AI-based image analysis for the detection of normal images

Patent metadata

Field	Value
Publication number	US-11935230-B2
Application number	US-202117237160-A
Country	US
Kind code	B2
Filing date	Apr 22, 2021
Priority date	Jun 3, 2020
Publication date	Mar 19, 2024
Grant date	Mar 19, 2024

Abstract

A system and method for identifying abnormal medical images. The system can be configured to receive a medical image, segment an anatomical structure from the medical image to define a segmented dataset, register the segmented dataset to a baseline dataset defining a normal anatomical structure, classify, by an abnormality classifier, whether the anatomical structure within the medical image as either abnormal or normal, wherein the abnormality classifier comprises a machine learning algorithm trained to distinguish between normal and abnormal versions of the anatomical structure in medical images, and based on whether the anatomical structure can be segmented from the medical image, whether the segmented dataset can be registered to the baseline dataset, or a classification associated with the medical image output by the abnormality classifier, flagging the medical image as either normal or abnormal.

First claim

The invention claimed is:

1. A computer-implemented method for identifying an abnormal medical image, the method comprising:
receiving a medical image;
segmenting an anatomical structure from the medical image to define a segmented dataset;
registering the segmented dataset to a baseline dataset defining a normal anatomical structure;
classifying, by an abnormality classifier, whether the anatomical structure within the medical image as either abnormal or normal, wherein the abnormality classifier comprises a machine learning algorithm trained to distinguish between normal and abnormal versions of the anatomical structure in medical images; and
based on whether the anatomical structure can be segmented from the medical image, whether the segmented dataset can be registered to the baseline dataset, or a classification associated with the medical image output by the abnormality classifier, flagging the medical image as either normal or abnormal; and
based on whether the medical image has been flagged as normal or abnormal, controlling whether the medical image is in a worklist associated with a radiologist, wherein the worklist is associated with a software platform, the software platform comprising at least one of a picture archiving and communication system, a radiology information system, or a clinical information system.

2. The method of claim 1 , further comprising:
classifying, by a lesion-specific classifier, whether the anatomical structure exhibits a lesion, wherein the lesion-specific classifier comprises a second machine learning algorithm trained to identify the lesion in medical images; and
wherein flagging the medical image as either normal or abnormal is further based on whether the lesion-specific classifier has identified the lesion in the medical image.

3. The method of claim 2 , wherein the lesion-specific classifier comprises at least one of a support vector machine, a decisions tree, a neural network, or a Bayes' model.

4. The method of claim 1 , wherein the medical image comprises at least one of a computed tomography (CT) image, a magnetic resonance imaging image, a positron emission tomography image, a single-photon emission CT image, an X-ray image, an ultrasound image, an optical coherence tomography image, or a photoacoustic image.

5. The method of claim 1 , wherein the abnormality classifier comprises at least one of a support vector machine, a decisions tree, a neural network, or a Bayes' model.

6. A computer system for identifying an abnormal medical image, the computer system comprising:
a processor; and
a memory coupled to the processor, the memory storing instructions that, when executed by the processor, cause the computer system to:
receive a medical image,
segment an anatomical structure from the medical image to define a segmented dataset,
register the segmented dataset to a baseline dataset defining a normal anatomical structure,
classify, by an abnormality classifier, whether the anatomical structure within the medical image as either abnormal or normal, wherein the abnormality classifier comprises a machine learning algorithm trained to distinguish between normal and abnormal versions of the anatomical structure in medical images, and
based on whether the anatomical structure can be segmented from the medical image, whether the segmented dataset can be registered to the baseline dataset, or a classification associated with the medical image output by the abnormality classifier, flagging the medical image as either normal or abnormal; and
wherein the memory stores further instructions that, when executed by the processor, cause the computer system to:
based on whether the medical image has been flagged as normal or abnormal, remove the medical image from a worklist associated with a radiologist, wherein the worklist is associated with a software platform, the software platform comprising at least one of a picture archiving and communication system, a radiology information system, or a clinical information system.

7. The computer system of claim 6 , further comprising an imaging machine, wherein the medical image is received from the imaging machine.

8. The computer system of claim 6 , wherein the abnormality classifier comprises at least one of a support vector machine, a decisions tree, a neural network, or a Bayes' model.

9. The computer system of claim 6 , wherein the memory stores further instructions that, when executed by the processor, cause the computer system to:
classify, by a lesion-specific classifier, whether the anatomical structure exhibits a lesion, wherein the lesion-specific classifier comprises a second machine learning algorithm trained to identify the lesion in medical images; and
wherein flagging the medical image as either normal or abnormal is further based on whether the lesion-specific classifier has identified the lesion in the medical image.

10. The computer system of claim 9 , wherein the lesion-specific classifier comprises at least one of a support vector machine, a decisions tree, a neural network, or a Bayes' model.

11. The computer system of claim 6 , wherein the medical image comprises at least one of a computed tomography (CT) image, a magnetic resonance imaging image, a positron emission tomography image, a single-photon emission CT image, an X-ray image, an ultrasound image, an optical coherence tomography image, or a photoacoustic image.

Assignees

• Siemens Healthineers Ag

Inventors

• Hölzer Philipp
• Frank Richard
• Schmidt Sebastian
• Sperl Jonathan

Classifications

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G06F18/2163

  Partitioning the feature space · CPC title

• G06F18/2433

  Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00

  Machine learning · CPC title