System and method for medical image analysis and probabilistic diagnosis
View Patent ↗Methods and systems for obtaining a probabilistic diagnosis based on medical images of a patient are disclosed. In exemplary embodiments, such methods include the steps of scanning some or all of a patient to obtain a medical image; evaluating the medical image for one or more designated key features; assigning discrete values to the one or more designated key features to form a patient scan key feature pattern; and transmitting the values of the one or more designated key features to a processor programmed to match the patient scan key feature pattern to one or more known disease-specific key feature patterns to create a probabilistic diagnosis and transmit the probabilistic diagnosis to a user. Associated systems include a medical imaging device that is capable of producing a medical image of some or all of a patient and a processor programmed to create a probabilistic diagnosis.
1. A method for obtaining a probabilistic diagnosis of a patient, the method comprising:
a medical imaging device scanning some or all of a patient to obtain a medical image;
identifying one or more designated key features in the medical image;
assigning discrete values to the one or more designated key features in the medical image to form a patient scan key feature pattern;
transmitting the patient scan key feature pattern to a processor;
the processor implementing an expert system to match the patient scan key feature pattern to one or more known disease specific key feature patterns to create a probabilistic diagnosis; and
the processor transmitting the probabilistic diagnosis to a user.
2. The method according to claim 1 , wherein the identifying step is performed by a human.
3. The method according to claim 1 , wherein the identifying step is performed by a computer.
4. The method according to claim 1 , wherein the values assigned to key features are quantitative.
5. The method according to claim 1 , wherein the values assigned to key features are semi-quantitative.
6. A method for obtaining a probabilistic diagnosis of a patient, the method comprising:
a medical imaging device scanning some or all of a patient to obtain a medical image;
transmitting the medical image to a processor programmed to create a probabilistic diagnosis by:
identifying one or more designated key features in the medical image;
assigning quantitative values to the one or more designated key features in the medical image to form a patient scan key feature pattern;
implementing an expert system that matches the patient scan key feature pattern to one or more known disease specific key feature patterns to create a probabilistic diagnosis; and
transmitting the probabilistic diagnosis to a user.
7. A system for obtaining a probabilistic diagnosis of a patient, the system comprising:
a medical imaging device adapted to produce a medical image of some or all of a patient; and
a processor that implements an expert system to create a probabilistic diagnosis by matching a patient scan key feature pattern comprising one or more quantitative or semi-quantitative values corresponding to designated key features extracted from the medical image to one or more known disease-specific key feature patterns and transmitting the probabilistic diagnosis to a user.
8. The system according to claim 7 , wherein the expert system creates a probabilistic diagnosis by extracting the one or more quantitative or semi-quantitative values corresponding to designated key features from the medical image before matching the patient scan key feature pattern.
9. The system according to claim 8 , wherein the processor is programmed to extract quantitative values corresponding to one or more designated key features by:
identifying one or more designated key features in the medical image; and
assigning quantitative values to the one or more designated key features to form a patient scan key feature pattern.
10. The system according to claim 7 , wherein the values corresponding to the key features are semi-quantitative rank ordered values.
11. The system according to claim 7 , wherein the values corresponding to designated key features are quantitative.
12. The system according to claim 7 , wherein the medical imaging device produces the medical image as a digital image in DICOM format.
13. The system according to claim 7 , wherein the expert system comprises a Bayesian network.
14. The system according to claim 7 , wherein the processor implements a report generator that transmits the probabilistic diagnosis to the user.
15. The system according to claim 7 , wherein the patient scan key feature pattern is created by deconstructing the medical image into a signal domain, a space domain, and a time domain, and the processor implements an algorithm to extract key features from the signal domain, the space domain, and the time domain.
16. The system according to claim 15 , wherein the expert system probabilistically matches key features extracted from the signal domain, the space domain, and the time domain with the one or more known disease-specific key feature patterns.
17. The system according to claim 15 , wherein the processor analyzes each domain separately and in parallel and statistically summarizes the key features of each domain into a Z score that is classified into categorical numerical or verbal descriptors having arbitrary dynamic ranges.
18. The system according to claim 7 , wherein the processor implements a tissue segmentation tool that subdivides the medical image into major tissue types of respective organs in the medical image.