Method of investigating mammograms for masses and calcifications, and apparatus for practicing such method

• US 5,212,637 A
• Filed: 11/22/1989
• Issued: 05/18/1993
• Est. Priority Date: 11/22/1989
• Status: Expired due to Term

• Current Assignee: Scanis, Inc.
• Original Assignee: Stereometrix Corporation
• Inventors: Saxena, Kripa C.
• Primary Examiner(s): Hayes, Gail O.
• Application Number:
  US07/441,452
• Time in Patent Office: 1,273 Days
• Field of Search: 364/413.13, 364/413.26, 382/6, 382/19, 382/22, 382/54, 378/99, 358/111, 358/167
• US Class Current: 600/407
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G16H 30/20 for handling medical images, e.g. DICOM, HL7 or PACS

Abstract

A method and apparatus as described facilitating the investigation of a human breast for a malignancy. The apparatus includes an optical assembly which includes a camera for acquiring information from a mammogram. This information is fed to a programmed computer which applied preselected criteria to identify suspicious masses and calcifications depicted by the mammogram. It further includes a graphics monitor for, among other things, displaying a reconstruction of the mammogram having the masses and the calcifications highlighted. The optical assembly also includes a zoom lens on the camera enabling one to obtain a magnified view on the monitor via the camera of a particular region it is desired to be investigated in more detail. The method of the invention includes optically analyzing a mammogram to acquire information defining characteristics of the breast desired to be analyzed, applying preselected criteria to such information, and thereafter displaying the results.

Claims

1. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) converting intuitive criteria used to investigate a human breast for malignancies to specific numerical criteria;

(b) programming a computer with said criteria;

(c) acquiring information in said programmed computer defining a human breast;

(d) using said computer to apply said specific numerical criteria defined by said program to said information to identify regions of said breast to be investigated, said information defining a calcification, and said step of using said computer includes applying criteria programmed into the same identifying regions having a predetermined number, size, and shape of calcifications.

2. The method of claim 1 wherein said preselected criteria that are applied, segregate regions of said breast having clusters of at least two calcifications within about one square centimeter from other portions of said breast.

3. Apparatus for assisting investigation of an object for a preselected condition, comprising:

(a) optoelectronic means for acquiring information from an image of said object;

(b) a table for positioning an image of said object in the field of view of said optoelectronic means;

(c) a computer operably connected to said optoelectronic means for acquiring said information from said means and for identifying a region of said object to be investigated, said computer also being adapted to provide a reconstruction of the image viewed by said camera; and

(d) means for outputting information defining said region.

4. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram defining said human breast with optoelectronic means, including the step of generating a digital representation of said mammogram;

(b) analyzing said information by applying preselected criteria to said information to identify regions of said breast defined by said information having either or both a mass and a calcification, said step of analyzing including detecting digital information in said representation defining masses in said human breast, and detecting digital information in said representation defining calcifications in said breast, and

(c) said step of detecting digital information defining masses including applying a spatial domain sharpening filter to said digital representation including the step of defining a subimage by selecting non-adjacent rows of pixels from said digital representation as row elements of said image and selecting non-adjacent columns of pixels from said digital representation as column elements of said subimage.

5. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram defining said human breast with optoelectronic means, including the step of generating a digital representation of said mammogram;

(b) analyzing said information by applying preselected criteria to said information to identify regions of said breast defined by said information having either or both a mass and a calcification, said step of analyzing including detecting digital information in said representation defining masses in said human breast; and

(c) detecting digital information in said representation defining calcifications in said breast, including the step of applying a spatial domain edge detection filter to said digital image.

6. The method of claim 5 wherein said step of applying said spatial domain edge detection filter includes the steps of:

defining a subimage of said digital image which is centered upon a pixel;

calculating a weighted value including the steps of weighting each pixel in said subimage by multiplying each pixel intensity by a fixed mask coefficient, where each fixed mask coefficient is defined along concentric patterns centered upon said center pixel and each pattern has a different fixed coefficient value, and summing all of said weighted pixel values;

assigning said weighted value of said subimage to said center pixel;

moving said subimage such that a different pixel is centered; and

repeating said applying and moving steps until all of said pixels of said digital image have been assigned said weighted value.

7. The method of claim 6 wherein said step of weighting each pixel within the subimage includes weighting the center pixel by multiplying the pixel value by the number 32, weighting all pixels adjacent to the center pixel by multiplying the pixel value by the number -2, and weighting the pixels adjacent to the -2 weighted pixels by multiplying the pixel value of the same by the number -1.

8. The method of claim 5 further including the steps of:

partitioning said filtered digital representation into two or more segments of substantially uniform intensity;

calculating a mean and standard deviation of said intensity in each of said segments;

determining a threshold in each of said segments based upon said calculated means and standard deviation;

segmenting said filtered digital representation by applying said threshold to create a second binary representation.

9. The method of claim 8 further including the step of applying a second preselected criteria to said second binary image by:

determining a cluster density for said calcifications in said second binary image;

comparing determined cluster density to a preselected cluster density criteria;

acquiring information about calcifications which fulfill said second preselected criteria.

10. The method of claim 9 wherein said second preselected cluster density criteria is fulfilled by clusters of two or more calcifications located in a specified area of said digital representation.

11. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram defining said human breast with optoelectronic means, including the step of generating a digital representation of said mammogram;

(b) analyzing said information by applying preselected criteria to said information to identify regions of said breast defined by said information having either or both a mass and a calcification, said step of analyzing including detecting digital information in said representation defining masses in said human breast, and detecting digital information in said representation defining calcifications in said breast;

further including the steps of;

applying a spatial domain sharpening filter to said digital representation;

partitioning said filtered digital representation into two or more segments of substantially uniform intensity;

calculating a means and standard deviation of said intensity in each of said segments;

determining a threshold in each of said segments based upon said calculated means and standard deviation; and

segmenting said filtered digital representation by applying said threshold to create a first binary representation.

12. The method of claim 11 including the step of applying a first preselected criteria to said first binary representation by:

determining an average and maximum width of said masses in said first binary image;

comparing said mass average and maximum width to a preselected width criteria;

determining the height of said masses in said first binary image;

comparing said height of said masses to a preselected height criteria;

determining the area of said masses in said first binary image;

comparing said area of said masses to a preselected area criteria; and

acquiring information about masses which fulfill said first preselected criteria.

13. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information with optoelectronic means from a mammogram defining said human breast; and

thereafter

(b) analyzing said information by applying preselected criteria to said information to identify regions of said breast defined by said information having either or both a mass and a calcification including the steps of;

generating a substantially rectangular window in said digital representation surrounding each of said masses;

measuring the height and width of said substantially rectangular windows;

determining the area of each of said masses;

calculating a perimeter of each of said masses;

calculating a factor having the steps of squaring said perimeter and dividing said squared perimeter by said area of each of said masses;

determining the length of a minor axis and a major axis of each of said masses;

calculating a ratio by dividing said minor axis length by said major axis length for each of said masses;

calculating a first mean intensity of all pixels within said substantially rectangular window;

incrementing the area of said substantially rectangular window by a fixed amount;

calculating a second means intensity of all pixels within said incremented substantially rectangular window;

determining a mean intensity difference by subtracting said first mean from said second mean;

comparing said window height and width and said mass factor, ratio, means intensity difference to a third preselected criteria; and

identifying said masses to be investigated which fulfill said third preselected criteria.

15. The method of claim 13 or 14 wherein said preselected criteria is determined by empirical data and radiologist expertise.

14. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram defining said human breast with optoelectronic means from a mammogram defining said human breast; and

thereafter

(b) analyzing said information by applying preselected criteria to said information to identity regions of said breast defined by said information having either or both a mass and a calcification, including the steps of;

determining clusters of said calcifications;

generating a substantially rectangular window surrounding each of said clusters of said calcifications;

counting the number of said calcifications in each of said substantially rectangular windows;

measuring a length of a major axis and a minor axis of each of said calcifications;

calculating a ratio by dividing said length of said minor axis by said length of said major axis for each of said calcifications;

calculating a density by dividing the area of said substantially rectangular window surrounding each of said clusters by said number of said calcifications within each of said substantially rectangular windows;

calculating a first average intensity of an area within said substantially rectangular window surrounding each of said clusters;

incrementing the area of said substantially rectangular window surrounding each of said clusters by a fixed amount;

calculating a second average intensity within said incremented area of said substantially rectangular window surrounding each of said clusters;

determining an average intensity difference by subtracting said first average intensity from said second average intensity;

comparing said major axis, minor axis, ratio, average intensity difference, and density to a fourth preselected criteria; and

identifying said clusters of said calcifications to be investigated with fulfill said preselected criteria.

16. In a method for assisting the investigation of a human breast, the steps of:

(a) acquiring information from a mammogram of said breast with optoelectronic means, which information defines said breast; and

thereafter

(b) applying preselected criteria to said information to identify regions of said breast defined by said information to be investigated, said step of acquiring information including the step of;

generating a digital representation of said mammogram, and said step of applying preselected criteria includes;

detecting digital information in said representation defining masses in said human breast by applying a spatial domain sharpening filter to said digital representation, and

wherein said step of applying said spatial domain sharpening filter includes the step of defining a subimage by selecting non-adjacent rows of pixels from said digital representation as row elements of said image and selecting non-adjacent columns of pixels from said digital representation as column elements of said subimage.

17. In a method for assisting the investigation of a human breast, the steps of:

(a) acquiring information from a mammogram of said breast with optoelectronic means, which information defines said breast; and

thereafter p1 (b) apply preselected criteria to said information to identify regions of said breast defined by said information to be investigated, said step of acquiring information including the step of;

generating a digital representation of said mammogram, and said step of applying preselected criteria including the step of;

detecting digital information in said representation defining masses in said human breast by applying a spatial domain sharpening filter to said digital representation, further including the steps of;

partitioning said filtered digital representation into two or more segments of substantially uniform intensity;

calculating a mean and standard deviation of said intensity in each of said segments;

determining a threshold in each of said segments based upon said calculated mean and standard deviation; and

segmenting said filtered digital representation by applying said threshold to create a first binary representation.

18. The method of claim 17 further including the step of applying a first preselected criteria to said first binary representation by:

determining an average and maximum width of said masses in said first binary image;

comparing said mass average and maximum width to a preselected width criteria;

determining the height of said masses in said first binary image;

comparing said height of said masses to a preselected height criteria;

determining the area of said masses in said first binary image;

comparing said area of said masses to a preselected area criteria; and

acquiring information about masses which fulfill said first preselected criteria.

19. In a method for assisting the investigation of a human breast, the steps of:

(a) acquiring information from a mammogram of said breast with optoelectronic means, which information defines said breast; and

thereafter

(b) applying preselected criteria to said information to identify regions of said breast defined by said information to be investigated, said step of acquiring information including the step of;

generating a digital representation of said mammogram; and

wherein said step of applying preselected criteria to said information includes the steps of;

generating a substantially rectangular window in said digital representation surrounding each of said masses determined in said analyzing step;

measuring the height and width of said substantially rectangular windows;

determining the area of each of said masses;

calculating a perimeter of each of said masses;

calculating a factor having the steps of squaring said perimeter and dividing said squared perimeter by said area of each of said masses;

determining the length of a minor axis and a major axis of each of said masses;

calculating a ratio by dividing said minor axis length by said major axis length for each of said masses;

calculating a first mean intensity of all pixels within said substantially rectangular window;

incrementing the area of said substantially rectangular window by a fixed amount;

calculating a second means intensity of all pixels within said incremented substantially rectangular window;

determining a means intensity difference by subtracting said first means from said second means;

comparing said window height and width and said mass factor, ratio, mean intensity difference to a third preselected criteria; and

identifying said masses to be investigated which fulfill said third preselected criteria.

20. The method of claim 19 wherein said preselected criteria is determined by empirical data and radiologist expertise.

21. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram of said human breast with optoelectronic means, including the step of generating a digital image of said mammogram;

(b) applying preselected criteria to said information to identify regions of said breast defined by said information to be investigated including the step of detecting information in said image defining calcifications in said human breast and applying a spatial domain edge detection filter to said digital image, and;

said step of applying said spatial edge detection filter includes the steps of;

defining a subimage of said digital image which is centered upon a pixel;

calculating a weighted sum value including the steps of weighting each pixel in said subimage by multiplying each pixel intensity by a fixed mask coefficient, where each fixed mask coefficient is defined along concentric patterns centered upon said center pixel and each pattern has a different fixed coefficient value, and summing all of said weighted pixel values;

assigning said weighted sum value of said subimage to said center pixel;

moving said subimage such that a different pixel is centered; and

repeating said applying and moving steps until all of said pixels of said digital image have been assigned said weighted sum value.

22. The method of claim 21 wherein said step of weighting each pixel within the subimage includes weighting the center pixel with the number 32, weighting all pixels adjacent to the center pixel with the number -2, and weighting the pixels adjacent to the -2 weighted pixels with the number -1.

23. In a method for assisting the investigation of a human breast for malignancies, the steps of:

(a) acquiring information from a mammogram of said human breast with optoelectronic means, including the step of generating a digital image representation of said mammogram;

(b) applying preselected criteria to said information to identify regions of said breast defined by said information to be investigated including the step of detecting information in said image defining calcifications in said human breast and applying a spatial domain edge detection filter to said digital image representation, further including the steps of;

partitioning said filtered digital representation into two or more segments of substantially uniform intensity;

calculating a mean and standard deviation of said intensity in each of said segments;

determining a threshold in each of said segments based upon said calculated mean and standard deviation; and

segmenting said filtered digital representation by applying said threshold to create a first binary representation.

24. The method of claim 23 further including the step of applying a second preselected criteria to said first binary image by:

determining a cluster density for said calcifications in said second binary image;

comparing determined cluster density to a preselected cluster density criteria; and

acquiring information about calcifications which will fulfill said selected criteria.

25. The method of claim 24 wherein said second preselected cluster density criteria is fulfilled by clusters of two or more calcifications located in a specified area of said digital representation.

26. In a method for assisting the investigation of a human breast for malignancy, the steps of:

(a) acquiring information from a mammogram of said human breast with optoelectronic means; and

thereafter

(b) analyzing said information by applying preselected criteria to said information to identify regions of said breast having calcifications, said step including the steps of;

determining clusters of said calcifications;

generating a substantially rectangular window surrounding each of said clusters of said calcifications;

counting the number of said calcifications in each of said substantially rectangular windows;

measuring a length of a major axis and a minor axis of each of said calcifications;

calculating a ratio by dividing said length of said minor axis by said length of said major axis for each of said calcifications;

calculating a density by dividing the area of said substantially rectangular window surrounding each of said clusters by said number of said calcifications within each of said substantially rectangular windows;

calculating a first average intensity of an area within said substantially rectangular window surrounding each of said clusters;

incrementing the area of said substantially rectangular window surrounding each of said clusters by a fixed amount;

calculating a second average intensity within said incremented area of said substantially rectangular window surrounding each of said clusters;

determining an average intensity difference by subtracting said first average intensity from said second average intensity;

comparing said major axis, minor axis, ratio, average intensity difference, and density to a fourth preselected criteria; and

identifying said clusters of said calcifications which fulfill said preselected criteria to be investigated.

27. The method of claim 26 wherein said preselected criteria is determined by empirical data and radiologist expertise.