Pathology case review, analysis and prediction

Patent metadata

Field	Value
Publication number	US-10733726-B2
Application number	US-201816029761-A
Country	US
Kind code	B2
Filing date	Jul 9, 2018
Priority date	Nov 6, 2013
Publication date	Aug 4, 2020
Grant date	Aug 4, 2020

Abstract

Systems and methods for personalized cancer therapy using analysis of pathology slides to target regions in a single sample that interrogates the feature data of a relatively large number of cells. The disclosure describes pathology case review tools of the future which include analysis, visualization and prediction modeling to provide novel information to the pathologist for the diagnosis of disease. This disclosure further describes a user interface to assist the physicians that make that diagnosis, pathologists. Complex computer learning algorithms will combine and mine these data sets to recommend optimal treatment strategies. A computer interface is provided which allows a pathologist to access those data instantly to make a more informed and accurate diagnosis.

First claim

What is claimed:

1. A method of simulating tumor characteristics over time, comprising:
obtaining digital slide image data of a tissue;
identifying regions of interest in the digital slide image data by performing a multidimensional statistical and morphometrical analysis to the digital slide image data to identify quantitative features of individual cells that are characteristic of specific tumors;
extracting object feature data from objects of interest, wherein the feature data represents parameters associated with the objects;
reconstructing 3D tissue architecture and metabolic landscape from object of interest identified in separate but consecutive tissue slices;
building a model space and parameterizing each object of interest; and
simulating the tissue over time in accordance with the object feature data.

2. The method of claim 1 , wherein identifying the regions of interest includes segmenting non-tumor regions from tumor regions.

3. The method of claim 1 , further comprising applying a spatial prediction modeling that is parameterized from histological feature data obtained from the slide image data.

4. The method of claim 1 , further comprising:
quantitatively classifying multidimensionality and heterogeneity of image-based data;
correlating results of the classifying with patient survival outcome data; and
determining a set of cellular features that distinguish between tumors that led to recurrence or metastasis from tumors that did not progress.

5. The method of claim 1 , further comprising reconstructing a tissue metabolic landscape by utilizing information from consecutive tissue slices to determine levels of different metabolites simultaneously in each region of the tissue.

6. The method of claim 1 , further comprising testing dynamical interactions between tumor cells, a microenvironment of the tumor cells and treatment of the tumor cells.

7. The method of claim 6 , further comprising generating dynamic simulations of at least one of a history a particular tumor prior to resection, a future progression of the particular tumor, and response of the particular tumor to therapeutic treatments.

8. The method of claim 1 , further comprising:
using a whole tumor tissue and a surrounding stroma after tumor resection together with a panel of molecular biomarkers; and
providing an analysis of a morphological and metabolic state of the whole tumor.

9. The method of claim 1 , further comprising combining static information from pathology samples with dynamical 2D/3D computer simulations using a computer model parameterized with individual patient data.

10. The method of claim 1 , further comprising applying a multidimensional statistical and morphometrical analysis to identify quantitative features of individual cells that are characteristic for a given tumor type, grade, stage or hormonal status.

11. A method of virtual pathology simulation, comprising:
using a model parameterization with patient-specific tumor morphology;
characterizing individual cellular features of each type of tumor;
simulating each tissue by systematically varying parameters defining cell doubling time, metabolic state and sensitivity to therapeutic compounds, wherein the model produces multi-parameter report charts of tumor response to a typical treatment schedule; and
analyzing outcomes produced by the model,
wherein the patient-specific tumor morphology includes at least one of a localization of tumor vasculature, localization of proliferating cells and regions of tumor hypoxia and tumor cells glycolytic state.

12. The method of claim 11 , further comprising accounting for tumor and stromal cell's morphometric parameters and expression levels of molecular biomarkers to provide a prediction of optimal tumor response.

13. A method of landscape pathology, comprising:
obtaining a slide image data of a tissue;
identifying regions of interest in the slide image data;
segmenting objects of interest in the slide image data;
extracting object feature data that represents parameters associated with each object of interest;
performing analytics on multiparametric tissue properties to determine environmental factors leading to tumor growth;
creating a multidimensional visualization of cell distributions in accordance with the object feature data;
quantifying heterogeneities within the cell distributions;
applying a topographical analysis of spatial heterogeneity; and
determining at least one of spatial pattern including point patterns, spatial patterns, network patterns and outlier patterns.

14. The method of claim 13 , wherein the regions of interest are non-tumor regions that are segmented from tumor regions.

15. The method of claim 13 , further comprising:
identifying patterns in the objects of interest using landscape ecology and connectivity; and
applying an outlier analysis to determine biological relationships in cellular distributions within the objects of interest.

16. The method of claim 13 , further comprising:
using one of a number of interactions with neighbors in a nearest neighbor analysis, a buffered analysis and a spatial autocorrelation to discern morphological or other similarities between nearby individuals; and
determining a size and connectivity of pockets of cells with distinct morphologies.

17. The method of claim 13 , further comprising measuring environmental factors that lead to tumor growth to quantify a microenvironment.

18. The method of claim 13 , further comprising extracting Darwinian interactions of environmental and cellular adaptations from the slide image data.

19. The method of claim 13 , wherein the spatial patterns are determined in accordance with clustered patterns.

20. The method of claim 13 , wherein a hill-and-valley map is used to show spatial distribution.

Assignees

• H Lee Moffitt Cancer Ct & Res
• H Lee Moffitt Cancer Center And Res Institute

Inventors

• Lloyd Mark Cassidy Cridlin
• Bui Marilyn Yuanxin Ma
• Rejniak Katarzyna Ann
• Puri Munish

Classifications

• G06F18/24

  Classification techniques · CPC title

• G06V2201/031

  of internal organs · CPC title

• G06V20/695

  Preprocessing, e.g. image segmentation · CPC title

• G06V20/698

  Matching; Classification · CPC title

• G06T7/11

  Region-based segmentation · CPC title