Method and apparatus for automatic cancer diagnosis scoring of tissue samples

• US 9,652,844 B2
• Filed: 03/18/2015
• Issued: 05/16/2017
• Est. Priority Date: 07/13/2012
• Status: Active Grant

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Zhang, Ximin, Liu, Ming-Chang
• Primary Examiner(s): Liew, Alex
• Application Number:
  US14/661,641
• Publication Number:
  US 20160189370A1
• Time in Patent Office: 790 Days
• Field of Search: 382128-134, 382192-194, 382221, 382271
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/10056 Microscopic image
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 10/462 Salient features, e.g. scale invariant feature transforms [SIFT]
• CPC Class Codes: G06V 20/69 Microscopic objects, e.g. biological cells or cellular parts
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

Certain aspects of an apparatus and method for automatic ER/PR scoring of tissue samples may include for determining a cancer diagnosis score comprising identifying a positive stained nucleus in a slide image of the tissue sample, identifying a negative stained nucleus in the slide image, computing a proportion score based on number of the positive stained nucleus identified and number of the negative stained nucleus identified and determining the cancer diagnosis score based on the proportion.

Claims

1. A non-transitory computer-readable storage medium storing one or more computer programs adapted to cause a processor based system to execute steps comprising:

converting at least a part of image information of a tissue sample to an optical domain image;

separating the at least a part of image information into positive stain image information and negative stain image information;

identifying a number of positive stained nuclei based on the positive stain image information and a number of negative stained nuclei based on the negative stain image information;

calculating a number of total nuclei by summing the number of positive stained nuclei and the number of negative stained nuclei, wherein a percentage score is calculated based on the number of positive stained nuclei and the number of total nuclei; and

displaying an indication of a cancer diagnosis score derived at least in part from the percentage score.

2. The non-transitory computer-readable storage medium of claim 1, wherein the image information includes RBG information.

3. The non-transitory computer-readable storage medium of claim 2, wherein the step of converting comprises:

transforming the at least the part of image information from an RGB domain to an optical domain.

4. The non-transitory computer-readable storage medium of claim 1, further comprising:

accepting one or more signals from a user input device to define a user-selected interest area within the image information of the tissue sample.

5. The non-transitory computer-readable storage medium of claim 1, wherein the step of converting comprises:

transforming the at least the part of image information to an optical domain.

6. The non-transitory computer-readable storage medium of claim 1, wherein the step of converting comprises:

calculating at least an optical density value for a positive stain based on the at least the part of image information.

7. The non-transitory computer-readable storage medium of claim 1, wherein the step of converting comprises:

separating the at least the part of image information into the positive stain image information and the negative stain image information.

8. The non-transitory computer-readable storage medium of claim 7, wherein the one or more computer programs are further adapted to cause the processor based system to execute steps comprising:

identifying a number of positive stained nuclei based on the positive stain image information and a number of negative stained nuclei based on the negative stain image information; and

calculating a number of total nuclei by summing the number of positive stained nuclei and the number of negative stained nuclei, wherein a proportion score is calculated based on the number of positive stained nuclei and the number of total nuclei.

9. The non-transitory computer-readable storage medium of claim 1, wherein the one or more computer programs are further adapted to cause the processor based system to execute steps comprising:

calculating a positive stained nuclei percentage based on a ratio of positive stained nuclei identified based on the positive stain image information to total nuclei included in the at least the part of image information, wherein a proportion score is calculated based on the positive stained nuclei percentage.

10. The non-transitory computer-readable storage medium of claim 9, wherein the one or more computer programs are further adapted to cause the processor based system to execute steps comprising:

determining an average intensity of positive stain in the positive stain image information, wherein the cancer diagnosis score is determined by summing the proportion score and an intensity score identified based on the average intensity.

11. The non-transitory computer-readable storage medium of claim 10, wherein the cancer diagnosis score includes an estrogen receptor or progesterone receptor (ER/PR) score.

12. The non-transitory computer-readable storage medium of claim 9, wherein

in a case that the positive stained nuclei percentage is more than 2/3, the proportion score is assigned to 5,

in a case that the positive stained nuclei percentage is more than 2/1, the proportion score is assigned to 4,

in a case that the positive stained nuclei percentage is more than 10%, the proportion score is assigned to 3,

in a case that the positive stained nuclei percentage is more than 1%, the proportion score is assigned to 2, and

in a case that the positive stained nuclei percentage is more than 0%, the proportion score is assigned to 1.

13. The non-transitory computer-readable storage medium of claim 2, wherein the one or more computer programs are further adapted to cause the processor based system to execute steps comprising:

selecting a salient region of the image information as the at least the part of image information by finding a region which is rich in both red and blue channels.

14. The non-transitory computer-readable storage medium of claim 8, wherein the positive stain image information and the negative stain image information are analyzed to obtain a size range parameter based on a proportion of positive stained pixels with pixel intensity below a predetermined positive threshold and negative stained pixels with pixel intensity below a predetermined negative threshold.

15. The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying the number of negative stained nuclei comprises:

applying a window averaging procedure to each pixel in the negative stain image information based on the size range parameter;

applying an inverse process on each pixel to obtain an inverse pixel and assigning the pixel value to the inverse pixel value;

determining a pixel to be a nucleus center if the pixel value is greater than a predetermined negative nucleus threshold; and

excluding a pixel as a negative stained nucleus center if the pixel value is equal or less than the predetermined negative nucleus threshold.

16. The non-transitory computer-readable storage medium of claim 15, wherein the step of applying the inverse process further comprises:

receiving as an input a window averaging of the intensity of the pixels of each of the positive and negative stain image information.

17. The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying the number of negative stained nuclei comprises:

filtering the negative stain image information using a low pass filter to generate a filtered negative stain image information;

applying a binary image generation process on each pixel of the filtered negative stain image information to produce a binary image information;

applying a distance transform to the binary image information to find a minimum distance between white and black pixels by generating a distance image information with the minimum distance as a pixel value; and

generating a new image information where each pixel value is the sum of the value at a center pixel of the negative stained nucleus and distance image pixel.

18. The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying the number of positive stained nuclei comprises:

applying a window averaging procedure to each pixel in the positive stain image information based on the size range parameter;

applying an inverse process on each pixel to obtain an inverse pixel and assigning the pixel value to the inverse pixel value

determining a pixel to be a nucleus center when the pixel value is greater than a positive nucleus threshold; and

excluding a pixel as a nucleus center when the pixel value is equal or less than the positive nucleus threshold.

19. The non-transitory computer-readable storage medium of claim 18, wherein the positive nucleus threshold is based on an average intensity of positive stain.