Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking
The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.
1. A computer-implemented method of determining non-calcified plaque from a non-contrast Computed Tomography (CT) image of a coronary region of a subject, the method comprising:
accessing, by a computer system, a non-contrast CT image of a coronary region of a subject;
identifying, by the computer system, epicardial fat on the non-contrast CT image;
segmenting, by the computer system, arteries on the non-contrast CT image using the identified epicardial fat as outer boundaries of the arteries;
identifying, by the computer system, a first set of pixels within the arteries on the non-contrast CT image, the first set of pixels comprising a Hounsfield unit radiodensity value below a predetermined radiodensity threshold;
classifying, by the computer system, the first set of pixels as a first subset of non-calcified plaque;
identifying, by the computer system, a second set of pixels within the arteries on the non-contrast CT image, the second set of pixels comprising a Hounsfield unit radiodensity value within a predetermined radiodensity range;
determining, by the computer system, a heterogeneity index of the second set of pixels and identifying a subset of the second set of pixels comprising a heterogeneity index above a predetermined heterogeneity index threshold, wherein the heterogeneity index of the second set of pixels is determined by generating spatial mapping of Hounsfield unit radiodensity values of the second set of pixels;
classifying, by the computer system, the subset of the second set of pixels as a second subset of non-calcified plaque; and
determining, by the computer system, non-calcified plaque from the non-contrast CT image by combining the first subset of non-calcified plaque and the second subset of non-calcified plaque,
wherein the computer system comprises a computer processor and an electronic storage medium.
2. The computer-implemented method of claim 1 , further comprising classifying, by the computer system, one or more pixels of the second set of pixels comprising a heterogeneity index below the heterogeneity index threshold as blood.
3. The computer-implemented method of claim 1 , wherein the predetermined radiodensity threshold comprises a Hounsfield unit radiodensity value of 30 Hounsfield units.
4. The computer-implemented method of claim 1 , wherein the predetermined radiodensity range comprises Hounsfield unit radiodensity values between 30 and 100 Hounsfield units.
5. The computer-implemented method of claim 1 , wherein identifying epicardial fat on the non-contrast CT image further comprises
determining a Hounsfield unit radiodensity value of each pixel within the non-contrast CT image; and
classifying as epicardial fat one or more pixels within the non-contrast CT image comprising a Hounsfield unit radiodensity value within a predetermined epicardial fat radiodensity range,
wherein the predetermined epicardial fat radiodensity range comprises a Hounsfield unit radiodensity value of −100.
6. The computer-implemented method of claim 1 , wherein the heterogeneity index of the second set of pixels is determined by generating a three-dimensional histogram of Hounsfield unit radiodensity values across a geometric region within the second set of pixels.
7. The computer-implemented method of claim 1 , further comprising determining, by the computer system, calcified plaque from the non-contrast CT image.
8. The computer-implemented method of claim 7 , further comprising generating a quantized color map of the coronary region of the subject by assigning a first color to the identified epicardial fat, assigning a second color to the segmented arteries, assigning a third color to the determined non-calcified plaque, and assigning a fourth color to the determined calcified plaque.
9. The computer-implemented method of claim 1 , further comprising:
identifying, by the computer system, a third set of pixels within the arteries on the non-contrast CT image comprising a Hounsfield unit radiodensity value above a predetermined calcified plaque radiodensity threshold; and
classifying, by the computer system, the third set of pixels as calcified plaque.
10. The computer-implemented method of claim 1 , further comprising determining, by the computer system, a proposed treatment for the subject based at least in part on the determined non-calcified plaque.
11. A computer-implemented method of determining low-attenuated plaque from a medical image of a subject, the method comprising:
accessing, by a computer system, a medical image of a subject, wherein the medical image comprises one or more of a computed tomography (CT) image, magnetic resonance (MR) image, ultrasound image, x-ray image, optical coherence tomography (OCT) image, nuclear medicine image, position-emission tomography (PET) image, single photon emission computed tomography (SPECT) image, or near-field infrared spectroscopy (NIRS) image;
identifying, by the computer system, epicardial fat on the medical image of the subject by:
determining a radiodensity value of each pixel or voxel within the medical image of the subject; and
classifying as epicardial fat one or more pixels or voxels within the medical image of the subject comprising a radiodensity value within a predetermined epicardial fat radiodensity range;
segmenting, by the computer system, arteries on the medical image of the subject using the identified epicardial fat as outer boundaries of the arteries;
identifying, by the computer system, a first set of pixels or voxels within the arteries on the medical image of the subject, the first set of pixels or voxels comprising a radiodensity value below a predetermined radiodensity threshold;
classifying, by the computer system, the first set of pixels or voxels as a first subset of low-attenuated plaque;
identifying, by the computer system, a second set of pixels or voxels within the arteries on the medical image, the second set of pixels or voxels comprising a radiodensity value within a predetermined radiodensity range;
determining, by the computer system, a heterogeneity index of the second set of pixels or voxels and identifying a subset of the second set of pixels or voxels comprising a heterogeneity index above a predetermined heterogeneity index threshold, wherein the heterogeneity index of the second set of pixels or voxels is determined by generating spatial mapping of radiodensity values of the second set of pixels or voxels;
classifying, by the computer system, the subset of the second set of pixels or voxels as a second subset of low-attenuated plaque; and
determining, by the computer system, low-attenuated plaque from the medical image of the subject by combining the first subset of low-attenuated plaque and the second subset of low-attenuated plaque,
wherein the computer system comprises a computer processor and an electronic storage medium.
12. The computer-implemented method of claim 11 , wherein the medical image comprises an image of a coronary region of the subject.
13. The computer-implemented method of claim 11 , further comprising determining, by the computer system, a proposed treatment for a disease based at least in part on the determined low-attenuated plaque.
14. The computer-implemented method of claim 13 , wherein the disease comprises one or more of arterial disease, renal artery disease, abdominal atherosclerosis, carotid atherosclerosis, disease associated with a lower extremity artery, disease associated with an upper extremity artery, or a disease associated with a cerebral artery.
15. A system for determining low-attenuated plaque from a medical image of a subject, the system comprising:
one or more computer non-transitory readable storage devices configured to store a plurality of computer executable instructions; and
one or more hardware computer processors in communication with the one or more computer readable storage devices and configured to execute the plurality of computer executable instructions in order to cause the system to:
access a medical image of a subject, wherein the medical image comprises one or more of a computed tomography (CT) image, magnetic resonance (MR) image, ultrasound image, x-ray image, optical coherence tomography (OCT) image, nuclear medicine image, position-emission tomography (PET) image, single photon emission computed tomography (SPECT) image, or near-field infrared spectroscopy (NIRS) image;
identify epicardial fat on the medical image of the subject by:
determining a radiodensity value of each pixel or voxel within the medical image of the subject; and
classifying as epicardial fat one or more pixels or voxels within the medical image of the subject comprising a radiodensity value within a predetermined epicardial fat radiodensity range;
segment arteries on the medical image of the subject using the identified epicardial fat as outer boundaries of the arteries;
identify a first set of pixels or voxels within the arteries on the medical image of the subject, the first set of pixels or voxels comprising a radiodensity value below a predetermined radiodensity threshold;
classify the first set of pixels or voxels as a first subset of low-attenuated plaque;
identify a second set of pixels or voxels within the arteries on the medical image, the second set of pixels or voxels comprising a radiodensity value within a predetermined radiodensity range;
determine a heterogeneity index of the second set of pixels or voxels and identify a subset of the second set of pixels or voxels comprising a heterogeneity index above a predetermined heterogeneity index threshold, wherein the heterogeneity index of the second set of pixels or voxels is determined by generating spatial mapping of radiodensity values of the second set of pixels or voxels;
classify the subset of the second set of pixels or voxels as a second subset of low-attenuated plaque; and
determine low-attenuated plaque from the medical image of the subject by combining the first subset of low-attenuated plaque and the second subset of low-attenuated plaque.
16. The system of claim 15 , wherein the medical image comprises a non-contrast CT image.
17. A computer-implemented method of determining non-calcified plaque from a medical image of a coronary region of a subject, the method comprising:
accessing, by a computer system, a medical image of a coronary region of a subject, the medical image comprising one or more of a Dual-Energy Computed Tomography (DECT) image, a spectral Computed Tomography (CT) image, or a photon counting detector CT image;
identifying, by the computer system, epicardial fat on the medical image of the subject;
segmenting, by the computer system, arteries on the medical image of the subject;
identifying, by the computer system, a first set of pixels or voxels within the arteries on the medical image of the subject, the first set of pixels or voxels comprising a radiodensity value below a predetermined radiodensity threshold;
classifying, by the computer system, the first set of pixels or voxels as a first subset of non-calcified plaque;
identifying, by the computer system, a second set of pixels or voxels within the arteries on the medical image of the subject, the second set of pixels or voxels comprising a radiodensity value within a predetermined radiodensity range;
classifying, by the computer system, a subset of the second set of pixels or voxels as a second subset of non-calcified plaque, wherein the subset of the second set of pixels or voxels is classified by determining a heterogeneity index of the second set of pixels or voxels and classifying the subset of the second set of pixels or voxels comprising a heterogeneity index above a heterogeneity index threshold, wherein the heterogeneity index of the second set of pixels or voxels is determined by generating spatial mapping of radiodensity values of the second set of pixels or voxels; and
determining, by the computer system, non-calcified plaque from the medical image of the subject by combining the first subset of non-calcified plaque and the second subset of non-calcified plaque,
wherein the computer system comprises a computer processor and an electronic storage medium.