System and method for detecting gene expression levels in tissue samples

Patent metadata

Field	Value
Publication number	US-12175658-B2
Application number	US-201916714890-A
Country	US
Kind code	B2
Filing date	Dec 16, 2019
Priority date	Feb 1, 2012
Publication date	Dec 24, 2024
Grant date	Dec 24, 2024

Abstract

A computer-based specimen analyzer ( 10 ) is configured to detect a level of expression of genes in a cell sample by detecting dots that represent differently stained genes and chromosomes in a cell. The color of the stained genes and the chromosomes is enhanced and filtered to produce a dot mask that defines areas in the image that are genes, chromosomes, or non-genetic material. Metrics are determined for the dots and/or pixels in the image of the cell in areas corresponding to the dots. The metrics are fed to a classifier that separates genes from chromosomes. The results of the classifier are counted to estimate the expression level of genes in the tissue samples.

First claim

The invention claimed is:

1. A computer system for detecting an expression level of genes in a tissue sample, the computer system comprising:
one or more data processors; and
a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform one or more operations including:
receiving an image of a tissue sample prepared using an in situ hybridization technique, wherein the image is a first image in an RGB color space;
processing the image to generate a set of ratio values corresponding to one or more candidate regions of the image, wherein processing the image comprises:
selecting the one or more candidate regions from the image, wherein a candidate region of the one or more candidate regions corresponds to a nucleus of a cell in the tissue sample;
for each candidate region of the one or more candidate regions:
detecting a plurality of image dots from the candidate region by converting the image to a second image in a L*a*b color space and computing a linear combination of L, a, and b values for each pixel of the second image;
 for each image dot of the plurality of image dots:
  identifying morphological metrics corresponding to the image dot; and
  processing the morphologic metrics of the image dot using a classifier to generate a result, the result indicating whether the image dot corresponds to a gene or a chromosome; and
computing a ratio value between a first count of image dots corresponding to the gene and a second count of image dots corresponding to the chromosome; and

aggregating the ratio value of the each candidate region of the one or more candidate regions to generate the set of ratio values; and

outputting the set of ratio values as a result that is indicative of a gene-expression level associated with the tissue sample.

2. The computer system of claim 1 , wherein each image dot of the plurality of image dots corresponds to an in-situ hybridization signal that indicates whether the image dot corresponds to the gene or the chromosome.

3. The computer system of claim 1 , wherein identifying the morphological metrics corresponding to the image dot includes evaluating at least one of an image-dot shape, an image-dot orientation, a first spatial relationship between the image dot and other image dots of the plurality of image dots, or a second spatial relationship between the image dot and another anatomical structure of the tissue sample.

4. The computer system of claim 1 , wherein the set of ratio values further indicates whether the gene is over-expressed in the tissue sample.

5. The computer system of claim 1 , wherein the one or more operations further include:
filtering the second image with a set of Difference of Gaussian filters to generate a set of filtered images;
generating a combined-filtered image based on the set of filtered images; and
processing the combined-filtered image to generate a dot-mask image, wherein the dot-mask image facilitates detection of the plurality of image dots for the candidate region.

6. The computer system of claim 1 , wherein the morphological metrics of the image dot are identified by analyzing a set of pixels within an area of the image dot.

7. A computer-implemented method of detecting an expression level of genes in a tissue sample, the method comprising:
receiving an image of a tissue sample prepared using an in situ hybridization technique, wherein the image is a first image in an RGB color space;
processing the image to generate a set of ratio values corresponding to one or more candidate regions of the image, wherein processing the image comprises:
selecting the one or more candidate regions from the image, wherein a candidate region of the one or more candidate regions corresponds to a nucleus of a cell in the tissue sample;
for each candidate region of the one or more candidate regions:
detecting a plurality of image dots from the candidate region by converting the image to a second image in a L*a*b color space and computing a linear combination of L, a, and b values for each pixel of the second image;
for each image dot of the plurality of image dots:
 identifying morphological metrics corresponding to the image dot; and
 processing the morphologic metrics of the image dot using a classifier to generate a result, the result indicating whether the image dot corresponds to a gene or a chromosome; and

computing a ratio value between a first count of image dots corresponding to the gene and a second count of image dots corresponding to the chromosome; and

aggregating the ratio value of the each candidate region of the one or more candidate regions to generate the set of ratio values; and

outputting the set of ratio values as a result that is indicative of a gene-expression level associated with the tissue sample.

8. The computer-implemented method of claim 7 , wherein each image dot of the plurality of image dots corresponds to an in-situ hybridization signal that indicates whether the image dot corresponds to the gene or the chromosome.

9. The computer-implemented method of claim 7 , wherein identifying the morphological metrics corresponding to the image dot includes evaluating at least one of an image-dot shape, an image-dot orientation, a first spatial relationship between the image dot and other image dots of the plurality of image dots, or a second spatial relationship between the image dot and another anatomical structure of the tissue sample.

10. The computer-implemented method of claim 7 , wherein the set of ratio values further indicates whether the gene is over-expressed in the tissue sample.

11. The computer-implemented method of claim 7 , further comprising:
filtering the second image with a set of Difference of Gaussian filters to generate a set of filtered images;
generating a combined-filtered image based on the set of filtered images; and
processing the combined-filtered image to generate a dot-mask image, wherein the dot-mask image facilitates detection of the plurality of image dots for the candidate region.

12. The computer-implemented method of claim 7 , wherein the morphological metrics of the image dot are identified by analyzing a set of pixels within an area of the image dot.

13. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform one or more operations including:
receiving an image of a tissue sample prepared using an in situ hybridization technique, wherein the image is a first image in an RGB color space;
processing the image to generate a set of ratio values corresponding to one or more candidate regions of the image, wherein processing the image comprises:
selecting the one or more candidate regions from the image, wherein a candidate region of the one or more candidate regions corresponds to a nucleus of a cell in the tissue sample;
for each candidate region of the one or more candidate regions:
detecting a plurality of image dots from the candidate region by converting the image to a second image in a L*a*b color space and computing a linear combination of L, a, and b values for each pixel of the second image;
for each image dot of the plurality of image dots:
 identifying morphological metrics corresponding to the image dot; and
 processing the morphologic metrics of the image dot using a classifier to generate a result, the result indicating whether the image dot corresponds to a gene or a chromosome; and

computing a ratio value between a first count of image dots correspondin

Assignees

• Ventana Med Syst Inc

Inventors

• Bamford Pascal
• Chukka Srinivas
• Martin Jim F
• Sarkar Anindya
• Sertel Olcay
• Suzue Ellen
• Varangaonkar Harshal

Classifications

• C12Q1/6813

  Hybridisation assays · CPC title

• G06V10/56

  relating to colour · CPC title

• G06V20/698

  Matching; Classification · CPC title

• G06T7/90

  Determination of colour characteristics · CPC title

• G06T2207/10056

  Microscopic image · CPC title