Image analysis – Applications – Biomedical applications
Reexamination Certificate
2000-12-05
2004-04-06
Patel, Jayanti K. (Department: 2625)
Image analysis
Applications
Biomedical applications
C600S334000, C600S358000, C600S363000
Reexamination Certificate
active
06718055
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to the medical imaging arts. It finds particular application in the area of detecting and quantifying the perfusion of an amount of medical imaging contrast agent through a region of interest delineated either manually, automatically, or semi-automatically.
There are many imaging modalities which give useful information when used for diagnostic imaging. The most widely used are X-rays CT, PET, SPECT and Magnetic Resonance (MR). The advent and wide use of contrast agents injected into the human body, such as Gadolinium, or Gad have enhanced the diagnostic utility such as in x-rays, CT, PET and MR. See e.g. U.S. Pat. No. 4,957,939 to Gries et al., entitled Sterile Pharmaceutical Compositions of Gadolinium Chelates Useful Enhancing NMR Imaging. When contrast agents are employed, it becomes significant to track their ability to perfuse in the tissues, so-called, perfusion through the region of interest. Indeed, this has given rise to the field of perfusion engineering which has played a vital role in early cancer detection and therapy.
While the use of contrast agents and MR scans has been in use for some time, especially in the area of brain scans, other regions of interest provide additional difficulties. For example, during a brain scan the brain can be held fairly stationary. On the other hand, breast, kidney and cardiac tissue tend to move during the exam, such as with the respiratory cycle. The motion includes rigid or translation motion as well as affine or nonrigid motion, such as flexing, rolling, compressing, and other nonlinear motions associated with soft tissue.
Additionally, contrast agents tend to be absorbed at different rates depending on tissue type. Again, while this is rarely a difficulty in the brain environment, the presence of different kinds of tissue classes such as glands, ducts, fat, ligaments, bone and the like introduce image processing requirements that quickly overcome even the most powerful imaging computers. Indeed, the images tend to be volume images of significant size. For example, for breast imaging, the image data set is large enough to encompass the entire breast. Moreover, in order to detect cancers in ducts and other fine regions, a resolution on the order of a cubic millimeter per pixel is desirable.
Thus, different contrast agent uptake rates and computationally expensive image processing, typically result in fitting mathematical models to estimate the contrast uptake curve. For example, researchers have tried modeling the ratio of measured signal intensity after the contrast injection to the pre-contrast signal intensity using the kinematics of the uptake based on exponential models. Some have proposed the compartmental model for this, others have used the Marquart algorithm for the analysis.
The present invention contemplates a new, improved method and apparatus for perfusion quantification and diagnostic imaging which overcomes the above-referenced difficulties and others.
SUMMARY OF THE INVENTION
In accordance with an embodiment of the present invention, a method of medical imaging includes correcting for motion in a series of temporally spaced images. A plurality of pixels are selected in a region of interest spatially within each of the images based on common characteristics of the pixels and for the selected pixels, a contrast agent uptake curve is computed indicative of a presence of the contrast agent in the pixels over time.
In accordance with another aspect of the present invention, the method further includes verifying correspondence of the selected pixels in each of the temporally spaced images.
In accordance with another aspect of the present invention, the selecting step includes classifying pixels in the region of interest based on a determinable tissue characteristic.
In accordance with another aspect of the present invention the selecting step further includes selecting one of the classifications of pixels as a class of pixels expected to react with the contrast agent.
In accordance with another aspect of the present invention the correcting step includes identifying a sub-set of pixels in each temporally spaced image, and deriving a relationship between successive sub-sets to describe motion between temporally spaced images.
In accordance with another aspect of the present invention the identifying step includes successively down sampling and smoothing pixels into progressively smaller sub-sets of pixels in the temporally spaced images until successive sub-sets converge.
In accordance with another aspect of the present invention the correcting step further includes minimizing entropy of the sub-set of pixels between successive temporally spaced images.
In accordance with another aspect of the present invention the verifying step includes computing co-occurrence matrices between successive images based on texture, each co-occurrence matrix indicative of similarity of contrast agent perfused pixels in successive images.
In accordance with another aspect of the present invention the method further includes generating binary masks from pixel classifications in the region of interest for establishing textural correspondence of the selected pixels.
In accordance with another aspect of the present invention the method further includes computing textural properties of the selected pixels from the generated binary mask.
In accordance with another embodiment of the present invention, a diagnostic imaging system includes an image memory for storing a plurality of temporally spaced digital image representations reconstructed from diagnostic data generated by an imaging device. The system further includes a perfusion processor including a motion correcting algorithm that registers the plurality of digital image representations, a filtration algorithm that spatially classifies pixels in a region of interest on each of the digital image representations, and a verification algorithm which establishes correspondence between selected classified pixels over successive digital image representations.
In accordance with another aspect of the present invention the motion correcting algorithm includes first and second windows including a selected number of pixels from a first digital image representation. The system further includes a comparator which recursively compares the first window and the second window until differences are minimized.
In accordance with another aspect of the present invention the comparator includes a transform defining a relationship between pixels in the first window and the second window when the differences are minimized, the comparator applying the transform to the digital image representations.
In accordance with another aspect of the present invention the filtration algorithm includes an algorithm which assigns a value to each pixel in the image representation based on a characteristic of the pixel. The filtration algorithm also includes an algorithm which selects pixels representing a characteristic of interest, and an algorithm which excludes pixels of other than the selected pixels.
In accordance with another aspect of the present invention the verification algorithm includes a co-occurrence matrix process which computes matrices based on a texture parameter of the selected pixels across successive image representations.
In accordance with another embodiment of the present invention, a medical imaging system includes an imaging device for producing a plurality of temporally spaced image representations, a memory for storing the image representations, and a processor for manipulating the image representations for viewing on a display. The processor being controlled by the computer implemented steps of: (a) aligning the plurality of temporally spaced image representations; (b) sorting selected pixels in a region of interest spatially within each of the image representations for further processing; (c) establishing correspondence between the selected pixels over successive image representations; and, (d) determining a curve indicative of a presence of a
Fay Sharpe Fagan Minnich & McKee LLP
Koninklijke Philips Electronics , N.V.
Patel Jayanti K.
Tabatabai Abolfazl
LandOfFree
Temporal and spatial correction for perfusion quantification... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Temporal and spatial correction for perfusion quantification..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Temporal and spatial correction for perfusion quantification... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3212848