# US 10,373,314 B2 · Systems and methods for diagnosing tumors in a subject by performing a quantitative analysis of texture-based features of a tumor object in a radiological image

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

- Number: US 10,373,314 B2
- Document ID: US10373314B2
- Status: Granted
- Kind: B2
- Application number: 15/907,509
- Filed: 2018-02-28
- Published: 2018-09-06
- Granted: 2019-08-06
- Earliest claimed: 2013-10-12
- Assignees: H. Lee Moffitt Cancer Center and Research Institute, Inc.; University of South Florida
- Applicants: H. Lee Moffitt Cancer Center and Research Institute, Inc.; University of South Florida
- Inventors: Robert J. Gillies; Lawrence O. Hall; Dmitry B. Goldgof
- Web page: https://patentgrep.com/patent/US10373314B2

## CPC

- G06T 7/0012: {Biomedical image inspection}
- A61B 6/03: Computed tomography [CT]
- A61B 6/032: {Transmission computed tomography [CT]}
- A61B 6/037: {Emission tomography}
- A61B 6/50: specially adapted for specific body parts; specially adapted for specific clinical applications
- A61B 6/5217: {extracting a diagnostic or physiological parameter from medical diagnostic data}
- G06T 7/41: based on statistical description of texture
- G16H 50/30: for calculating health indices; for individual health risk assessment
- A61B 6/469: {for selecting a region of interest [ROI]}
- G06T 2207/10081: Computed x-ray tomography [CT]
- G06T 2207/30061: Lung
- G06T 2207/30096: Tumor; Lesion
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

An example method for diagnosing tumors in a subject by performing a quantitative analysis of a radiological image can include identifying a region of interest (ROI) in the radiological image, segmenting the ROI from the radiological image, identifying a tumor object in the segmented ROI and segmenting the tumor object from the segmented ROI. The method can also include extracting a plurality of quantitative features describing the segmented tumor object, and classifying the tumor object based on the extracted quantitative features. The quantitative features can include one or more texture-based features.

## Claims

### Claim 1

A method for diagnosing tumors in a subject by performing a quantitative analysis of a radiological image, comprising:
identifying, using a processor, a region of interest (ROI) in the radiological image;
segmenting the ROI from the radiological image using the processor;
identifying a tumor object in the segmented ROI using the processor;
segmenting the tumor object from the segmented ROI using the processor;
extracting, by the processor, a plurality of quantitative features describing the segmented tumor object; and
classifying the tumor object based on the extracted quantitative features using the processor, wherein the quantitative features include one or more shape-based features.

### Claim 2 (depends on 1)

The method of claim 1, wherein classifying the tumor object based on the extracted quantitative features further comprises predicting whether the tumor object is a malignant or benign tumor.

### Claim 3 (depends on 1, 2)

The method of any of claims 1-2, wherein classifying the tumor object based on the extracted quantitative features further comprises using a decision tree algorithm, a nearest neighbor algorithm or a support vector machine.

### Claim 4 (depends on 1)

The method of claim 1, wherein the quantitative features include one or more texture-based features.

### Claim 5 (depends on 4)

The method of claim 4, wherein each of the texture-based features describes a spatial arrangement of image intensities within the tumor object.

### Claim 6 (depends on 4)

The method of claim 4, wherein the texture-based features include at least one of a run-length texture feature, a co-occurrence texture feature, a Laws texture feature, a wavelet texture feature or a histogram texture feature.

### Claim 7 (depends on 4)

The method of claim 4, wherein a number of texture-based features is greater than approximately 150.

### Claim 8 (depends on 1)

The method of claim 1, wherein each of the shape-based features describes a location, a geometric shape, a volume, a surface area, a surface-area-to-volume ratio or a compactness of the tumor object.

### Claim 9 (depends on 1, 2, 3, 4, 5, 6, 7, 8)

The method of any of claims 1-8, wherein a total number of quantitative features is greater than approximately 200.

### Claim 10 (depends on 1, 2, 3, 4, 5, 6, 7, 8, 9)

The method of any of claims 1-9, wherein the ROI is a lung field.

### Claim 11 (depends on 1, 2, 3, 4, 5, 6, 7, 8, 9, 10)

The method of any of claims 1-10, wherein the radiological image is a low-dose computed tomography (CT) image.

### Claim 12 (depends on 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11)

The method of any of claims 1-11, further comprising:
reducing the extracted quantitative features to a subset of extracted quantitative features; and
classifying the tumor object based on the subset of extracted quantitative features.

### Claim 13 (depends on 12)

The method of claim 12, wherein the subset of extracted quantitative features includes one or more quantitative features that are predictive of the classification of the tumor object.

### Claim 14 (depends on 13)

The method of claim 13, further comprising determining the one or more quantitative features that are predictive of the classification of the tumor object using at least one of a Recursive Elimination of Features (Relief-F) algorithm, a Correlation-Based Feature Subset Selection for Machine Learning (CFS) algorithm or a Relief-F with Correlation Detection algorithm.

### Claim 15 (depends on 12)

The method of claim 12, wherein the subset of extracted quantitative features includes one or more non-redundant quantitative features having adequate reproducibility and dynamic range.

### Claim 16 (depends on 15)

The method of claim 15, wherein reducing the extracted quantitative features to a subset of extracted quantitative features further comprises eliminating one or more of the extracted quantitative features having a reproducibility metric between a baseline radiological image and a subsequent radiological image less than a predetermined reproducibility value, the subsequent radiological image being captured a fixed period of time after the baseline radiological image was captured.

### Claim 17 (depends on 16)

The method of claim 16, wherein the reproducibility metric is a concordance correlation coefficient.

### Claim 18 (depends on 16, 17)

The method of any of claims 16-17, wherein the predetermined reproducibility value is less than 0.90.

### Claim 19 (depends on 16, 17, 18)

The method of any of claims 16-18, wherein reducing the extracted quantitative features to a subset of extracted quantitative features further comprises eliminating one or more of the extracted quantitative features having a dynamic range less than a predetermined dynamic range value.

### Claim 20 (depends on 16, 17, 18, 19)

The method of any of claims 16-19, wherein reducing the extracted quantitative features to a subset of extracted quantitative features further comprises eliminating one or more of the extracted quantitative features that are redundant quantitative features.

### Claim 21 (depends on 20)

The method of claim 20, further comprising calculating a coefficient of determination (R2Bet) between at least two quantitative features, wherein the R2Bet of each of the redundant quantitative features is less than a predetermined redundancy value.

### Claim 22 (depends on 21)

The method of claim 21, wherein predetermined redundancy value is 0.95.

### Claim 23 (depends on 1)

The method of claim 1, wherein a total number of quantitative features is greater than approximately 10.