System and method for the noninvasive identification and display of breast lesions and the like

• US 5,003,979 A
• Filed: 02/21/1989
• Issued: 04/02/1991
• Est. Priority Date: 02/21/1989
• Status: Expired due to Term

• Current Assignee: Linda Merickel
• Original Assignee: University of Virginia
• Inventors: Merickel, Michael B., Adams, Ann H., Brookeman, James R.
• Primary Examiner(s): Howell, Kyle L.
• Assistant Examiner(s): ZELE, KRISTA MICHELLE
• Application Number:
  US07/313,183
• Time in Patent Office: 770 Days
• Field of Search: 128/653, 128/664-665, 128/736, 128/915, 128/660.09, 128/660.06, 128/660.01, 128/661.02, 378/37, 324/309, 364/413.13, 364/413.14, 364/413.25
• US Class Current: 600/410
• CPC Class Codes: G01R 33/465 applied to biological material, e.g. in vitro testing
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06T 2207/10088 Magnetic resonance imaging [MRI]
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 10/478 Contour-based spectral representations or scale-space representations, e.g. by Fourier analysis, wavelet analysis or curvature scale-space [CSS]
• CPC Class Codes: G06V 2201/032 of protuberances, polyps nodules, etc.
• CPC Class Codes: Y10S 128/915 Ultrasound mammography

Abstract

There is disclosed an image processing, pattern recognition and computer graphics system and method for the noninvasive identification and evaluation of female breast cancer including the characteristic of the boundary thereof using multidimensional Magnetic Resonance Imaging (MRI). The system and method classifies the tissue using a Fisher linear classifier followed by a refinement to show the boundary shape and whether the surface of the carcinoma is lobulated or spiculated. The results are a high information content display which aids in the diagnosis and analysis of breast cancer and to assist in any surgical or other remedial planning. The high information content display also assists in the assessment of the effectiveness of therapies showing any reduction or increase in the size of the carcinoma.

Claims

1. A method for the noninvasive determination and display of a body part composed of different tissue types of interest of a mammalian live body where the boundary of said tissue types encloses an area or volume and where the boundary shape is of significance using information obtained from multiple image slices produced by a body scan such as MRI or PET comprising the following steps:

produce multiple sequences of image slices by noninvasive scanning a mammalian body;

classify at least some of the different tissue types in said body part by comparing by computer said tissue types in said image slices to classifier parameters of said tissue types created by data obtained from other similar mammalian bodies;

determine the boundary shape of one or more of said tissue types of interest which have been classified; and

display said one or more tissue types of interest including the boundary shape thereof.

2. The method of claim 1, wherein said body part is a human female breast with said classification step classifying the tissue types into fat, fibroadenoma, carcinoma and cyst and said tissue types of interest include carcinoma.

3. The method of claim 2, wherein the classification step is accomplished by using a statistical classifier.

4. The method of claim 3, wherein said statistical classifier is a Fisher Linear Classifier.

5. The method of claim 3, wherein said body scan is MRI and at least a T₁ weighted image, T₂ weighted image and Dixon opposed image is made of said image slices.

6. The method of claim 2, wherein said boundary shape is determined as being either smooth, lobulated or spiculated by converting the boundary of said tissue types of interest to a polar format thereof.

7. The method of claim 6, wherein said polar format includes the steps of:

performing a Fourier transform on said polar format to generate a resultant having frequency components; and

analyzing the said frequency components of said resultant from said Fourier transform.

8. The method of claim 2, wherein said boundary shape is defined by determining a shape measure based on a ratio of the boundary shapes peripheral length or area to the enclosed area or volume, respectively.

9. The method of claim 1, including the following steps:

reconstructing the shape and volume of said one or more tissue types of interest into a 3-D image by converting the sequence of tissue images to quad trees; and

computing and indicating volumes and areas and creating a 3-D display showing the boundary shape of said one or more tissue types of interest.

10. The method of claim 9, wherein:

said body part is a human female breast and said tissue types of interest include carcinoma with said classification step classifying the tissue types into fat, fibroadenoma, carcinoma and cyst;

the classification step is accomplished by using a statistical classifier; and

said body scan is MRI and at least a T₁ weighted image, T₂ weighted image and Dixon opposed image is made of said image slices.

11. A noninvasive imaging analysis system for the display of a body part composed of different tissue types of interest of a mammalian live body where the boundary of said tissue types encloses an area or volume and where the boundary shape is of significance using information obtained from multiple image slices produced by a body scan such as MRI and PET comprising:

an electronic digital storage means for storing said information as well as other information;

a computer for processing said information from said storage means;

a graphics display station for displaying an image resulting from said processed information;

said computer for processing including

means for taking multiple sequenced image slices resulting from the body scan and determining the boundary shape of one or more of said tissue types of interest;

means for classifying at least some of said tissue types in said body part by comparing said tissue types to classifier parameters of said tissue types stored in said storage means;

means for determining the boundary shape of one or more of said tissue types of interest which have been classified; and

means for displaying said one or more tissue types of interest including its location, size and boundary shape thereof.

12. The system of claim 11, wherein:

said body part is a human female breast with said classifying means classifying the tissue types into fat fibroadenoma, carcinoma and cyst and said tissue types of interest include carcinoma.

13. The system of claim 12, wherein the classification means is accomplished by using a statistical classifier.

14. The system of claim 13, wherein said statistical classifier is a Fisher Linear Classifier.

15. The system of claim 13, wherein:

said body scan is MRI which generates at least a T₁ weighted image, T₂ weighted image and Dixon opposed image of said image slices.

16. The system of claim 12, wherein said boundary shape is determined as being either smooth, lobulated or spiculated by means for converting the boundary of said tissue type of interest to a polar format.

17. The system of claim 16, wherein:

said converting means includes means for performing a Fourier transform on said polar format to generate a resultant having frequency components and analyzing said frequency components of said resultant from said Fourier transform.

18. The system of claim 12, wherein said boundary shape determining means calculates a shape measure based on a ratio of the boundary to the enclosed area or volume.

19. The system of claim 11 where said computer for processing includes:

means for reconstructing the shape and volume of said one or more tissue types of interest into a 3-D image by converting the sequence of tissue images to quad trees; and

means for computing and indicating volumes and areas and creating a 3-D display showing the boundary shape of said one or more tissues types of interest.

20. The system of claim 19, wherein:

said body part is a human female breast and said tissue types of interest include carcinoma with said classification means classifying the tissue types into fat, fibroadenoma, carcinoma and cyst which is accomplished by using a statistical classifier; and

said body scan is MRI which generates a T₁ weighted image, T₂ weighted image and Dixon opposed image of said image slices.