# US 11,443,433 B2 · Quantification and staging of body-wide tissue composition and of abnormal states on medical images via automatic anatomy recognition

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

- Number: US 11,443,433 B2
- Document ID: US11443433B2
- Status: Granted
- Kind: B2
- Application number: 16/271,949
- Filed: 2019-02-11
- Published: 2019-08-22
- Granted: 2022-09-13
- Earliest claimed: 2018-02-10
- Assignees: The Trustees of the University of Pennsylvania
- Applicants: The Trustees of the University of Pennsylvania
- Inventors: Jayaram K. Udupa; Tiange Liu; Drew A. Torigian; Dewey Odhner; Yubing Tong
- Web page: https://patentgrep.com/patent/US11443433B2

## CPC

- G06T 7/0012: {Biomedical image inspection}
- A61B 5/4519: {Muscles (measuring muscular strength A61B5/224; using electromyography [EMG] A61B5/389)}
- A61B 5/4842: {Monitoring progression or stage of a disease}
- A61B 5/4872: {Body fat}
- A61B 5/7267: {involving training the classification device}
- A61B 5/7275: {Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor}
- A61B 6/032: {Transmission computed tomography [CT]}
- A61B 6/037: {Emission tomography}
- A61B 6/505: {for diagnosis of bone}
- A61B 6/5217: {extracting a diagnostic or physiological parameter from medical diagnostic data}
- A61B 34/10: Computer-aided planning, simulation or modelling of surgical operations
- G06F 16/51: Indexing; Data structures therefor; Storage structures
- G06F 18/24323: {Tree-organised classifiers}
- G06T 7/11: Region-based segmentation
- G06T 7/187: involving region growing; involving region merging; involving connected component labelling
- G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI]
- G06V 10/764: using classification, e.g. of video objects
- G16H 30/40: for processing medical images, e.g. editing
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- A61B 5/4869: {Determining body composition}
- A61B 6/5235: {combining images from the same or different ionising radiation imaging techniques, e.g. PET and CT}
- G06T 2207/30004: Biomedical image processing
- G06V 2201/031: of internal organs
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- 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

## Abstract

Quantification of body composition plays an important role in many clinical and research applications. Radiologic imaging techniques such as Dual-energy X-ray absorptiometry, magnetic resonance imaging (MRI), and computed tomography (CT) imaging make accurate quantification of the body composition possible. This disclosure presents an automated, efficient, accurate, and practical body composition quantification method for low dose CT images; method for quantification of disease from images; and methods for implementing virtual landmarks.

## Claims

### Claim 1

A computerized method of providing, based on medical images, quantification of body-wide tissue composition including one or more of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), muscle tissue, and bone tissue via automatic anatomy recognition, the method comprising the steps of:
determining, based on automatically identifying a plurality of virtual landmarks in a set of medical images, a body region of interest, wherein determining the body region of interest comprises determining, based on one or more of the plurality of virtual landmarks, a first image slice indicative of a first boundary of the body region of interest and determining, based on one or more of the plurality of virtual landmarks, a second image slice indicative of a second boundary of the body region of interest;
generating, based on at least a portion of the set of medical images associated with the body region of interest, a fuzzy anatomy model of the body region of interest; and
using, based on one or more patient images and the fuzzy anatomy model, automatic anatomy recognition (AAR) to recognize and delineate one or more tissue objects, in the body region of interest of a patient, wherein the tissue objects comprise one or more of a subcutaneous adipose tissue (SAT) object, a visceral adipose tissue (VAT) object, a muscle tissue object, or a bone tissue object.

### Claim 2 (depends on 1)

The computerized method of claim 1, further comprising:
determining, using at least positron emission tomography imaging data from the one or more patient images, a disease burden in the body region of interest of the patient.

### Claim 3 (depends on 1)

The computerized method of claim 1, wherein generating a fuzzy anatomy model of the body region of interest comprises gathering image data from the set of medical images determining definitions of each body region and organ and delineating them following said definitions, building hierarchical fuzzy anatomy models of organs and/or tissues for each body region, and selecting from the hierarchical fuzzy anatomy models the fuzzy anatomy model of the body region of interest.

### Claim 4 (depends on 3)

The computerized method of claim 3, wherein using AAR to recognize and delineate one or more tissue objects in the body region of interest comprises recognizing and locating organs and/or tissues in the one or more patient images by employing the hierarchical fuzzy anatomy models, and delineating tissue objects indicating specific tissue types in the body region of interest following the hierarchy.

### Claim 5 (depends on 1)

The computerized method of claim 1, wherein the body region of interest comprises a region from a body torso.

### Claim 6

A system that provides, based on medical images, quantification of body-wide tissue composition including one or more of subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), muscle tissue, and bone tissue via automatic anatomy recognition, the system comprising:
a database that stores image data from medical image sets;
a memory storing computer instructions; and
a processor that processes the computer instructions to:
determine, based on automatically identifying a plurality of virtual landmarks in a medical image set of the image data, a body region of interest, wherein determining the body region of interest comprises determining, based on one or more of the plurality of virtual landmarks, a first image slice indicative of a first boundary of the body region of interest and determining, based on one or more of the plurality of virtual landmarks, a second image slice indicative of a second boundary of the body region of interest;
generate, based on at least a portion of the image data associated with the body region of interest, a fuzzy anatomy model of the body region of interest; and
use automatic anatomy recognition (AAR) to recognize and delineate, based on one or more patient images and the fuzzy anatomy model, one or more tissue objects in the body region of interest for a particular patient, wherein the one or more tissue objects comprise one or more of a subcutaneous adipose tissue (SAT) object, a visceral adipose tissue (VAT) object, a muscle tissue object, and a bone tissue object.

### Claim 7 (depends on 6)

The system of claim 6, wherein the processor is further configured to determine, using at least image data from the one or more patient images, a disease burden in the body region of interest of the patient.

### Claim 8

A computerized method of recognizing, staging, and quantifying, via automatic anatomy recognition and based on patient images, abnormal states within objects body-wide, the method comprising the steps of:
determining, based on automatically identifying using an iterative principal component analysis process a plurality of virtual landmarks in a set of medical images, a body region of interest;
locating, based on determining the body region of interest, an object in the body region of interest of a particular patient; and
determining, using at least image data from the set of medical images, an indication of a disease in the located object.

### Claim 9 (depends on 8)

The computerized method of claim 8, wherein the indication of disease comprises an indication of a disease burden, wherein the indication of the disease burden is calculated based on determining, for each voxel associated with the object, a disease severity value indicating presence of a disease in the voxel.

### Claim 10 (depends on 8)

The computerized method of claim 8, wherein the indication of the disease comprises one or more of a metric quantifying the disease, a disease stage, a re-stage state, or a predictive state analysis.

### Claim 11 (depends on 8)

The computerized method of claim 8, further comprising:
determining, using at least image data from the set of medical images, a treatment plan for the particular patient.

### Claim 12 (depends on 8)

The computerized method of claim 8, wherein the virtual landmarks are not limited to a surface of the object.

### Claim 13 (depends on 8)

The computerized method of claim 8, wherein determining the body region of interest is based on using a machine learning model trained to correlate virtual landmarks with corresponding boundaries of body regions of interest.

### Claim 14 (depends on 8)

The computerized method of claim 8, wherein the iterative principal component analysis process comprises performing a first level principal component analysis defining one or more octants and performing a first level principal component analysis within at least one of the one or more octants.

### Claim 15 (depends on 1)

The computerized method of claim 1, wherein each of the one or more tissue objects comprises a set of voxels in the body region of interest that are identified as comprising a tissue type associated with the tissue object.

### Claim 16 (depends on 1)

The computerized method of claim 1, wherein using AAR to recognize and delineate one or more tissue objects comprises determining on a voxel-by-voxel basis an indication of whether a tissue type of the tissue object is located in the corresponding voxel.

### Claim 17 (depends on 1)

The computerized method of claim 1, wherein the set of medical images comprises one or more of positron emission tomography, computed tomography images, or low-dose computed tomography images.

### Claim 18 (depends on 1)

The computerized method of claim 1, wherein automatically identifying the plurality of virtual landmarks in the set of medical images comprises:
performing an iterative process for determining the plurality of virtual landmarks, wherein at least one iteration of the iterative process comprises a predetermined process that automatically determines virtual landmark locations based on a prior virtual landmark without further analysis of any objects in the set of medical images.

### Claim 19 (depends on 1)

The computerized method of claim 1, wherein the first boundary comprises a superior boundary and the second boundary comprises an inferior boundary.

### Claim 20 (depends on 8)

The computerized method of claim 8, wherein the virtual landmarks are able to be located in an image at any of the following locations: on a surface of the object, outside the object, and inside the object.