Quantitative imaging for detecting histopathologically defined plaque fissure non-invasively
Systems and methods for analyzing pathologies utilizing quantitative imaging are presented herein. Advantageously, the systems and methods of the present disclosure utilize a hierarchical analytics framework that identifies and quantify biological properties/analytes from imaging data and then identifies and characterizes one or more pathologies based on the quantified biological properties/analytes. This hierarchical approach of using imaging to examine underlying biology as an intermediary to assessing pathology provides many analytic and processing advantages over systems and methods that are configured to directly determine and characterize pathology from underlying imaging data.
1. A method for computer aided detection of fissure for a pathology using an enriched radiological dataset, the method comprising:
receiving a radiological dataset for a patient, wherein the radiological dataset is obtained non-invasively;
enriching the dataset by performing analyte measurement and/or classification of one or more of: (i) anatomic structure, (ii) shape or geometry or (iii) tissue characteristic, type or character, with objective validation for a set of analytes relevant to a pathology, wherein the analyte measurement and/or classification of anatomic structure, shape, or geometry and/or tissue characteristic, type, or character includes semantic segmentation to identify and classify regions of interest in the radiological dataset, wherein the regions of interest are identified with respect to cross-sections of a tubular structure in the radiological dataset; and
using a machine learned classification approach based on known ground truths to process the enriched dataset and determine a fissure for the pathology.
2. The method of claim 1 , wherein enriching the dataset further includes spatial transformations of the dataset to accentuate biologically-significant spatial context.
3. The method of claim 1 , wherein enriching the dataset includes both (i) semantic segmentation to identify and classify regions of interest in cross-sections of a tubular structure in the radiological dataset to produce an annotated dataset and (ii) spatial transformation by spatial unwrapping the annotated dataset with respect to the cross-sections of a tubular structure to produce an unwrapped dataset.
4. The method of claim 1 , wherein an image volume in the radiological dataset is preprocessed to form a region of interest containing a physiological target, lesion, and/or set of lesions that is to be analyzed.
5. The method of claim 4 , wherein the region of interest and/or the physiological target, lesion, and/or set of lesions are at least one of (i) automatically determined from the radiological dataset or (ii) identified by a user from the radiological dataset.
6. The method of claim 4 , wherein the region of interest includes one or more cross sections, each composed of projections through that volume.
7. The method of claim 4 , wherein the pre-processing the image volume includes deblurring or restoring using a patient-specific point spread determination algorithm to mitigate artifacts or image limitations that result from the image formation process.
8. The method of claim 1 , wherein the machine learned classification approach is use of a trained convolutional neural network (CNN), wherein the CNN is based on a refactoring of AlexNET, Inception, CaffeNet, or other open source or commercially available framework.
9. The method of claim 1 , wherein the dataset is enriched by visually using different colors to represent different analyte sub-regions.
10. The method of claim 1 , wherein dataset enrichment includes ground truth annotation of analyte sub-regions as well as providing a spatial context of how such analytes sub-regions present in cross-section.
11. The method of claim 10 , wherein the spatial context provides a common basis for analysis of enhanced dataset relative to histological cross-sections.
12. The method of claim 10 , wherein the spatial context includes providing a coordinate system based on polar coordinates relative to a centroid of each cross-section.
13. The method of claim 1 , wherein dataset enrichment includes ground truth annotation of analyte sub-regions using ex vivo classification independent of or in conjunction with image-based classification based on a common spatial context between the radiological dataset and ex vivo data.
14. The method of claim 13 , wherein enriched dataset is visualized using different colors to represent different analyte sub-regions, wherein colors for visualizing the enhanced dataset are selected to correspond to colors utilized in the ex vivo data.
15. The method of claim 13 , wherein the enriched dataset includes a visual overlay of ex-vivo data over the radiological data.
16. The method of claim 1 , wherein the pathology is related to the vasculature, wherein the machine learned classification approach provides for atherosclerotic plaque fissure.
17. The method of claim 16 , wherein a ground truth basis used in training the atherosclerotic plaque fissure is the Stary, Virmani, or other clinically accepted plaque typing system.
18. The method of claim 16 , wherein plaque is classified as either stable or unstable.
19. The method of claim 16 , wherein a ground truth basis used in training the atherosclerotic plaque fissure involves calcification, lipid-rich, necrotic, hemorrhagic, fibrotic, ulcerated, remodeled, thrombosed, and/or other atherogenic characteristic, singly or in combination.
20. The method of claim 1 , wherein the radiological dataset includes computed tomography (CT), dual energy computed tomography (DECT), spectral computed tomography (spectral CT), computed tomography angiography (CTA), cardiac computed tomography angiography (CCTA), magnetic resonance imaging (MRI), multi-contrast magnetic resonance imaging (multi-contrast MRI), ultrasound (US), positron emission tomography (PET), intra-vascular ultrasound (IVUS), optical coherence tomography (OCT), near-infrared radiation spectroscopy (NIRS), and/or single-photon emission tomography (SPECT) diagnostic images.
21. The method of claim 20 , wherein enriching the dataset includes using image deblurring or restoring is used to identify lesions of interest and extract plaque composition quantitatively.
22. The method of claim 20 , wherein enriching the dataset includes spatially transforming cross-sectional segmented images into an ‘unwrapped’ reference frame.
23. The method of claim 1 , further comprising using a machine learned classification approach based on known ground truths to process the enriched dataset and determine a predictive outcome related to the pathology.