US-12616370-B2 - System and method for determining atherosclerotic pathological tissue types in a coronary artery OCT image using trained engines

Abstract

There is described a system for determining an atherosclerotic pathological tissue type of a coronary artery, the system comprising: an optical coherence tomography (OCT) imaging system being configured for acquiring an OCT image of tissue within said coronary artery; and a controller configured for: using a trained fully convolutional engine stored on said memory and having a plurality of convolutional layers with respective dilation rates different than unity, extracting pathological tissues regardless their type in at least a region of interest of said OCT image; using a trained auto-encoder classification engine stored on said memory and having a layer characterized with a sparsity regularization parameter, determining an atherosclerotic pathological tissue type associated to said region of interest of said OCT image based on said extracted pathological tissues; and outputting said atherosclerotic pathological tissue type of said coronary artery.

Inventors

• Atefeh ABDOLMANAFI
• Luc Duong
• Nagib DAHDAH
• Ragui IBRAHIM

Assignees

• ECOLE DE TECHNOLOGIE SUPERIEURE

Dates

Publication Date

20260505

Application Date

20211005

Claims (16)

1. 1 . A system for determining an atherosclerotic pathological tissue type of a coronary artery, the system comprising: an optical coherence tomography (OCT) imaging system being configured for acquiring an OCT image of tissue within said coronary artery; and a controller having a processor and a memory having instructions stored thereon that when executed by said processor perform the steps of: using a trained fully convolutional engine stored on said memory and having a plurality of convolutional layers with respective dilation rates different than unity, extracting a plurality of pathological tissues regardless their type in at least a region of interest of said OCT image; using a trained auto-encoder classification engine stored on said memory and having a layer characterized with a sparsity regularization parameter, determining an atherosclerotic pathological tissue type associated to said region of interest of said OCT image based on said extracted pathological tissues; and outputting said atherosclerotic pathological tissue type of said coronary artery.
2. 2 . The system of claim 1 wherein said atherosclerotic pathological tissue type is selected from the group consisting of: fibrous plaque, fibrocalcific, fibroatheroma, acute thrombus, and micro-vessels.
3. 3 . The system of claim 1 wherein at least one of said dilation rates is greater than a dilation rate threshold.
4. 4 . The system of claim 3 wherein said dilation rate threshold is the unity.
5. 5 . The system of claim 1 wherein the layer of the auto-encoder classification engine is a hidden layer.
6. 6 . The system of claim 1 wherein said sparsity regularization parameter ranges within a given sparsity regularization parameter range.
7. 7 . The system of claim 1 wherein said outputting includes generating an output image having the atherosclerotic pathological tissue type overlaid over said OCT image, with a lead line leading to the region of interest.
8. 8 . The system of claim 1 wherein upon finding at least one atherosclerotic pathological tissue type in said OCT image, associating a tag indicative of unhealthiness to the OCT image.
9. 9 . A method for determining an atherosclerotic pathological tissue type of a coronary artery, the method comprising: using an optical coherence tomography (OCT) imaging system, acquiring an OCT image of tissue within said coronary artery; using a controller, using a trained fully convolutional engine stored on a memory of said controller and having a plurality of convolutional layers with respective dilation rates, extracting a plurality of pathological tissues regardless their type in at least a region of interest of said OCT image; using a trained auto-encoder classification engine stored on said memory and having a layer characterized with a sparsity regularization parameter, determining an atherosclerotic pathological tissue type associated to said region of interest of said OCT image based on said extracted pathological tissues; and outputting said atherosclerotic pathological tissue type of said coronary artery.
10. 10 . The method of claim 9 wherein said atherosclerotic pathological tissue type is selected from the group consisting of: fibrous plaque, fibrocalcific, fibroatheroma, acute thrombus, and micro-vessels.
11. 11 . The method of claim 9 wherein at least one of said dilation rates is greater than a dilation rate threshold.
12. 12 . The method of claim 11 wherein said dilation rate threshold is the unity.
13. 13 . The method of claim 9 wherein the layer of the auto-encoder classification engine is a hidden layer.
14. 14 . The method of claim 9 wherein said sparsity regularization parameter ranges within a given sparsity regularization parameter range.
15. 15 . The method of claim 7 wherein said outputting includes generating an output image having the atherosclerotic pathological tissue type overlaid over said OCT image, with a lead line leading to the region of interest.
16. 16 . The method of claim 7 wherein upon finding at least one atherosclerotic pathological tissue type in said OCT image, associating a tag indicative of unhealthiness to the OCT image.

Description

FIELD The improvements generally relate to the field of pathological tissue type determination in coronary artery optical coherence tomography (OCT) images, and more specifically involve computer-implemented engines trained using artificial intelligence. BACKGROUND OCT imaging techniques involve the use of coherent light which can penetrate well into coronary artery tissues, thus allowing not only an inside wall of the coronary artery to be imaged, but also allowing imaging of deeper layers of the coronary artery. Although existing OCT imaging techniques are satisfactory to a certain degree, there remains room for improvement. For instance, as acquiring a multitude of OCT images of coronary artery tissue of a patient can be relatively straightforward using existing OCT imaging systems, examining these OCT images can require a significant amount of time from highly qualified physicians. SUMMARY It is an aim of the present disclosure to describe methods and systems suited for determining pathological tissue type(s) associated with OCT images representing a coronary artery of a patient with atherosclerosis. The methods and systems involve the use of computer-implemented trained engines to extract pathological tissues regardless their types from a coronary artery OCT image. Then, the extracted pathological tissues are classified into respective atherosclerotic pathological tissue types. It was found that there was a need for computer-implemented trained engines which can perform the pathological tissue extraction and classification using raw, unprocessed OCT images of the coronary artery of the patient while still achieving satisfactory results. It is noted that the examples described herein have been used for determining and identifying atherosclerotic plaque tissue(s) in one or more regions of interest of a given coronary artery OCT image. In accordance with a first aspect of the present disclosure, there is provided a system for determining an atherosclerotic pathological tissue type of a coronary artery, the system comprising: an OCT imaging system being configured for acquiring an OCT image of tissue within said coronary artery; and a controller having a processor and a memory having instructions stored thereon that when executed by said processor perform the steps of: using a trained fully convolutional engine stored on said memory and having a plurality of convolutional layers with respective dilation rates different than unity, extracting a plurality of pathological tissues regardless their type in at least a region of interest of said OCT image; using a trained auto-encoder classification engine stored on said memory and having a layer characterized with a sparsity regularization parameter, determining an atherosclerotic pathological tissue type associated to said region of interest of said OCT image based on said extracted pathological tissues; and outputting said atherosclerotic pathological tissue type of said coronary artery. Further in accordance with the first aspect of the present disclosure, said atherosclerotic pathological tissue type can for example be selected from the group consisting of: fibrous plaque, fibrocalcific, fibroatheroma, acute thrombus, and micro-vessels. Still further in accordance with the first aspect of the present disclosure, at least one of said dilation rates can for example be greater than a dilation rate threshold. Still further in accordance with the first aspect of the present disclosure, said dilation rate threshold can for example be the unity. Still further in accordance with the first aspect of the present disclosure, the layer of the auto-encoder classification engine can for example be a hidden layer. Still further in accordance with the first aspect of the present disclosure, said sparsity regularization parameter can for example range within a given sparsity regularization parameter range. Still further in accordance with the first aspect of the present disclosure, said outputting can for example include generating an output image having the atherosclerotic pathological tissue type overlaid over said OCT image, with a lead line leading to the region of interest. Still further in accordance with the first aspect of the present disclosure, upon finding at least one atherosclerotic pathological tissue type in said OCT image, the method can for example include a step of associating a tag indicative of unhealthiness to the OCT image. In accordance with a second aspect of the present disclosure, there is provided a method for determining an atherosclerotic pathological tissue type of a coronary artery, the method comprising: using an OCT imaging system, acquiring an OCT image of tissue within said coronary artery; using a controller, using a trained fully convolutional engine stored on a memory of said controller and having a plurality of convolutional layers with respective dilation rates, extracting a plurality of pathological tissues regardless their types in a