Facilitating artificial intelligence integration into systems using a distributed learning platform

Patent metadata

Field	Value
Publication number	US-12040075-B2
Application number	US-202117187016-A
Country	US
Kind code	B2
Filing date	Feb 26, 2021
Priority date	Dec 31, 2018
Publication date	Jul 16, 2024
Grant date	Jul 16, 2024

Abstract

Techniques are described that facilitate integrating artificial intelligence (AI) informatics in healthcare systems using a distributed learning platform. In one embodiment, a computer-implemented is provided that comprises interfacing, by a system operatively coupled to a processor, with a medical imaging application that provides for viewing medical image data. The method further comprises, facilitating, by the system, generation of structured diagnostic data according to a defined ontology in association with usage of the imaging application to perform a clinical evaluation of the medical image data. The method further comprises providing, by the system, the structured diagnostic data to one or more machine learning systems, wherein based on the providing, the one or more machine learning systems employ the structured diagnostic data as training data to generate or train one or more diagnostic models configured to provide artificial intelligence-based diagnostic evaluations of new medical image data.

First claim

What is claimed is:

1. A system that facilitates integrating artificial intelligence (AI) informatics into healthcare systems, the system comprising:
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a data development component that integrates with a clinical application and provides one or more tools that facilitate generating structured training in association with reviewing medical data via a graphical user interface of the clinical application, wherein the structured training data comprises structured annotation data associated with the medical image data, and wherein the one or more tools comprise:
an annotation component that controls terms for inclusion in the structured annotation data based on a defined ontology for a type of the medical image data, wherein the annotation component determines one or more annotation terms that are relevant to the medical image data and adhere to the defined ontology, provides the one or more annotation terms to an annotator via the graphical user interface, and restricts selection of the terms for inclusion in the structured annotation data by the annotator to the one or more annotation terms, and
a crowd sourcing component that initiates a poll to a group of specialists requesting respective diagnostic interpretations of the medical image data in response to a confidence score associated with an automated interpretation of the medical image data generated via one or more diagnostic models being below a threshold confidence score; and

a data collection component that aggregates the structured training data provides the structured training data to a plurality of AI model development subsystems that train different AI models to generate different inference outputs using different subsets of the structured training data.

2. The system of claim 1 , wherein the computer executable components further comprise:
a workflow assistance component that facilitates identifying and calling one or more of the different AI models applicable to new medical image data accessed via the clinical application, applying the one or more AI models to the new medical image data to generate one or more inference outputs, and providing the one or more inference outputs for review in association with usage of the clinical application to facilitate performance of a clinical workflow.

3. The system of claim 2 , wherein the one or more AI models are configured to determine at least one of, feature information that identifies and characterizes one or more defined features reflected in the medical image data or diagnosis information regarding a medical condition reflected in the medical image data in accordance with the defined ontology.

4. The system of claim 3 , wherein the workflow assistance component applies the one or more models to the new medical image data to generate at least one of, the feature information or the diagnosis information in accordance with the defined ontology.

5. The system of claim 4 , wherein the data development component further facilitates receiving feedback information representative of an interpretation of the accuracy of the feature information or the diagnosis information in accordance with the defined ontology.

6. The system of claim 5 , wherein the data development component generates additional structured training data that associates the new medical image data with the feedback information and the accuracy information or the diagnosis information.

7. The system of claim 6 , wherein corresponding subsystems for the one or more AI models of the plurality of AI model development subsystems refine the one or more AI models based on the additional structured training data using one or more machine learning techniques.

8. A method for integrating artificial intelligence (AI) informatics into healthcare systems, the system characterized by a distributed learning architecture, the method comprising:
integrating, by a system operatively coupled to a processor, with a clinical application and providing one or more functions that facilitate generating structured training in association with reviewing medical data via a graphical user interface of the clinical application, wherein the structured training data comprises structured annotation data associated with the medical image data, and wherein the one or more functions comprise:
an annotation function that controls textual terms for inclusion in the structured annotation data based on a defined ontology for a type of the medical image data, wherein the annotation function determines one or more annotation terms that are relevant to the medical image data and adhere to the defined ontology, provides the one or more annotation terms to an annotator via the graphical user interface, and restricts selection of the textual terms for inclusion in the structured annotation data by the annotator to the one or more annotation terms, and
a crowd sourcing function that initiates a poll to a group of specialists requesting respective diagnostic interpretations of the medical image data in response to a confidence score associated with an automated interpretation of the medical image data generated via one or more diagnostic models being below a threshold confidence score; and

providing, by the system, the structured training data to a plurality of AI model development subsystems for usage thereof to train different AI models to generate different inference outputs using different subsets of the structured training data.

9. The method of claim 8 , further comprising:
identifying and calling, by the system, one or more of the different AI models applicable to new medical image data accessed via the clinical application;
applying, by the system, the one or more AI models to the new medical image data to generate one or more inference outputs; and
providing, by the system, the one or more inference outputs for review in association with usage of the clinical application to facilitate performance of a clinical workflow.

10. The method of claim 9 , wherein the one or more AI models are configured to determine at least one of, feature information that identifies and characterizes one or more defined features reflected in the new medical image data or diagnosis information regarding a medical condition reflected in the new medical image data in accordance with the defined ontology.

11. The method of claim 9 , further comprising:
receiving, by the system, feedback information representative of an interpretation of the accuracy of the one or more inference outputs;
generating, by the system, additional structured training data that associates the feedback information with the one or more corresponding inference outputs; and
storing, by the system, the structure training data and the additional structured training data in a network accessible structured training database.

12. The method of claim 11 , further comprising:
providing, by the system, the plurality of AI model development subsystems access to the network accessible structured training database for refining the AI models.

13. The system of claim 1 , wherein the structured training data comprises measurement data that defines a measurement for one or more features in the medical image data, and wherein the one or more tools comprise:
a measurement component that generates the measurement data based on a measurement marking applied to the one or more features in the medical image data as displayed via the graphical user interface.

14. The system of claim 1

Assignees

• Ge Prec Healthcare Llc

Inventors

• Kalafut John
• Dreyer Keith
• Michalski Mark
• Pomerantz Stuart
• Doyle Sean
• Tenenholtz Neil

Classifications

• G06T11/60

  Creating or editing images; Combining images with text · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G06F40/30

  Semantic analysis · CPC title

• G06N20/00

  Machine learning · CPC title

• G16H50/70

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