Method and device for analyzing pathological slide image

Patent metadata

Field	Value
Publication number	US-12530874-B2
Application number	US-202318178233-A
Country	US
Kind code	B2
Filing date	Mar 3, 2023
Priority date	Mar 3, 2022
Publication date	Jan 20, 2026
Grant date	Jan 20, 2026

Abstract

Provided is a computing device including at least one memory, and at least one processor configured to obtain a first pathological slide image one of a first object and biological information of the first object, generate training data by using at least one first patch included in the first pathological slide image, and the biological information, train a first machine learning model based on the training data, and analyze a second pathological slide image of a second object by using the trained first machine learning model.

First claim

What is claimed is:

1 . A computing device comprising:
at least one memory; and at least one processor configured to:
obtain a first pathological slide image of a first object, and spatial transcriptomics information of the first object,
generate training data including type information of at least one cell expressed in at least one first patch included in the first pathological slide image by using the at least one first patch and the spatial transcriptomics information, wherein the type information of the at least one cell is obtained based on gene expression information corresponding to the at least one first patch,
train a first machine learning model based on the training data to predict the type information of the at least one cell, and analyze a second pathological slide image of a second object by using the trained first machine learning model.

2 . The computing device of claim 1 , wherein the gene expression information includes gene expression information for each of a plurality of regions respectively associated with the at least one cell, and
the at least one processor is further configured to train the first machine learning model by using the training data as ground-truth data.

3 . The computing device of claim 2 , wherein the at least one processor is further configured to generate the second machine learning model configured to identify the type information of the at least one cell included in the first object, by adding or removing at least one layer in the trained first machine learning model.

4 . The computing device of claim 1 , wherein the at least one processor is further configured to predict a therapeutic reaction of a subject corresponding to the second pathological slide image by using spatial transcriptomics information of the second object expressed in the second pathological slide image.

5 . The computing device of claim 4 , wherein the predicting of the therapeutic reaction is performed by a third machine learning model, and
the spatial transcriptomics information of the second object comprises at least one of spatial transcriptomics information obtained by the trained first machine learning model and separately obtained spatial transcriptomics information.

6 . The computing device of claim 5 , wherein the third machine learning model is trained to predict the therapeutic reaction, by using a feature vector extracted from at least one layer included in the trained first machine learning model.

7 . The computing device of claim 5 , wherein the third machine learning model is trained to predict the therapeutic reaction, by using gene expression information included in the spatial transcriptomics information of the second object and position information corresponding to the gene expression information.

8 . The computing device of claim 1 , wherein the at least one processor is further configured to:
receive an user input for selecting at least one cell for correcting a labeling of the at least one cell; update the labeling of the at least one cell, based on the user input; and train the first machine learning model based on the updated labeling.

9 . A method of analyzing a pathological slide image, the method comprising:
obtaining a first pathological slide image of a first object and spatial transcriptomics information of the first object; generating training data including type information of at least one cell expressed in at least one first patch included in the first pathological slide image by using the at least one first patch and the spatial transcriptomics information, wherein the type information of the at least one cell is obtained based on gene expression information corresponding to the at least one first patch; training a first machine learning model based on the training data to predict the type information of the at least one cell; and analyzing a second pathological slide image of a second object by using the trained first machine learning model.

10 . The method of claim 9 , wherein the gene expression information includes gene expression information for each of a plurality of regions respectively associated with the at least one cell, and the training comprises training the first machine learning model by using the training data as ground-truth data.

11 . The method of claim 10 , wherein the training further comprises generating the second machine learning model configured to identify the type information of the at least one cell included in the first object, by adding or removing at least one layer in the trained first machine learning model.

12 . The method of claim 9 , further comprising predicting a therapeutic reaction of a subject corresponding to the second pathological slide image by using spatial transcriptomics information of the second object expressed in the second pathological slide image.

13 . The method of claim 12 , wherein the predicting of the therapeutic reaction is performed by a third machine learning model, and
the spatial transcriptomics information of the second object comprises at least one of spatial transcriptomics information obtained by the trained first machine learning model, and separately obtained spatial transcriptomics information.

14 . The method of claim 13 , wherein the third machine learning model is trained to predict the therapeutic reaction, by using a feature vector extracted from at least one layer included in the trained first machine learning model.

15 . The method of claim 13 , wherein the third machine learning model is trained to predict the therapeutic reaction, by using gene expression information included in the spatial transcriptomics information of the second object and position information corresponding to the gene expression information.

16 . A computer-readable recording medium having recorded thereon a program for causing a computer to execute the method of claim 11 .

Assignees

• Lunit Inc

Inventors

• Yoo Dong Geun
• Song Sang Hoon
• Ock Chan Young
• Jung Won Kyung
• Cho Soo Ick
• Paeng Kyung Hyun

Classifications

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T2207/30024

  Cell structures in vitro; Tissue sections in vitro · CPC title

• G06V2201/03

  Recognition of patterns in medical or anatomical images · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06V20/698

  Matching; Classification · CPC title