Automated report generation based on cognitive classification of medical images

Patent metadata

Field	Value
Publication number	US-10916341-B2
Application number	US-201715844249-A
Country	US
Kind code	B2
Filing date	Dec 15, 2017
Priority date	Dec 15, 2017
Publication date	Feb 9, 2021
Grant date	Feb 9, 2021

Abstract

Methods and systems for automatically triaging an image study of a patient generated as part of a medical imaging procedure. One system includes a computing device including an electronic processor. The electronic processor is configured to receive, from a cognitive system applying a model developed using computer vision and machine learning techniques based on deep learning methodology to classify image studies, a classification assigned to the image study using the model, and automatically generate a structured report for the image study based on the classification assigned by the model, the structured report accessible by a radiologist via a structured reporting system.

First claim

What is claimed is:

1. A system for automatically triaging a first image study of a patient generated as part of a medical imaging procedure, the system comprising:
a computing device including an electronic processor configured to
receive, from a cognitive system applying a model developed using computer vision and machine learning techniques based on deep learning methodology to classify image studies, a classification assigned to the first image study using the model,
assign a severity classification to the first image study based on the classification assigned by the model, a classification assigned to a second image study of the patient, and a comparison of an image study with an image included in the second image study, wherein the second image study for the patient was generated using a different modality than the first image study, and
in response to the severity classification being associated with a predetermined level of risk, automatically generate a structured report for the image, the structured report including one or more fillable fields and accessible by a radiologist via a structured reporting system.

2. The system of claim 1 , wherein the classification includes a BI-RADS classification.

3. The system of claim 1 , wherein the second image study is an image study of the patient generated prior to the first image study.

4. The system of claim 1 , wherein the electronic processor is configured to generate the structured report by populating at least one of the one or more fillable fields included in the structured report with the severity classification.

5. The system of claim 1 , wherein the electronic processor is further configured to submit the structured report to the radiologist for review and approval.

6. The system of claim 1 , wherein the electronic processor is configured to generate the structured report by populating at least one of the one or more fillable fields included in the structured report with the classification assigned by the model.

7. The system of claim 1 , wherein the electronic processor is further configured to, in response to the classification representing a second predetermined level of risk, automatically communicate with a resource allocation system to reserve at least one medical resource for treating the patient.

8. The system of claim 7 , wherein the resource allocation system includes a hospital system for reserving at least one selected from a group consisting of staff, a facility, and equipment.

9. The system of claim 7 , wherein the at least one resource includes a resource for performing a biopsy of the patient.

10. The system of claim 1 , wherein the electronic processor is further configured to, in response to the classification representing a second predetermined level of risk, automatically generate a worklist, the worklist prioritizing a plurality of tasks for treating the patient.

11. The system of claim 1 , wherein the electronic processor is further configured to, in response to the classification representing a second predetermined level of risk, automatically determine a differential diagnosis for the patient associated with the image study.

12. The system of claim 1 , wherein the electronic processor is further configured to automatically generate the structured report in response to the classification representing the predetermined level of risk based on the at least one rule associated with at least one selected from a group consisting of the patient, a facility, a radiologist, a network, a geographical area, and a type of imaging modality.

13. Non-transitory computer-readable medium including instructions that, when executed by an electronic processor, perform a set of functions, the set of functions comprising:
receiving, from a cognitive system applying a model developed using computer vision and machine learning techniques based on deep learning methodology to classify image studies, a classification assigned to a first image study using the model;
comparing an image included in the first image study to an image included in a second image study of the patient to determine a patient change, the second image study generated prior to the first image study;
receiving a classification assigned to a third image study of the patient, the third image study generated by a different imaging modality than the first image study;
generating a severity classification for the first image study based on the classification assigned to the first image study, the patient change, and the classification assigned to the third image study; and
in response to the severity classification representing a predetermined level of risk, automatically generating a structured report for the first image, the structured report including one or more fillable fields.

14. A method of automatically analyzing an image study of a patient generated as part of a medical imaging procedure, the method comprising:
receiving, with an electronic processor, a classification from a cognitive system for the image study, the cognitive system applying a model developed using computer vision and machine learning techniques based on deep learning methodology to classify image studies based on a classification schema;
in response to the classification representing a predetermined level of risk, perform at least one action based on the application of at least one rule, wherein the at least one action includes automatically, with the electronic processor, generating a structured report for the image study, the structured report including one or more fillable fields,
wherein the at least one rule is associated with at least one selected from a group consisting of the patient, a facility, a radiologist, a network, a geographical area, and a type of imaging modality.

Assignees

• IBM

Inventors

• Stoval Iii William Murray
• Sati Marwan
• Azabagic Andjela
• Covell Grant

Classifications

• A61B5/7264 Primary

  Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems · CPC title

• G16H30/20 Primary

  for handling medical images, e.g. DICOM, HL7 or PACS · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H30/40

  for processing medical images, e.g. editing · CPC title

• A61B5/0035

  adapted for acquisition of images from more than one imaging mode, e.g. combining MRI and optical tomography · CPC title