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

Patent metadata

Field	Value
Publication number	US-12579642-B2
Application number	US-202318139632-A
Country	US
Kind code	B2
Filing date	Apr 26, 2023
Priority date	Dec 3, 2014
Publication date	Mar 17, 2026
Grant date	Mar 17, 2026

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 imaging a cell sample comprising, the method comprising:
i. acquiring image data with a multi-spectral imaging system from the cell sample, the acquired image data comprising multi-spectral pixels; ii. spectrally unmixing the multi-spectral pixels providing a number of ‘n’ single channel images; and iii. performing a cluster analysis comprising applying an unsupervised, non-parametric, density-based clustering algorithm in the n+2 dimensional space given by the ‘n’ single channel images, thereby generating a plurality of clusters displaying a representation of a clustering result obtained by the clustering.

2 . The method of claim 1 , further comprising displaying on a display device at least one of the ‘n’ single channel images and displaying the representation of the clustering result as an overlay on the at least one of the ‘n’ single channel images.

3 . The method of claim 1 , wherein the number ‘n’ of single channel images is greater than 3 or greater than 4.

4 . The method of claim 1 , wherein the spectrally unmixed image data is low pass filtered before performing the clustering algorithm.

5 . The method of claim 1 , wherein the spectrally unmixed image data is binned before performing the clustering algorithm.

6 . The method of claim 5 , wherein the displayed representation has the resolution of the unbinned image data.

7 . A method of characterizing heterogeneity in a cell sample comprising one or more analytes labeled with a detectable marker, the method comprising analyzing an image of the cell sample on a computer apparatus comprising a computer processor programmed to apply a cluster analysis to a dataset obtained from the image of the cell sample to create a cluster map comprising a plurality of clusters of expression patterns, wherein:
(a) the dataset comprises an image stack for at least a portion of the image of the cell sample, wherein the image stack comprises a x-axis, a y-axis, and a z-axis, wherein the x-axis and the y-axis represent spatial coordinates within the portion of the image, and the z-axis comprises a number of ‘n’ layers, wherein each layer of the z axis comprises intensity data for a single detectable marker at a plurality of x,y coordinates; (b) the cluster analysis comprises applying an unsupervised, non-parametric, density-based clustering algorithm to the image stack, wherein the clustering algorithm clusters points (defined by the dataset in an at least n+2 dimensional space, where each point is given by x,y coordinates and ‘n’ intensity data values from the n layers at the respective x,y coordinate in the image stack, thereby generating the plurality of clusters, (c) outputting of output data being representative of the result of the cluster analysis, the data being indicative of the heterogeneity in the cell sample, wherein the image is a multi-channel image and wherein the ‘n’ layers are obtained by unmixing the multi-channel image.

8 . The method of claim 7 , wherein the outputting of the data is performed by displaying the multi-channel image and/or one or more single channel images, wherein a single channel image is given by displaying one of the n layers, by visualizing the delimitations of the clusters in the multi-channel image and/or the at least one single channel image.

9 . The method of claim 7 , further comprising generating a cluster histogram that indicates the proportional area for each cluster in the dataset, the cluster histogram being the data that is output in step (c).

10 . The method of claim 7 , further comprising binning the points of the n+2 dimensional space for low pass filtering the image stack and reducing the xy resolution to provide a binned image stack, wherein the cluster analysis in step b is performed on the binned image stack.

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

• G06V10/56

  relating to colour · CPC title

• G06V10/25

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

• G06V20/698

  Matching; Classification · CPC title