Image analysis – Applications – Biomedical applications
Reexamination Certificate
1999-11-23
2003-06-24
Patel, Jayanti K. (Department: 2725)
Image analysis
Applications
Biomedical applications
C600S407000, C250S474100
Reexamination Certificate
active
06584216
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to magnetic resonance imaging (MRI), image normalization, image display, and image processing.
BACKGROUND OF THE INVENTION
Magnetic Resonance Imaging (MRI) has revolutionized radiological imaging of the internal structures of the human body. It has the advantage of being noninvasive with no known health hazards. A variety of MRI protocols are currently available, with and without the use of contrast agents, such as T1, T1 with a contrast agent, T2 and proton density (Pd) with spin-echo (SE) or fast spin-echo (FSE) sequences, magnetization transfer (MT), FLAIR, SPGR, and GRASS. These protocols allow a variety of contrasts to be set up among the different tissues within the same organ system.
Ironically, this richness of image acquisition modalities presents a major problem. The image intensities in MRI do not have a fixed meaning, not even within the same protocol for the same body region obtained on the same scanner for the same patient. This is due to a variety of scanner-dependent variations. Unlike other imaging methods, such as x-ray computerized tomography, MR images lack a standard and quantifiable interpretation of image intensities. Consequently, absolute intensity values lack a fixed meaning. This implies that MR images cannot be displayed at preset windows. One may have to re-adjust the window settings for each case, perhaps even for the same patient. The lack of a meaning for intensities also poses problems in image segmentation and quantification (Bezdek et al.,
Med Phys
. 20 (4):1033-1048 (1993); Clarke et al.,
Magn. Reson. Imaging
, 11 (1):95-106 (1993); Kikinis et al.,
J. Magn. Reson. Imaging
, 2:619-629 (1992); Udupa et al.,
IEEE Trans. Med Imaging
, 16(5):598-609 (1997)).
Most visualization and analysis methods have parameters. The exceptions are perhaps manual methods, wherein the human knowledge can be considered to be representative of the parameters. However, the results of segmentation by two physicians are likely to differ because of the differences in their training. Setting values for the parameters for the non-manual methods becomes more difficult without the same protocol-specific intensity meaning. Thus, there has been a recognized need in the art, with regard to protocols that were the same or “close” to each other, to try to develop ways that the resulting images would also be “close.”
Attempts have been made in the past to calibrate MR signal characteristics at the time of acquisition using phantoms (Edelstein et al.,
Med. Phys
. 11 (2):180-185 (1984); Yamamoto et al.,
Radiology
209(P):582 (1998)). Although it is feasible to do such a calibration of each patient scan, it would obviously be too cumbersome to be practical or efficient. Moreover, such a technique is not applicable to image data that have already been acquired without the required calibration phantoms. Post-processing techniques that could be applied to the image data without special acquisition requirements would clearly be more practical and attractive to the practitioner. Such methods would not only make acquiring images simpler, but also permit processing already-acquired data.
There has been a natural tendency to think that a simple scaling of the minimum to maximum intensity range of the given image to a fixed standard range could solve this problem. However, as illustrated in
FIG. 1
, simple scaling is usually ineffective in achieving a similarity of intensities.
A post-processing technique to automatically adjust the contrast and brightness of MR images (i.e., “windowing”) for image display has been presented in Wendt,
J. Digit. Imaging
7:95-97 (1994). However, although such automatic windowing may achieve display uniformity, it is inadequate for quantitative image analysis, since the intensities still do not have tissue-specific meaning after the windowing transformation. There does not seem to have been any serious attempt to address this latter problem in the past.
Therefore, although the need for image consistency was recognized, solutions to the problem were not forthcoming until the present invention. Until now, there remained a need for a simple, quick and reliable means and method by which MR images could be transformed under conditions by which there is a significant gain in the similarity and the meaning of intensities of the resulting images.
SUMMARY OF THE INVENTION
The present invention provides a method for standardizing MR image intensity scales by post-processing intensity transformation techniques on routinely acquired images, without requiring specialized acquisition protocols and calibration phantoms. The standardizing method offers previously unattainable consistency of intensity meaning of tissues by devising a transformation that is specific to a given MRI protocol, P&egr;
, and for any body region, D&egr;
, to provide standardized images,
si
. Essentially, the histogram of a given volume image is deformed to match a “standard” histogram for the corresponding protocol and body-region. The parameters of the standard histogram are “learnt” in a training step. This permits standardizing and fixing “windows” by protocol, body region, and tissue regions, thereby minimizing or eliminating the human interaction required in the per-case manual window adjustments that are currently required in visualizing MR images at physician viewing stations. The method offers significantly more consistent tissue meaning for MR image intensities than the images,
i
, before standardization.
Accordingly, it is an object of the invention to provide a method for standardizing MR image intensity scales for imaging a region of the body of a patient, wherein the method comprises post-processing intensity transforming a routinely acquired MR image of the region of the patient's body, wherein the image is acquired by any selected MRI protocol, and wherein the method requires no specialized acquisition protocol or calibration phantoms.
It is a further object of the invention to provide a method for standardizing MR image intensity scales, wherein a standardized intensity histogram is achieved for the selected MRI protocol, and for the selected region of the patient's body, by the steps comprising: a) estimating for the selected protocol and region of the patient's body, certain landmarks from a given set of image histograms to form a standard histogram, a standardizer; (b) imaging the region of the patient's body by the selected protocol to form an image histogram, and (c) computing the standardized intensity histogram by mapping landmarks determined from the image histogram to corresponding landmarks of the standard scale histogram.
Preferred embodiments of the invention are provided, wherein the standardized images of similar intensities have similar tissue meaning, independent of variations within or between patients. The standardization is specific to a MRI protocol and/or to a specific body region. The resulting standarized images are independent of variations within or between MRI scanners.
Theoretical guidelines are provided, accompanied by a practical demonstration of how to utilize them for selecting the values of the parameters of the method. In addition, the present invention offers proofs that lossless intensity transformation and order is guaranteed if choices are made in accordance with the provided guidelines. Several variants of the method are provided, which utilize different, representative scale parameters and landmarks. Because median and other percentile values are more robust than the mode, using the landmarks of the present invention results in a more robust standardizer. “Mode” refers to the image intensity, which occurs most frequently within the body region in the acquired image. Consequently, more consistent meaning of intensities and better-defined ranges for different tissues on the standard scale are possible by the present methods than were previously possible, permitting a practitioner to set better intensity of interest ranges, while retaining the ability to distinguish releva
Nyúl László G.
Udupa Jayaram K.
Choobin Barry
Dilworth Paxson LLP
McConathy Evelyn H.
Patel Jayanti K.
The Trustees of the University of Pennsylvania
LandOfFree
Method for standardizing the MR image intensity scale does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for standardizing the MR image intensity scale, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for standardizing the MR image intensity scale will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3144851