System and method for tumor characterization

Patent metadata

Field	Value
Publication number	US-12536223-B2
Application number	US-202418750510-A
Country	US
Kind code	B2
Filing date	Jun 21, 2024
Priority date	May 29, 2019
Publication date	Jan 27, 2026
Grant date	Jan 27, 2026

Abstract

A method of treating a subject comprises administering a treatment to a subject identified as having a high probability of distant metastatic recurrence, wherein the probability of distant metastatic recurrence was determined by a process, comprising acquiring at least one image of a tissue sample comprising a plurality of cells, taken from a subject, classifying each of the plurality of cells into categories, dividing the at least one image into a plurality of patches, calculating values for a plurality of morphological features based on the patches, and calculating a distant metastatic recurrence probability based on the values. A computer-implemented method of training a neural network and a system for characterizing a cancer in a subject are also described.

First claim

What is claimed is:

1 . A computer-implemented method of training a neural network for characterizing a cancer in a body of a subject, comprising:
acquiring at least one first image of a first tissue sample comprising a plurality of cells, taken from a first region of the body of the subject; classifying each of the plurality of cells into categories; dividing the at least one image into a plurality of patches, each patch comprising multiple cells of the plurality of cells; calculating a cell density of each of the plurality of patches; calculating values for a plurality of morphological features based on the patches, wherein the morphological features comprise at least one of (Count of immune cells in “large” cluster)/(Total count of immune cells); (Count of immune cells in “large” cluster)/(Total count of tumor+immune); (Count of immune cells in “large” cluster)/(Count of tumor cells in “large” cluster); (Count of immune cells in “large” cluster)/(Count of immune cells in “small” cluster); (Total count of immune cells)/(Total count of tumor+immune); (Immune cell total area)/(Tumor cell total area); and (Immune cell total area)/(Total Immune area+Tumor area); calculating a distant metastatic recurrence probability based on the values and the cell density; and training at least one neural network using a combination of the values and a low-dimensional representation of a sequence generated by a DNN.

2 . The method of claim 1 , wherein the patches have a size between 100×100 and 1000×1000.

3 . The method of claim 1 , further comprising the step of down sampling at least one of the plurality of patches.

4 . The method of claim 1 , further comprising discarding patches having no tumor cell information from the plurality of patches.

5 . The method of claim 1 , further comprising assigning cells in each of the plurality of patches to one or more clusters using a clustering algorithm to produce cluster data for each patch, wherein the cluster data comprises a cluster size.

6 . The method of claim 5 , wherein the calculating the distant metastatic recurrence is further based on the clustering data.

7 . The method of claim 1 , further comprising calculating an estimated characteristic of the cancer based on the plurality of morphological features.

8 . The method of claim 1 , wherein the image is divided into patches at least in part by random selection.

9 . The method of claim 1 , further comprising at least one second image of a second tissue sample taken from a second region of the body of the subject.

10 . The method of claim 1 , wherein the distant metastatic recurrence probability is calculated by aggregating a set of votes for each patch in the plurality of patches based on the values and the cell density.

11 . A computer-implemented method of training a neural network for characterizing a cancer in a subject, comprising:
acquiring at least one image of a tissue sample comprising a plurality of cells, taken from a subject; classifying each of the plurality of cells into categories; dividing the at least one image into a plurality of patches at least partially by random selection; discarding patches having no tumor cell information from the plurality of patches; calculating values for a plurality of morphological features based on the patches; and training at least one neural network using a combination of the values and a low-dimensional representation of a sequence generated by a DNN; wherein the morphological features comprise at least one of (Count of immune cells in “large” cluster)/(Total count of immune cells); (Count of immune cells in “large” cluster)/(Total count of tumor+immune); (Count of immune cells in “large” cluster)/(Count of tumor cells in “large” cluster); (Count of immune cells in “large” cluster)/(Count of immune cells in “small” cluster); (Total count of immune cells)/(Total count of tumor+immune); (Immune cell total area)/(Tumor cell total area); and (Immune cell total area)/(Total Immune area+Tumor area).

12 . The method of claim 11 , further comprising calculating a cell density of each of the plurality of patches and discarding patches having a cell density below a threshold.

13 . The method of claim 11 , wherein each patch comprises a plurality of cells.

14 . The method of claim 11 , further comprising calculating an estimated characteristic of the cancer based on the plurality of morphological features.

15 . The method of claim 14 , wherein the estimated characteristic is a probability of distant metastatic recurrence.

16 . The method of claim 11 , wherein the categories comprise tumor cells, non-tumor cells, and immune cells.

17 . The method of claim 11 , further comprising at least one second image of a second tissue sample taken from a second region of the body of the subject.

18 . The method of claim 11 , further comprising calculating an estimated characteristic of the cancer based on a combination of the values and the low-dimensional representation.

Assignees

• Univ New York

Inventors

• Wang Jing
• Kulkarni Prathamesh
• Robinson Eric

Classifications

• G06V2201/03

  Recognition of patterns in medical or anatomical images · CPC title

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H30/40

  for processing medical images, e.g. editing · CPC title

• G16H30/20

  for handling medical images, e.g. DICOM, HL7 or PACS · CPC title

• G06V20/698

  Matching; Classification · CPC title