System and Method for Dynamically Adaptable Learning Medical Diagnosis System

• US 20100063947A1
• Filed: 05/15/2009
• Published: 03/11/2010
• Est. Priority Date: 05/16/2008
• Status: Abandoned Application

• Current Assignee: Wisconsin Alumni Research Foundation (University of Wisconsin System)
• Original Assignee: Wisconsin Alumni Research Foundation (University of Wisconsin System)
• Inventors: Burnside, Elizabeth S., Page, Charles D., de Morais Santos Costa, Vitor Manuel, Davis, Jesse J.
• Application Number:
  US12/467,140
• Publication Number:
  US 20100063947A1
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 30/20 for handling medical images, e.g. DICOM, HL7 or PACS
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A system and method for determining a likelihood of a disease presence in a particular patient includes a patient history database containing records. Each record includes a plurality of data fields related to a particular patient. An analyzing network is provided having access to the patient history database and having features based on the plurality of data fields included in the records to analyze the plurality of data fields and determine a likelihood of disease presence based on the plurality of features. A learning network is provided that has access to the analyzing network to review the likelihood of disease presence determined by the analyzing network and the plurality of data fields included in the records and automatically identify, evaluate, and add new features to the analyzing network that improve determinations of a likelihood of the disease.

Claims

1. A system for determining a likelihood of a disease:

a patient history database containing records each having a plurality of data fields related to a particular patient;

an analyzing network having access to the patient history database and having a plurality of features based on the plurality of data fields included in the records to analyze the plurality of data fields to determine a likelihood of the disease based on the plurality of features; and

a learning network having access to the analyzing network to review the likelihood of the disease determined by the analyzing network and the plurality of data fields included in the records and automatically identify, evaluate, and add new features to the analyzing network that at least improve determinations of a likelihood of the disease.

2. The system of claim 1 wherein the learning network is configured to only add a new feature to the analyzing network if the new feature would improve the results of determining a likelihood of the disease by greater than a threshold.

3. The system of claim 2 wherein the threshold includes a two percent increase in detection results.

4. The system of claim 2 wherein the threshold includes improving the results of determining a likelihood of the disease in at least five percent of the determinations.

5. The system of claim 1 wherein the at least one of the plurality of data fields related to a particular patient includes a medical image.

6. The system of claim 5 wherein the medical image includes a mammogram and another of the plurality of data fields includes a CAD report.

7. The system of claim 1 wherein the learning network is further configured to add new data fields to the database to store data corresponding to the new features.

8. The system of claim 1 wherein the analyzing network includes a Bayesian network.

9. The system of claim 1 wherein the analyzing network can be selectively enabled and disabled.

10. The system of claim 1 wherein the disease is cancer.

11. A method for developing a system for determining a likelihood of a disease:

providing a database of patient records;

building a Bayesian network to access the database of patient records, analyze a particular patient record in the database, and provide a likelihood of the disease in a patient corresponding to the particular patient record; and

automatically augmenting the Bayesian network using a learning network having access to the Bayesian network to review the likelihood of the disease determined by the analyzing network and the patient records, wherein the augmentation includes adding new features to the Bayesian network that improve determinations of a likelihood of the disease.

12. The method of claim 11 wherein the step of adding new features to the Bayesian network includes adding new features not corresponding to fields in the patient records.

13. The method of claim 12 further comprising adding fields to the patient records corresponding to the new features.

14. The method of claim 11 wherein the addition of a new feature to the analyzing network is only performed if the new feature would improve the results of determining a likelihood of the disease by greater than a threshold.

15. A system for determining a disease state:

a patient history database containing records each having a plurality of data fields related to a particular patient;

a Bayesian network having access to the patient history database and having a plurality of features based on the plurality of data fields included in the records to analyze the plurality of data fields and determine a disease state of a particular patient; and

a learning network having access to the Bayesian network to review the determined disease state and the plurality of data fields included in the records and automatically identify and evaluate potential new features that, if added to the Bayesian network, would improve determinations of the disease state.

16. The system of claim 15 wherein the learning network is configured to only add a potential new feature to the Bayesian network if the potential new feature would improve the determinations of the disease state by greater than a threshold amount.

17. The system of claim 16 wherein the threshold includes a two percent increase in the proper determination of the disease state.

18. The system of claim 15 wherein the threshold includes causing an increase in the proper determination of the disease state in at least five percent of the historical determinations.

19. The system of claim 15 wherein the patient history database includes Breast Imaging Reporting and Data System (BI-RADS) information.

20. The system of claim 15 wherein the learning network utilizes a score as you use (SAYU) protocol.