Surgery – Diagnostic testing – Detecting nuclear – electromagnetic – or ultrasonic radiation
Reexamination Certificate
1999-10-01
2001-08-21
Lateef, Marvin M. (Department: 3737)
Surgery
Diagnostic testing
Detecting nuclear, electromagnetic, or ultrasonic radiation
C600S443000
Reexamination Certificate
active
06277074
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to a method and apparatus for improving motion estimation within a biological tissue and specifically to local Warping for improving contrast and reducing noise in strain imaging.
2. Background Art
Ultrasonics
It is generally known to use ultrasonics for non-invasive imaging of soft biological tissue. An acoustic transducer is used as both a transmitter and receiver of sound pulses in the 1-50 MHz frequency range. The transducer is used in a pulse-echo system where the transducer converts between electrical and acoustic energy to transmit and receive sound pulse to and from the target tissue. Received signals are translated by a computer to produce an image.
Elastography
A technique called elastography was developed in concert with ultrasonics in order to detect inhomogeneities or stiff regions within soft biological tissues. Elastography is a remote sensing technique for imaging elastic properties of biological tissues. Elastograms are formed by comparing two-dimensional arrays of RF echo signals recorded before and after applying a weak compression to the biological tissue. When the force is applied to the surface of the biological tissue, a stress field is generated that deforms the tissue volume within. An elastogram is an image of the resulting deformation or strain of the associated tissue. To create an elastogram, a displacement field is estimated from windowed segments of ultrasonic echo signals acquired before and after a weak compressive force is applied along the axis of the sound beam. Strain Images are formed by taking the gradient of the displacement map obtained via comparison between ultrasound RF echo fields. The displacement deformation is estimated by cross correlating these echo signals. The cross correlation between the amplitude line (A-line) segments in the pre-compression data and the corresponding A-line segments in the post-compression data is calculated. The lag at the peak of the cross correlation reveals the displacement of the tissue caused by compression.
In principle, any diagnostic imaging modality can be used to estimate this displacement. In practice, however, ultrasound and magnetic resonance imaging provide the best motion estimations. Local tissue displacement is the fundamental measurement of strain imaging that determines visibility of targets. This procedure is useful because the motion of a stiff region, like a tumor under compression, is different than motion of normal soft biological tissue under compression. This difference is reflected in relative changes in ultrasonic signals recorded before and after compression.
1-D Companding
Accurate estimates of displacement using only the cross correlation method described above require there to be little strain over the echo segment. Unfortunately, the deformation of biological tissue is typically very complex and large decorrelation errors persist. An early method for reducing some decorrelation errors involved temporally stretching the compressed echo field globally by the amount of the physical compression prior to cross correlation. Because cross correlation methods assume there is only rigid-body motion over the duration of the data window, stretching signals by the applied strain conditions the data to better satisfy this critical assumption. This global temporal stretching is known in the art of radar imaging and sound recording.
Temporal stretching, however, fails to remove significant decorrelation errors because when a nearly incompressible material such as a biological tissue is compressed in one direction, for example, along the axis of the ultrasonic beam, it necessarily bulges in directions not axial to the beam. This causes the echo field to be distorted in three dimensions.
It is known in the art of radar imaging and sound recording to use a technique called 1 -D companding to reduce decorrelation errors in one direction in an attempt to improve image quality. Just like temporal stretching, 1-D companding is a pre-processing technique applied to the pre or post-compression data prior to cross correlation. 1-D companding refers to a spatially-variable signal scaling along the axis of the beam through not just temporal stretching, but compression and expansion of the signal by an estimation of the applied strain to that signal. This technique attempts to reduce estimation noise by restoring coherence between the waveforms to be cross-correlated. The improvement in image noise using 1-D companding, however, is highly dependent on the duration of the data window, the elastic heterogeneity of the medium, and the amount of applied strain.
Accordingly, five main problems exist with the current acoustic motion estimation and strain imaging techniques: 1) Noise; 2) Contrast; 3) Out of Plane Displacement; 4) Shear Motion; and 5) Rotational Motion.
Noise
Noise is a consequence of the disagreement between the pre and post-compression echo fields strain (lack of coherence). Noise increases with the amount of applied compression. Two main reasons for this increase are: 1) the windows used for cross correlation analysis may contain significant strain due to the size of the window; and 2) the segment of the biological tissue under investigation may rotate, shear and/or move laterally from its original location so that the corresponding A-lines in pre and post-compression images no longer contain correlated data from the same region of the object. Correspondingly, it is well known in the art to compress biological tissue only a small percentage (typically less than 1%) of its total height to maintain strain noise at an acceptable level.
Contrast
Contrast is the difference in brightness between the target within the biological tissue and the background of the biological tissue itself. An increase in compression increases the contrast, which allows for better viewing of the target in the strain image. As noted above, however, increased compression also increases noise, which distorts the strain image. Also, as noted above, compression is generally limited to 1% of the targets total height in order to limit noise. Limiting compression also limits contrast making abnormal tissue more difficult to detect.
Out of Plane Displacement, Shear and Rotational Motion
Out of plane displacement occurs when the target moves perpendicularly out of the scan plane upon compression. The directions perpendicular to the scan plane, parallel with the ultrasound beam, and perpendicular to the beam in the scan plane are referred to as elevational, axial, and lateral respectively. Shear motion occurs when motion in one direction, e.g., axially, depends on the position in another direction, e.g., laterally. Rotational motion occurs when the target rotates away from its original orientation either within or out of the scan plane. While there are many elements on a 1-D linear array transducer for sending pulses and receiving echoes in the lateral direction, there is only one row of elements on the transducer in the elevational direction. Consequently, a target that may have moved in the elevational direction with respect to the transducer can not be tracked because the post-compression echo field does not contain any data from such a target; resulting in axial decorrelation. Similarly, shear and rotation result in greater decorrelation noise when simple techniques such as 1-D stretching and 1-D cross correlation are used.
Unfortunately, biological tissues do not move simply in only one dimension when compressed. Correspondingly, the known use of temporal stretching and 1-D companding is insufficient to avoid decorrelation errors. Extending the concept of global temporal stretching to 2-D and 3-D is not viable because it is rare that the biological tissue is homogeneous and incompressible at all spatial scales as required.
Accordingly, it is a primary object of the present invention to provide a method and apparatus for improving motion estimation within a biological tissue.
It is another object of the present invention to prov
Chaturvedi Pawan
Hall Timothy J.
Insana Michael F.
Zhu Yanning
Imam Ali M.
Lateef Marvin M.
Sonnenschein Nath & Rosenthal
University of Kansas Medical Center
LandOfFree
Method and apparatus for motion estimation within biological... 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 and apparatus for motion estimation within biological..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for motion estimation within biological... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2534108