Computer Diagnosis of Malignancies and False Positives

• US 20080107323A1
• Filed: 10/23/2007
• Published: 05/08/2008
• Est. Priority Date: 10/25/2006
• Status: Active Grant

• Current Assignee: Siemens Computer Aided Diagnosis Ltd. (Siemens AG)
• Original Assignee: Siemens Computer Aided Diagnosis Ltd. (Siemens AG)
• Inventors: Bamberger, Philippe, Ratner, Eliahu
• Granted Patent: US 8,238,637 B2
• Time in Patent Office: Days
• US Class Current: 382/132
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/30 Determination of transform parameters for the alignment of images, i.e. image registration

Abstract

A method for diagnosing malignancy of suspect regions in a 2D imaging projection of a body organ of a subject comprising the steps of: normalizing the image projection to map regions thereof onto a standard shape; extracting location of regions mapped on the standard shape and using normalized position of a suspect region to assess likelihood that it is malignant.

Claims

1. A method for diagnosing malignancy of suspect regions in a 2D imaging projection of a body organ of a subject comprising the steps of:

normalizing the image projection to map regions thereof onto a standard shape;

extracting location of regions mapped on the standard shape and using normalized position of a suspect region to assess likelihood that it is malignant.

2. The method of claim 1 wherein two images of the body organ are taken from different angles, to allow extraction of 3 dimensional data for locating position of suspect region in normalized 3D space.

3. The method of claim 1, wherein the body organ is a mammary gland.

4. The method of claim 1 wherein the suspect region is suspected as being cancerous.

5. The method of claim 1 wherein the 2D imaging projection is a digital X-ray image.

6. The method of claim 1 wherein the assessment of likelihood of malignancy of a region comprises comparing location of suspected malignancy with statistical data representing a sample set of a population.

7. The method of claim 6 wherein the statistical data comprises a pre-established probability mapping of suspected malignancies projected onto the standardized shape.

8. The method of claim 7 wherein the malignancy probability mapping is created by a method comprised of the following steps:

providing a training set of digital images with regions of known malignancy status;

determining the normalized coordinates of each region in the training set of digital images;

dividing each normalized digital image into a pre-determined number of cells;

calculating the malignancy occurrence frequency for each cell; and

generating a malignancy probability map based on the frequency of occurrence in each cell.

9. The method of claim 8 wherein the organ is generally one of a pair and the digital images comprises images of left and right organs of organs occurring in pairs and the probability mapping comprises laterally reflecting images to pool data from both sides of the body.

10. The method of claim 6 wherein the statistical data comprises a probability mapping of false positives onto the standardized shape.

11. The method of claim 10, wherein the probability mapping of false positives are created by the method comprised the steps of:

providing a training set of digital images featuring detected false positives;

determining the normalized coordinates of each false positive in each digital images within the training set;

dividing each of the digital images into a pre-determined number of cells and, based on the normalized coordinates of all the false candidates in the set of images, calculating the false object occurrence frequency for each cell; and

generating a false positive probability map based on the frequency of occurrence of false positives in each cell.

12. The method of claim 6 wherein the sample set is selected from the list comprising:

a large random population, a large same sex population, a population within a specific age range, a population sharing genetic similarities, and an ethnic population.

13. The method of claim 6 wherein the sample set is a set of one and the method comprises comparing the normalized image of the body organ mapped onto a single normalized reference image.

16. The method of claim 13 wherein one of said normalized images is laterally reflected to cancel effect of asymmetry of the organ.

18. The method of claim 1 wherein the step of normalizing comprises identifying positions of a suspect region in polar coordinates.

19. The method of claim 18 further comprises converting the polar coordinates into Cartesian coordinates.

20. The method according to claim 18, wherein the step of normalizing the coordinates in polar coordinates further comprises the steps of:

segmenting the projection image of the body organ from its background;

defining a unit vector from an origin point in the digital image to any tissue contour point resulting from said segmenting and

determining the origin point in the digital image and an extreme point therein, thereby forming a zero angle vector;

normalizing vectors of all suspect regions in terms of polar angles based on the zero angle vector and on the unit vector crossing each suspect region.

21. The method of claim 20 further including the step of smoothing the probability maps.

14. The method of clam 13 wherein the reference image is a previously obtained image of the same organ.

15. The method of clam 13 wherein the organ is one of a pair and the normalized image of the body organ is compared with a normalized image of the paired organ.

17. The method of clam 13 wherein the reference image is an image of an equivalent organ for a related subject.

22. A malignancy probability mapping of a projection image of a body organ, created by the method comprising the steps of:

providing a training set of digital projection images of body organs including suspect regions of known malignancy status;

determining normalized coordinates of each suspect region in the training set of digital projection images;

dividing each normalized digital projection image into a pre-determined number of cells;

calculating the malignancy occurrence frequency for each cell; and

generating a malignancy probability map based on the frequency of occurrence in each cell.

23. The method of claim 22 wherein the step of determining normalized coordinates comprises segmenting the projection image of the body organ from its background;

defining a unit vector from an origin point in the digital image to any tissue contour point resulting from said segmenting and

determining the origin point in the digital image and an extreme point therein, thereby forming a zero angle vector;

normalizing vectors of all suspect regions in terms of polar angles based on the zero angle vector and on the unit vector crossing each suspect region.

25. The method of claim 22 wherein the step of determining normalized coordinates comprises segmenting the projection image of the body organ from its background;

defining a unit vector from an origin point in the digital image to any tissue contour point resulting from said segmenting and

determining the origin point in the digital image and an extreme point therein, thereby forming a zero angle vector;

normalizing vectors of all false positives in terms of polar angles based on the zero angle vector and on the unit vector crossing each false positive.

24. A false positive probability mapping of a projection image of a body organ, fabricated by the method comprising the steps of:

providing a training set of digital projection images of body organs including suspect regions recognized as being false positives;

determining normalized coordinates of each false positive in the training set of digital projection images;

dividing each normalized digital projection image into a pre-determined number of cells;

calculating the false positive frequency for each cell; and

generating a false positive probability map based on the frequency of occurrence in each cell.