# US 12,008,747 B2 · Population-specific prediction of prostate cancer recurrence based on stromal morphology features

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,008,747 B2
- Document ID: US12008747B2
- Status: Granted
- Kind: B2
- Application number: 16/886,966
- Filed: 2020-05-29
- Published: 2021-02-04
- Granted: 2024-06-11
- Earliest claimed: 2019-07-29
- Assignees: Case Western Reserve University
- Applicants: Case Western Reserve University
- Inventors: Anant Madabhushi; Hersh Bhargava; Patrick Leo; Priti Lal
- Web page: https://patentgrep.com/patent/US12008747B2

## CPC

- G06T 7/0012: {Biomedical image inspection}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- G06N 20/20: Ensemble learning
- G06T 7/11: Region-based segmentation
- G06V 10/7796: {based on specific statistical tests}
- G16H 30/40: for processing medical images, e.g. editing
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/30: for calculating health indices; for individual health risk assessment
- G06F 18/2113: by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
- G06N 3/045: Combinations of networks
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06T 2207/20076: Probabilistic image processing
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30024: Cell structures in vitro; Tissue sections in vitro
- G06T 2207/30081: Prostate
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Embodiments discussed herein facilitate determination of one of a probability of prostate cancer recurrence-free survival or a risk factor associated with prostate cancer based on intra-tumor stromal morphology. Example embodiments can perform operations comprising: accessing a digitized histological image of a prostate of a patient, wherein the histological image comprises a region of interest associated with prostate cancer; identifying nuclei of intra-tumoral stromal cells within the region of interest; extracting, for the region of interest of the digitized histological image, one or more features describing the structure of the intra-tumoral stromal cells; and generating, via a model based at least on the one or more features, one of a probability of prostate cancer recurrence-free survival or a risk score associated with prostate cancer for the patient based at least on the extracted one or more features.

## Claims

### Claim 1

A non-transitory computer-readable medium storing computer-executable instructions that, when executed, cause a processor to perform operations, comprising:
accessing a digitized histological image of a prostate of a patient, wherein the digitized histological image comprises a region of interest associated with prostate cancer;
generating a first probability map that indicates a likelihood that each pixel in the digitized histological image is part of a nucleus;
determining nuclear coordinates using the first probability map;
generating a second probability map that indicates a likelihood that each pixel in the digitized histological image is part of a stromal compartment;
combining information from the second probability map with the nuclear coordinates to identifying nuclei of intra-tumoral stromal cells within the region of interest;
extracting, for the region of interest of the digitized histological image, one or more features describing a structure of the intra-tumoral stromal cells; and
generating, via a model based at least on the one or more features, a risk score associated with prostate cancer for the patient based at least on the one or more features.

### Claim 2 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the operations further comprise:
segmenting nuclei within the region of interest; and
segmenting intra-tumoral stroma within the region of interest,
wherein identifying the nuclei of the intra-tumoral stromal cells is based at least in part on the segmented nuclei and the segmented intra-tumoral stroma.

### Claim 3 (depends on 2)

The non-transitory computer-readable medium of claim 2, wherein the nuclei and the intra-tumoral stroma are segmented via one or more deep learning models.

### Claim 4 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the model is based at least on the one or more features for a training set of histological images associated with a population sub-group, wherein the patient is a member of the population sub-group.

### Claim 5 (depends on 4)

The non-transitory computer-readable medium of claim 4, wherein the population sub-group is African American.

### Claim 6 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the one or more features comprise at least one of: one or more Voronoi tessellation features, one or more Delauney triangulation features, one or more minimum spanning tree features, one or more local nuclear cluster graph features, one or more nuclear shape features, one or more cell orientation entropy features, one or more sub-graph features, or one or more texture features.

### Claim 7 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the model is a machine learning model.

### Claim 8 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the model is an elastic-net penalized Cox regression model.

### Claim 9 (depends on 1)

The non-transitory computer-readable medium of claim 1, further comprising:
determining a likelihood of prostate cancer recurrence based upon the risk score.

### Claim 10

A method of assessing prostate cancer for a patient, comprising:
accessing an image of a prostate of a patient, wherein the image comprises a region of interest associated with prostate cancer;
generating a first probability map that indicates a likelihood that each pixel in the image is part of a nucleus;
determining nuclear coordinates using the first probably map;
generating a second probability map representing a likelihood that each pixel in the image is part of a stromal compartment;
combining information from the second probability map with the nuclear coordinates to determine coordinates of nuclei of intra-tumoral stromal cells;
identifying boundaries of the nuclei of the intra-tumoral stromal cells and the stromal compartment within the region of interest;
extracting, from the region of interest of the image, one or more features describing a shape, an orientation, and an arrangement of nuclei within the intra-tumoral stromal cells, wherein the boundaries of the nuclei of the intra-tumoral stromal cells and the stromal compartment are used in extracting the one or more features; and
generating, via a model based at least on the one or more features, a risk score associated with prostate cancer for the patient.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
determining a likelihood of recurrence of the prostate cancer based upon the risk score.

### Claim 12 (depends on 10)

The method of claim 10, further comprising:
comparing the risk score to a risk score threshold to determine a likelihood of recurrence of the prostate cancer.

### Claim 13 (depends on 10)

The method of claim 10, wherein the one or more features comprise metrics derived from a stromal texture, a connectivity graph of the nuclei of the intra-tumoral stromal cells, nuclear centroids, descriptors of nuclear shape and orientation, and sub-graph features.

### Claim 14 (depends on 10)

The method of claim 10, further comprising:
assessing a prognostic capability of the one or more features with respect to Biochemical Recurrence-Free Survival (BRFS) using a univariate Cox proportional hazards regression.

### Claim 15 (depends on 10)

The method of claim 10, wherein the one or more features further describe a shape of the intra-tumoral stromal cells, an orientation of the intra-tumoral stromal cells, and a texture of the intra-tumoral stromal cells.

### Claim 16 (depends on 10)

The method of claim 10, further comprising:
segmenting nuclei of the image within the region of interest; and
segmenting the intra-tumoral stromal cells of the image within the region of interest, wherein identifying the boundaries of the nuclei of the intra-tumoral stromal cells is based at least in part on the segmented nuclei of the image and the segmented intra-tumoral stromal cells of the image.

### Claim 17 (depends on 10)

The method of claim 10, wherein the one or more features are extracted after identifying the boundaries of nuclei of the intra-tumoral stromal cells and the stromal compartment within the region of interest.

### Claim 18 (depends on 10)

The method of claim 10, further comprising:
segmenting nuclei using a convolutional neural network to generate the first probability map;
and
segmenting the intra-tumoral stromal cells using a U-Net convolution neural network to generate the second probability map.

### Claim 19 (depends on 10)

The method of claim 10, wherein the model is based at least on the one or more features associated with a population sub-group, wherein the patient is a member of the population sub-group.

### Claim 20 (depends on 19)

The method of claim 19, wherein the population sub-group is African American.

### Claim 21 (depends on 10)

The method of claim 10, wherein the one or more features comprise at least one of: one or more Voronoi tessellation features, one or more Delauney triangulation features, one or more minimum spanning tree features, one or more local nuclear cluster graph features, one or more nuclear shape features, one or more cell orientation entropy features, one or more sub-graph features, or one or more texture features.

### Claim 22

An apparatus for assessing prostate cancer within a patient, comprising:
a memory configured to store an image of a prostate of a patient, wherein the image comprises a region of interest associated with prostate cancer;
a one or more circuits configured to:
generate a first probability map that indicates a likelihood that each pixel in the image is part of a nucleus;
determine nuclear coordinates using the first probably map;
generate a second probability map representing a likelihood that each pixel in the image is part of a stromal compartment;
combine information from the second probability map with the nuclear coordinates to determine coordinates of a nuclei of intra-tumoral stromal cells;
identify boundaries of the nuclei of the intra-tumoral stromal cells and the stromal compartment within the region of interest using convolutional neural network to represent a likelihood each pixel in the image is part of the stromal compartment;
extract, from the region of interest of the image, one or more features describing a structure of the intra-tumoral stromal cells, wherein the boundaries of the nuclei of the intra-tumoral stromal cells and the stromal compartment are used in extracting the one or more features; and
generate, via a model based at least on the one or more features, a risk score associated with prostate cancer within the patient.

### Claim 23 (depends on 22)

The apparatus of claim 22, wherein the one or more features comprise metrics derived from a stromal texture, a connectivity graph of the nuclei of the intra-tumoral stromal cells, nuclear centroids, descriptors of nuclear shape and orientation, and sub-graph features.

### Claim 24 (depends on 22)

The apparatus of claim 22, wherein the one or more features describe a shape of the intra-tumoral stromal cells, an orientation of the intra-tumoral stromal cells, an arrangement of the nuclei within the intra-tumoral stromal cells, and a texture of the intra-tumoral stromal cells.

### Claim 25 (depends on 22)

The apparatus of claim 22, wherein the one or more circuits are further configured to extract the one or more features after identifying the boundaries of nuclei of the intra-tumoral stromal cells and the stromal compartment within the region of interest.

### Claim 26 (depends on 22)

The apparatus of claim 22, wherein the one or more circuits are further configured to
determine a likelihood of recurrence of the prostate cancer based upon the risk score.