# US 10,499,866 B2 · Mammography apparatus

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

- Number: US 10,499,866 B2
- Document ID: US10499866B2
- Status: Granted
- Kind: B2
- Application number: 15/750,514
- Filed: 2016-08-04
- Published: 2018-08-16
- Granted: 2019-12-10
- Earliest claimed: 2015-08-06
- Assignees: Tel HaShomer Medical Research Infrastructure and Services Ltd.
- Applicants: Tel HaShomer Medical Research, Infrastructure and Services Ltd.
- Inventors: Miriam Sklair-Levy; Arnaldo Mayer; Shmuel Yitzhak Pfeffer
- Web page: https://patentgrep.com/patent/US10499866B2

## CPC

- A61B 6/502: {for diagnosis of breast, i.e. mammography}
- A61B 6/0414: {with compression means}
- A61B 6/0435: {with means for imaging suspended breasts}
- A61B 6/481: {involving the use of contrast agents}
- A61B 6/5205: {involving processing of raw data to produce diagnostic data}
- G06T 7/0012: {Biomedical image inspection}
- G06T 7/11: Region-based segmentation
- G06T 7/13: Edge detection
- G06T 7/44: using image operators, e.g. filters, edge density metrics or local histograms
- G06V 10/421: {by analysing segments intersecting the pattern}
- G06V 10/46: Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features (colour feature extraction G06V10/56)
- G06V 10/50: by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
- A61B 6/12: Arrangements for detecting or locating foreign bodies
- G06T 2207/10116: X-ray image
- G06T 2207/20081: Training; Learning
- G06T 2207/30068: Mammography; Breast
- G06V 2201/03: Recognition of patterns in medical or anatomical images
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method of diagnosing breast cancer, the method comprising: acquiring a contrast enhanced region of interest (CE-ROI) comprised in an X-ray image of a patient's breast, the X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image; generating a texture feature vector (TF) having components based on the indications of intensity provided by a plurality of X-ray pixels in the CE-ROI; and using a classifier to classify the texture feature vector TF to determine whether the CE-ROI is malignant.

## Claims

### Claim 1

A method of diagnosing breast cancer, the method comprising:
acquiring a contrast enhanced region of interest (CE-ROI) comprised in an X-ray image of a patient's breast, the X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image;
determining a texture neighborhood for each of the plurality of X-ray pixels, the texture neighborhood for a given X-ray pixel of the plurality of X-ray pixels extending to a bounding pixel radius of BPR pixels from the given pixel;
selecting a plurality of R different pixel radii PR(r), 1≤r≤R, in the texture neighborhood, for which PR(r)<PR(r+1) and PR(R)=BPR;
determining values for a number of M representative bins of a texture vector histogram HTV(T) having a bin length in number of bins T equal to a sum over r from 1 to R of 2{circumflex over ( )}(r), where J(r) is a number of X-ray pixels at radius r selected for processing, and values HTV(t) 1≤t≤T that are a function of texture vectors TV determined for the texture neighborhoods of a plurality of pixels in the CE-ROI
generating a texture feature vector (TF) having components based on at least a portion of the values determined for the M representative bins; and
using a classifier to classify the texture feature vector TF to determine whether the CE-ROI is malignant.

### Claim 2 (depends on 1)

The method according to claim 1 wherein BPR is greater than or equal to about 25, 50, or 75.

### Claim 3 (depends on 1)

The method according to claim 1 wherein R is greater than or equal to 2, 10, or 20.

### Claim 4 (depends on 1)

The method according to claim 1 and comprising for each pixel radius assigning a binary number to each of the J(r) X-ray pixels at the pixel radius, wherein the assigned binary number is 1 if the indication of intensity of the X-ray pixel is greater than the indication of intensity of the given X-ray pixel and 0 otherwise and comprising for each pixel radius PR(r) generating a binary string based on the assigned binary numbers.

### Claim 5 (depends on 4)

The method according to claim 4 and comprising converting the binary strings for the pixel radii PR(r), 1≤r≤R to decimal numbers N(r), 1≤r≤R respectively and determining a texture vector (TV) for the texture neighborhood of the given X-ray pixel comprising a set of the decimal numbers {N(1), N(2), . . . N(R)}.

### Claim 6 (depends on 1)

The method according to claim 1 wherein if P(0)=0, and P(r)=2{circumflex over ( )}(1)+2{circumflex over ( )}J(2), . . . , +2{circumflex over ( )}J(r), for 1≤r≤R, a value HTV(t) of a t-th bin in the histogram for which P(r)<t≤P(r+1) is equal to a number of the texture neighborhoods of the given X-ray pixels of the plurality of X-ray pixels for which N(r)=(t−(P(r)).

### Claim 7 (depends on 1)

The method according to claim 1 and comprising determining components of the texture feature vector (TF) to comprise a set of at least a portion of the values determined for the M representative bins.

### Claim 8 (depends on 1)

The method according to claim 1 wherein M is less than or equal to about 50, 20, or 5.

### Claim 9 (depends on 1)

The method according to claim 1 wherein the X-ray image is a spectral contrast enhanced digital mammography (SCEDM) image.

### Claim 10 (depends on 1)

The method according to claim 1 wherein the classifier comprises a support vector machine and/or a neural network and/or a convolutional neural network.

### Claim 11 (depends on 1)

The method according to claim 1 and comprising processing the CE-ROI to determine a time dependent component of a feature vector and using the time dependent component to determine whether the CE-ROI indicates a malignancy.

### Claim 12 (depends on 1)

The method according to claim 1 and comprising using a convolutional neural network (CNN) to automatically identify and determine which features of the CE-ROI to use to generate the feature vector.

### Claim 13

A method of diagnosing breast cancer, the method comprising:
acquiring a contrast enhanced region of interest (CE-ROI) comprised in an X-ray image of a patient's breast, the X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image;
determining a texture neighborhood for each of the plurality of X-ray pixels, the texture neighborhood for a given X-ray pixel of the plurality of X-ray pixels extending to a bounding pixel radius of BPR pixels from the given pixel;
generating a texture feature vector (TF) having components based on the indications of intensity provided by a plurality of X-ray pixels in the CE-ROI that are located within the texture neighborhood;
determining a contour for the CE-ROI and a contour feature vector CF based on the contour and concatenating the texture feature vector TF and the contour feature vector CF to provide a feature vector CF-TF; and
using a classifier to classify CF-TF to determine whether the CE-ROI is malignant.

### Claim 14

A method of diagnosing breast cancer, the method comprising:
acquiring a contrast enhanced region of interest (CE-ROI) comprised in an X-ray image of a patient's breast, the X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image;
determining a texture neighborhood for each of the plurality of X-ray pixels, the texture neighborhood for a given X-ray pixel of the plurality of X-ray pixels extending to a bounding pixel radius of BPR pixels from the given pixel;
generating a texture feature vector (TF) having components based on the indications of intensity provided by a plurality of X-ray pixels in the CE-ROI that are located within the texture neighborhood;
determining a profile feature vector (PF) based on a personal profile of a patient for which the X-ray image was acquired;
concatenating the profile feature vector PF with the texture feature vector TF to form a feature vector PF-TF; and
using a classifier to classify PF-TF to determine whether the CE-ROI indicates a malignancy.

### Claim 15 (depends on 14)

The method according to claim 14 and comprising determining a contour for the CE-ROI and a contour feature vector CF based on the contour, determining a profile feature vector PF based on a personal profile of a patient for which the X-ray image was acquired, and concatenating the profile feature vector PF with the contour feature vector CF and the texture feature vector TF to form a feature vector PF-CF-TF, and using a classifier to classify PF-CF-TF to determine whether the CE-ROI indicates a malignancy.

### Claim 16

Apparatus for diagnosing breast cancer, the apparatus comprising:
a source of X-rays;
a controller configured to control the X-ray source to acquire an SCEDM X-ray image of a patient's breast the SCEDM X-ray image comprising X-ray pixels that indicate intensity of X-rays that passed through the breast to generate the image; and
a controller comprising a set of instructions executable to process the SCEDM X-ray image to:
acquire a contrast enhanced region of interest (CE-ROI) in the SCEDM X-ray image;
determine a texture neighborhood for each of a plurality of X-ray pixels, in the CE-ROI the texture neighborhood for a given X-ray pixel of the plurality of X-ray pixels extending to a bounding pixel radius of BPR pixels from the given pixel;
generate a texture feature vector (TF) having components based on the indications of intensity provided by a plurality of X-ray pixels in the CE-ROI that are located within the texture neighborhood; and
use a classifier to classify the texture feature vector TF to determine whether the CE-ROI is malignant.