Surgery – Diagnostic testing – Detecting nuclear – electromagnetic – or ultrasonic radiation
Reexamination Certificate
1999-05-26
2001-05-29
Lateef, Marvin M. (Department: 3737)
Surgery
Diagnostic testing
Detecting nuclear, electromagnetic, or ultrasonic radiation
Reexamination Certificate
active
06238342
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to ultrasonic imaging, and more particularly relates to methods and apparatus for classifying tissue-type in accordance with spectral and clinical data.
BACKGROUND OF THE INVENTION
Ultrasonic imaging is an important and cost effective medical diagnostic tool. Ultrasonic spectrum analysis has been used to diagnose and monitor disease in experimental settings. However, ultrasonic techniques which can reliably differentiate among tissues, e.g., between cancerous tissue and non-cancerous tissue, have proven elusive.
It is well established that early detection of cancer is a key element to successful treatment. In the case of prostate cancer, where nearly 180,000 new cases are expected to be detected in 1999 in the United States alone, cancer detection usually requires undergoing a needle biopsy under the guidance oftransrectal ultrasound (TRUS). However, while TRUS can effectively display the outline of the prostate, thus insuring that the biopsy needle is properly directed into the glandular portion of the prostate where cancer generally arises, current TRUS is limited in its ability to define areas within the prostate of high cancer risk. As a result, the person performing the biopsy takes multiple biopsy samples somewhat randomly using the ultrasonically visible anatomic features of the gland as guides in selecting biopsy sites. However, the fact that nearly 30% of patients who are cancer negative in a first biopsy are cancer positive in a second biopsy taken within a one year period, suggests that many cancers are missed by current TRUS guided biopsy. Thus, a vast opportunity exists for improving the techniques used to select the location of the biopsy sites used for the detection of cancer in the prostate.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a method of characterizing various material-types using ultrasound measurements in connection with non-ultrasound data related to or indicative of the material types.
It is an object of the present invention to provide a method of characterizing various tissue-types using ultrasound measurements and clinical data.
It is another object of the present invention to provide a method of identifying various tissue-types using ultrasound measurements and clinical data.
It is a further object of the present invention to provide a method of displaying a real-time image depicting different tissues differently, e.g., suspect tissue differently than non-suspect tissue.
It is another object of the present invention to provide an ultrasound apparatus which performs real-time imaging of an area undergoing a biopsy procedure which can display suspect tissue regions in a manner distinct from non-suspect tissue regions.
It is a further object of the present invention to provide three dimensional rendering of different tissues, i.e., suspect tissue, using ultrasound techniques.
In accordance with the present invention, an ultrasound apparatus for performing material classification includes an ultrasound scanner for acquiring RF echo signals related to properties of a material being tested, such as biological tissue. The RF echo signals are presented to a digitizer which converts the RF signals to digital signals, representing a plurality of spatial points in a scanned plane. A processor operatively coupled to the digitizer and ultrasound scanner is responsive to the digital signals from the digitizer and extracts spectral parameters relating to the RF signals. An input device is also included for providing material-related data to the processor. A classifier, which is responsive to at least a portion of the spectral parameters and at least a portion of the material-related data, is employed to assign a tissue classification score to the spatial points of an ultrasound scan. A display is provided for displaying the assigned classification scores.
The material can take the form of biological tissue and the material related data can take the form of clinical data related to a test subject. The display preferably takes the form of an image of the ultrasound scan.
Preferably, the classifier includes a look-up table which is created using a neural network trained with a large database of clinical, spectral and histological data from patients who have undergone an ultrasound guided biopsy.
The classification scores serve as an indicator of material type, e.g., tissue type. In one embodiment, the classification scores distinguish cancerous tissue from non-cancerous tissue. The classification scores define a range of values which is preferably divided into a plurality of sub-ranges, with each sub-range representing a most-likely material type or condition, tissue type or level of suspicion of cancer. Each level of suspicion is preferably displayed with a unique visual parameter, such as color or specific grey scale value.
The spectral values used preferably include at least a portion of slope value, mid-band value and intercept value extracted from an approximation of a power spectrum, e.g., a linear regression of the spectrum of the RF echo data. When the tissue being classified is in the prostate, the clinical data preferably includes a prostate specific antigen (PSA) level and age and the PSA level is adjusted in accordance with expected ranges of normal related to a patients age.
Also in accordance with the present invention, an ultrasound apparatus for building a database and training a classifier for a real-time classification system is defined. The apparatus includes an ultrasound scanner for acquiring RF echo signals related to material properties. A digitizer is responsive to the ultrasound scanner and converts the RF signals to digital signals representing a plurality of spatial points in an ultrasound image. A processor, which is operatively coupled to both the ultrasound scanner and digitizer, is responsive to the digital signals from the digitizer and extracts spectral parameters relating to the RF signals. An input device is provided for entering material-related data to the apparatus. A non-volatile computer storage device is provided for storing records of a database, the records include material-related (e.g., clinical data) as entered using the input device, spectral data provided by the processor and corresponding conclusive data, e.g., histological data from biopsies taken with the guidance of the ultrasound scanner.
A classifier having a plurality of inputs and providing a classification output value is operatively coupled to the database. To train the classifier, at least a portion of said spectral parameters and at least a portion of said material-related data are used as input parameters and the conclusive data is used as an expected output value and the classifier iteratively adapts the processing of the input parameters until the output value substantially matches the expected output value for each record in the database.
Preferably, the classifier includes a look-up table which relates each combination of said portion of spectral parameters and said portion of clinical data to a predetermined memory location, and during a training operation, each memory location is provided a tissue classification score corresponding to the unique combination of input parameters.
The spectral values used preferably include at least a portion of slope value, mid-band value and intercept value extracted from a linear estimate of a power spectrum of the RF echo data. When the material being classified is tissue-type in the prostate, the clinical data preferably includes a prostate specific antigen (PSA) level and age.
A method of material classification ultrasound imaging of a subject, in accordance with the present invention, includes the steps of receiving material-related data regarding the subject; performing an ultrasound scan of the subject and collecting RF echo data therefrom; extracting spectral parameters from the RF echo data at a plurality of points of the ultrasound scan; applying at least a portion of the material-related data and at least a
Feleppa Ernest Joseph
Liu Tian
Lizzi Frederic Louis
BakerBotts L.L.P.
Imam Ali M.
Lateef Marvin M.
Riverside Research Institute
LandOfFree
Ultrasonic tissue-type classification and imaging methods... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Ultrasonic tissue-type classification and imaging methods..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ultrasonic tissue-type classification and imaging methods... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2566545