices
Specific application
Computerized tomography
C378S008000, C378S901000
Reexamination Certificate
active
06539074
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to generation of computed tomographic images. More particularly, the invention relates to reconstruction of multiple, tomographic slice images to represent an imaged object in three spatial dimensions, using four-dimensional tomographic projection data.
Tomography (or computed tomography, abbreviated “CT”) is a class of technologies for generating images that accurately represent the interiors of objects. Unlike direct images (such as photographic images, conventional radiographs, etc.), a tomographic image is generated by reconstructing image data from projection data. The “projection data” are data that represent properties of illuminating rays transmitted through the object along numerous non-parallel paths. Reconstruction of the projection data into image data may be carried out through an appropriate computational process for image reconstruction. As used herein, the term “tomographic image” will mean any image generated by a process of tomography. A three-dimensional image of the interior of the object is therefore a “tomographic image,” also, if reconstructed from projection data. A tomographic image depicting a cross section of the object is sometimes called a tomographic “slice” image.
Medical applications of tomography have become commonplace and are widely known to provide images of almost photographic detail. A trained physician can draw detailed diagnostic conclusions from a single tomographic slice image. More generally, tomographic imaging has become a useful tool in such diverse areas of application as nondestructive testing, microseismic mapping of underground geological formations, and three-dimensional imaging with electron microscopy.
In the context of medical applications, the challenges of early diagnosis and treatment of disease continue to demand improved assessment methods. For example, serious limitations have continued to exist in the available approaches for assessment of coronary artery disease. Coronary artery disease principally in the form of atherosclerosis leading to vessel stenosis, is a leading cause of myocardial infarction and ischemia.
Several imaging modalities have been available for assessing coronary artery disease. The standard technique for assessment of atherosclerotic stenosis in the coronary vasculature is cardiac catheterization. This technique requires the placement of a catheter in a particular branch of the coronary vessel, injecting a radio-opaque contrast agent while X-ray fluoroscopic images are acquired using an image intensifier or a digital detector array.
Catheterization procedures are effective but are also highly invasive. Moreover, conventional techniques for assessment by catheterization involve acquiring two dimensional (2D) projection images of the 3D vasculature. Hence the data that is acquired may be confounded by structures in the chest or by contortions of the vessels themselves.
Various forms of tomography have been proposed as alternatives to angiographic catheterization. However, most previous efforts in this direction have yielded mixed results, due to issues of image quality. The acquisition of data for accurate tomographic images depends upon the patient being stationary while data is being acquired, to ensure that the acquired data are mathematically consistent. Data inconsistency is well known to create artifacts (such as streaks and blurring) in the resulting reconstructed images. Blurring of image details (such as vessel walls) is a known problem that tends to reduce the effectiveness of the existing approaches to quantification.
The patient's respiration is a source of motion that is readily controlled by acquiring the data during periods when the patient's breath is held (such periods being called “breath-holds”). However, unlike other organs, the beating heart undergoes motion even during breath-holds. Effective data acquisition for tomographic imaging therefore calls for measures to compensate for the concurrent motion of the heart.
Magnetic resonance (MR) imaging has been implemented with special data acquisition protocols that provide some compensatory measures. The acquired data can be reconstructed to generate a three-dimensional (3D) representation of the relevant structures of the subject's heart. The coronary vasculature may be segmented from the myocardium, whereby vascular stenosis may then be quantified by conventional techniques.
On the other hand, a basic problem with generating volumetric images of cardiac structures is to address simultaneously four competing data collection criteria. First, generating a volumetric representation of a cardiac structure entails acquiring sufficient projection data for reconstruction of a plurality of stacked slice images, or else for a three-dimensional reconstruction. Second, the spatial resolution and fidelity of the reconstructed image data are directly tied to the number of distinct projections in the acquired projection data. Hence, reconstructing usable images entails acquiring numerous projections. Third, because the heart is in motion during the data acquisition, the acquired data variously represent the heart in all different phases of the cardiac cycle. Reconstruction of an accurate three-dimensional representation thus entails some form of cardiac gating in the data acquisition process, so that the projections used to reconstruct the image data all represent the heart structures at approximately the same phase of the cardiac cycle. Fourth, shorter data acquisition times are more desirable to minimize the patient's discomfort, the radiation dose (for CT scans using ionizing radiation), and motion blur due to respiration or multiple breath-holds.
Existing cardiovascular MR techniques have simultaneously addressed at most three of these four criteria. Effective study of the coronary vasculature has depended upon images with high resolution and high fidelity. To obtain sufficient data with existing techniques, the data acquisition time for a three-dimensional representation can amount to several minutes and has entailed both respiratory gating and cardiac gating. U.S. Pat. No. 5,997,883 to Epstein, et al., notes that generation of multiple slice images from conventional spin echo sequence data results in slice images representing the heart at different phases of the cardiac cycle. The same patent describes methods for generating time-lapse (i.e., “cine”) images by using short repetition times and retrospective data resorting to bin the data according to cardiac phase. However, such methods are specifically directed toward imaging the time evolution of the heart at a single spatial location and thus are not readily applicable to volumetric imaging.
X-ray tomography (XCT) has also been considered for assessment of cardiovascular disease. Recent developments in image reconstruction algorithms have improved this situation somewhat by enabling CT techniques to achieve better time resolution than with the standard operation mode. The basic approach of these techniques has been to segment consistent projection data collected from multiple heart cycles based on predetermined sectors of the cardiac cycle. Such data sectoring methods have been shown to improve significantly the time resolution of the collected data. In practice, however, these existing methods have been found to introduce some image artifacts due to inconsistencies in the data.
On the other hand, it would be desirable to generate three-dimensional images using as much information as possible from the volumetric projection data.
It is therefore apparent that a continuing need exists for an efficient technique to generate three-dimensional images from volumetric projection data.
Such a method desirably would provide acceptable spatial resolution in the resulting images, without requiring extensive additional data acquisition. In imaging contexts where the imaged object is in motion, such as in cardiac imaging, the method would desirably overcome the artifact problems that have existed with data segmentation methods.
Cline Harvey Ellis
Yavuz Mehmet
Bruce David V.
General Electric Company
Hale Lester R.
Patnode Patrick K.
LandOfFree
Reconstruction of multislice tomographic images from... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Reconstruction of multislice tomographic images from..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Reconstruction of multislice tomographic images from... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3010982