Methods, systems, and apparatuses for quantitative analysis of heterogeneous biomarker distribution

Patent metadata

Field	Value
Publication number	US-11049247-B2
Application number	US-202016811688-A
Country	US
Kind code	B2
Filing date	Mar 6, 2020
Priority date	Dec 3, 2014
Publication date	Jun 29, 2021
Grant date	Jun 29, 2021

Abstract

Methods, systems, and apparatuses for detecting and describing heterogeneity in a cell sample are disclosed herein. A plurality of fields of view (FOV) are generated for one or more areas of interest (AOI) within an image of the cell sample are generated. Hyperspectral or multispectral data from each FOV is organized into an image stack containing one or more z-layers, with each z-layer containing intensity data for a single marker at each pixel in the FOV. A cluster analysis is applied to the image stacks, wherein the clustering algorithm groups pixels having a similar ratio of detectable marker intensity across layers of the z-axis, thereby generating a plurality of clusters having similar expression patterns.

First claim

The invention claimed is:

1. A method of generating a dataset for use in a cluster analysis comprising:
(a) computing a field of view (FOV) sampling grid for each of a plurality of areas of interest (AOI) within obtained image data;
(b) acquiring multi-spectral data and/or hyper-spectral data at single or multiple z-planes for each computed FOV;
(c) unmixing signal data corresponding to one or more detectable markers from the collected multi-spectral data and/or hyper-spectral data for each computed FOV;
(d) identifying FOVs to be compared as a group in the cluster analysis; and
(e) storing selected features of interest of the identified FOVs in the dataset, wherein the features of interest are selected by automatically segmenting the unmixed signals corresponding to the one or more detectable marker signals in each of the selected FOVs.

2. The method of claim 1 , wherein the FOV sampling grid for each of the plurality of AOIs is calculated by generating a regularly spaced grid of FOVs.

3. The method of claim 1 , wherein the FOV sampling grid for each of the plurality of AOIs is calculated by assigning an FOV to one or more regions within each AOI having a pre-determined set of morphological features.

4. The method of claim 3 , wherein the one or more regions are tumorous regions.

5. The method of claim 1 , wherein the unmixed signals corresponding to the one or more detectable marker signals in each of the selected FOVs are automatically segmented based on morphometric properties.

6. The method of claim 1 , wherein the unmixed signals corresponding to the one or more detectable marker signals in each of the selected FOVs are automatically segmented based on photometric properties.

7. The method of claim 1 , wherein the selected features are stored in a nested data structure or database together with one or more metadata attributes selected from the group consisting of patient information, assay information, biopsy information, section information, AOI position information, and FOV position information.

8. The method of claim 1 , wherein the cluster analysis comprises applying an unsupervised, non-parametric, density-based clustering algorithm to the dataset.

9. The method of claim 8 , wherein the unsupervised, non-parametric, density-based clustering algorithm is a Mean-Shift clustering algorithm.

10. The method of claim 8 , further comprising generating an expression cluster map comprising a plurality of clusters, wherein each pixel in the expression cluster map is grouped into one cluster of the plurality of clusters, and where each cluster comprises a different expression pattern for each of the detectable markers.

11. The method of claim 1 , wherein:
(a4a) the FOVs identified to be compared as a group correspond to different tumor foci in the same tissue section; or
(a4b) the FOVs identified to be compared as a group are determined on the basis of a biopsy taken from the same patient for comparison to a different biopsy taken from the same patient; or
(a4c) the FOVs identified to be compared as a group are determined on the basis of tumor location; or
(a4d) the FOVs identified to be compared as a group are determined based on the patient for comparison to another patient; or
(a4e) the FOVs identified to be compared as a group are determined on the basis of tumor genotype.

12. The method of claim 1 , wherein at least one of the one or more detectable markers is coupled to at least one antibody that specifically binds to at least one phosphorylated protein.

13. The method of claim 12 , wherein the at least one phosphorylated protein is a member of a PI-3 kinase signal transduction pathway or a MAP kinase signal transduction pathway.

14. The method of claim 13 , wherein the at least one phosphorylated protein is selected from the group consisting of AKT, PRAS40, S6, EIF4G, and ERK1/2.

15. A method of characterizing a biological sample according to a phosphorylation profile comprising:
(a) obtaining image data derived from a biological sample, wherein the biological sample is stained for the presence of at least a phosphorylated form of a signal transduction protein;
(b) generating a field of view (FOV) image stack for each of a plurality of FOVs within one or more areas of interest (AOI) within the obtained image data;
(c) deriving a 3D tiled image that represents an entire AOI of the biological sample, wherein the 3D tiled image is derived from the generated FOV image stack; and
(d) applying an unsupervised, non-parametric, density-based clustering algorithm to the 3D tiled image to calculate an x,y array, where the x and y coordinates of the x,y array are the spatial coordinates of the 3D tiled image stack, and the value at each x,y position is a label that indicates the cluster to which a given pixel belongs.

16. The method of claim 15 , further comprising computing an expression cluster map comprising a plurality of clusters, wherein each pixel in the expression cluster map is grouped into one cluster of the plurality of clusters, and where each cluster comprises a different expression pattern for each of the detectable markers.

17. The method of claim 15 , wherein the at least one phosphorylated form of the signal transduction protein is a member of a PI-3 kinase signal transduction pathway or a MAP kinase signal transduction pathway.

18. The method of claim 17 , wherein the at least one phosphorylated form of the signal transduction protein is selected from the group consisting of AKT, PRAS40, S6, EIF4G, and ERK1/2.

19. The method of claim 15 , further comprising characterizing the biological sample according to an activation state of the phosphorylated form of the signal transduction protein.

20. The method of claim 15 , the unsupervised, non-parametric, density-based clustering algorithm is a Mean-Shift clustering algorithm.

Assignees

• Ventana Med Syst Inc
• Univ Oregon Health & Science

Inventors

• Barnes Michael
• Chafin David
• Garsha Karl
• Grogan Thomas M
• Roberts Esteban
• Stevens Benjamin
• Ventura Franklin
• Chefd'Hotel Christophe
• Shanmugam Kandavel
• Gray Joe
• Ramunno-Johnson Damien
• Vu Tothu Tania
• Druker Brian J
• Jacob Thomas

Classifications

• G06F18/23211

  with adaptive number of clusters · CPC title

• G06F18/2321

  using statistics or function optimisation, e.g. modelling of probability density functions · CPC title

• G06F18/213

  Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title

• G06V10/56

  relating to colour · CPC title