Medical evaluation machine learning workflows and processes
Systems and methods for processing electronic imaging data obtained from medical imaging procedures are disclosed herein. Some embodiments relate to data processing mechanisms for medical imaging and diagnostic workflows involving the use of machine learning techniques such as deep learning, artificial neural networks, and related algorithms that perform machine recognition of specific features and conditions in imaging data. In an example, a deep learning model is selected for automated image recognition of a particular medical condition on image data, and applied to the image data to recognize characteristics of the particular medical condition. Based on the characteristics recognized by the automated image recognition on the image data, an electronic workflow for performing a diagnostic evaluation of the medical imaging study may be modified, updated, or prioritized.
1. A method of artificial intelligence data processing for medical imaging data, comprising operations performed by a computing device, with the operations comprising:
obtaining imaging data originating from a medical imaging procedure of a human subject;
classifying, using a trained image recognition model, at least one identifiable condition from at least one image of the imaging data, wherein a selection of the trained image recognition model is provided by selecting the trained image recognition model from a plurality of trained image recognition models, and wherein the selection and operation of the trained image recognition model is based on at least one characteristic of the medical imaging procedure indicated by metadata originating from the medical imaging procedure; and
defining properties of an electronic workflow that performs a diagnostic evaluation of the imaging data from the medical imaging procedure, based on at least one classified characteristic of the identifiable condition.
2. The method of claim 1 , wherein the selection and operation of the trained image recognition model is based on the imaging data generated by the medical imaging procedure in combination with the metadata originating from the medical imaging procedure.
3. The method of claim 2 , wherein the metadata indicates an anatomical area or anatomical feature that is represented in the at least one image, and
wherein the operation of the trained image recognition model on the at least one image is modified based on the anatomical area or anatomical feature that is represented in the at least one image.
4. The method of claim 1 , wherein the classified characteristic of the identifiable condition includes at least one of: an identification of an area in an image in which the identifiable condition is detected, a likelihood of a medical condition being present in an image, a measurement in an image associated with a medical condition, a correlation of a first medical condition to a second medical condition detected in the imaging data, or a measurement of a: frequency, severity, or urgency of a medical condition detected in an image.
5. The method of claim 1 , wherein the properties of the electronic workflow are further defined based on results of natural language processing, the results of natural language processing being produced from analysis of data associated with the medical imaging procedure or data associated with a prior medical imaging procedure of the human subject.
6. The method of claim 1 , wherein the classified characteristic of the identifiable condition includes a detection value that corresponds to a level of feature recognition in the imaging data for a negative or positive finding of the identifiable condition, wherein at least one property in the electronic workflow is modified based on the detection value.
7. The method of claim 1 , wherein defining the electronic workflow includes:
communicating data to establish, with the electronic workflow, a first assignment of the imaging data to a first evaluator; and
communicating data to establish, with the electronic workflow, a second assignment of the imaging data to a second evaluator;
wherein respective properties of the first assignment and the second assignment are defined based on the classified characteristic of the identifiable condition.
8. The method of claim 1 , wherein defining the electronic workflow includes:
determining an assignment, for at least a first part of the imaging data, within a first electronic worklist associated with a first evaluator, wherein the classified characteristic of the identifiable condition includes an indication of an urgent medical condition; and
communicating an indication of the urgent medical condition in connection with the assignment for the first part of the imaging data.
9. The method of claim 8 , wherein the urgent medical condition is indicated as a trauma or other time-sensitive medical condition, and wherein defining the electronic workflow includes determining a second assignment, for at least a second part of the imaging data, within a second electronic worklist associated with a second evaluator, wherein the first part of the imaging data and the second part of the imaging data are identified based on respective anatomical regions captured by the imaging data.
10. The method of claim 1 , further comprising:
identifying, from the imaging data, the at least one image as being associated with an anatomical classification or area, wherein the operation of the trained image recognition model is further based on the anatomical classification or area.
11. A non-transitory machine-readable storage medium including instructions that, when executed by at least one processor of a computing device, causes the computing device to perform operations comprising:
obtaining imaging data originating from a medical imaging procedure of a human subject;
classifying, using a trained image recognition model, at least one identifiable condition from at least one image of the imaging data, wherein a selection of the trained image recognition model is provided by selecting the trained image recognition model from a plurality of trained image recognition models, wherein the selection and operation of the trained image recognition model is based on at least one characteristic of the medical imaging procedure indicated by metadata originating from the medical imaging procedure; and
defining properties of an electronic workflow that performs a diagnostic evaluation of the imaging data from the medical imaging procedure, based on at least one classified characteristic of the identifiable condition.
12. The machine-readable storage medium of claim 11 , wherein the selection and operation of the trained image recognition model is based on the imaging data generated by the medical imaging procedure in combination with the metadata originating from the medical imaging procedure, and wherein the metadata indicates an anatomical area or anatomical feature that is represented in the at least one image.
13. The machine-readable storage medium of claim 11 , wherein the classified characteristic of the identifiable condition includes at least one of: an identification of an area in an image in which the identifiable condition is detected, a likelihood of a medical condition being present in an image, a measurement in an image associated with a medical condition, a correlation of a first medical condition to a second medical condition detected in the imaging data, or a measurement of a: frequency, severity, or urgency of a medical condition detected in an image.
14. The machine-readable storage medium of claim 11 , wherein the properties of the electronic workflow are further defined based on results of natural language processing, the results of natural language processing being produced from analysis of data associated with the medical imaging procedure or data associated with a prior medical imaging procedure of the human subject.
15. The machine-readable storage medium of claim 11 , wherein defining the electronic workflow includes:
communicating data to establish, with the electronic workflow, a first assignment of the imaging data to a first evaluator; and
communicating data to establish, with the electronic workflow, a second assignment of the imaging data to a second evaluator;
wherein respective properties of the first assignment and the second assignment are defined based on the classified characteristic of the identifiable condition; and
wherein the classified characteristic includes a detection value that corresponds to a negative or positive finding of the identifiable condition, and wherein at least one property in the electronic workflow is modified based on the detection value.
16. The machine-readable storage medium of claim 11 , wherein defining the electronic workflow includes:
determining an assignment, for at least a first part of the imaging data, within a first electronic worklist associated with a first evaluator, wherein the classified characteristic of the identifiable condition includes an indication of an urgent medical condition; and
communicating an indication of the urgent medical condition in connection with the assignment for the first part of the imaging data;
determining a second assignment, for at least a second part of the imaging data, within a second electronic worklist associated with a second evaluator;
wherein the first part of the imaging data and the second part of the imaging data are identified based on respective anatomical regions captured by the imaging data.
17. The machine-readable storage medium of claim 11 , the operations further comprising:
identifying, from the imaging data, the at least one image as being associated with an anatomical classification or area, wherein the operation of the trained image recognition model is further based on the anatomical classification or area.
18. A computing system, comprising:
processing circuitry; and
memory comprising instructions stored thereon, which when executed by the processing circuitry, configure the computing system to perform operations comprising:
obtaining imaging data originating from a medical imaging procedure of a human subject;
classifying, using a trained image recognition model, at least one identifiable condition from at least one image of the imaging data, wherein a selection of the trained image recognition model is provided by selecting the trained image recognition model from a plurality of trained image recognition models, and wherein the selection and operation of the trained image recognition model is based on at least one characteristic of the medical imaging procedure indicated by metadata originating from the medical imaging procedure; and
defining properties of an electronic workflow that performs a diagnostic evaluation of the imaging data from the medical imaging procedure, based on at least one classified characteristic of the identifiable condition.
19. The computing system of claim 18 , wherein the selection and operation of the trained image recognition model is based on the image data generated by the medical imaging procedure in combination with the metadata originating from the medical imaging procedure, and wherein the metadata indicates an anatomical area or anatomical feature that is represented in the at least one image.
20. The computing system of claim 18 , wherein the classified characteristic of the identifiable condition includes at least one of: an identification of an area in an image in which the identifiable condition is detected, a likelihood of a medical condition being present in an image, a measurement in an image associated with a medical condition, a correlation of a first medical condition to a second medical condition detected in the imaging data, or a measurement of a: frequency, severity, or urgency of a medical condition detected in an image.
21. The computing system of claim 18 , wherein the properties of the electronic workflow are further defined based on results of natural language processing, the results of natural language processing being produced from analysis of data associated with the medical imaging procedure or data associated with a prior medical imaging procedure of the human subject.
22. The computing system of claim 18 , wherein defining the electronic workflow includes:
communicating data to establish, with the electronic workflow, a first assignment of the imaging data to a first evaluator; and
communicating data to establish, with the electronic workflow, a second assignment of the imaging data to a second evaluator;
wherein respective properties of the first assignment and the second assignment are defined based on the classified characteristic of the identifiable condition; and
wherein the classified characteristic includes a detection value that corresponds to a negative or positive finding of the identifiable condition, and wherein at least one property in the electronic workflow is modified based on the detection value.
23. The computing system of claim 18 , wherein defining the electronic workflow includes:
determining an assignment, for at least a first part of the imaging data, within a first electronic worklist associated with a first evaluator, wherein the classified characteristic of the identifiable condition includes an indication of an urgent medical condition; and
communicating an indication of the urgent medical condition in connection with the assignment for the first part of the imaging data;
determining a second assignment, for at least a second part of the imaging data, within a second electronic worklist associated with a second evaluator;
wherein the first part of the imaging data and the second part of the imaging data are identified based on respective anatomical regions captured by the imaging data.
24. The computing system of claim 18 , the operations further comprising:
identifying, from the imaging data, the at least one image as being associated with an anatomical classification or area, wherein the operation of the trained image recognition model is further based on the anatomical classification or area.