No Cover Image

E-Thesis 557 views 188 downloads

Digital Twin of Cardiovascular Systems / NEERAJ CHAKSHU

Swansea University Author: NEERAJ CHAKSHU

  • 2023_Chakshu_NK.final.63221.pdf

    PDF | E-Thesis - restricted access

    Copyright: The Author, Neeraj K. Chakshu, 2023. Distributed under the terms of a Creative Commons Attribution-ShareAlike 4.0 International License (CC-BY-SA).

    Download (7.76MB)

DOI (Published version): 10.23889/SUthesis.63221

Abstract

Patient specific modelling using numerical methods is widely used in understanding diseases and disorders. It produces medical analysis based on the current state of patient’s health. Concurrently, as a parallel development, emerging data driven Artificial Intelligence (AI) has accelerated patient c...

Full description

Published: Swansea, Wales, UK 2021
Institution: Swansea University
Degree level: Doctoral
Degree name: Ph.D
Supervisor: Sazonov, Igor. and Nithiarasu, Perumal.
URI: https://cronfa.swan.ac.uk/Record/cronfa63221
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2023-04-27T10:26:35Z
last_indexed 2023-04-27T10:26:35Z
id cronfa63221
recordtype RisThesis
fullrecord <?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>63221</id><entry>2023-04-20</entry><title>Digital Twin of Cardiovascular Systems</title><swanseaauthors><author><sid>bc2dc7d21001bbc11a29dc6dc1d16813</sid><firstname>NEERAJ</firstname><surname>CHAKSHU</surname><name>NEERAJ CHAKSHU</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-04-20</date><abstract>Patient specific modelling using numerical methods is widely used in understanding diseases and disorders. It produces medical analysis based on the current state of patient’s health. Concurrently, as a parallel development, emerging data driven Artificial Intelligence (AI) has accelerated patient care. It provides medical analysis using algorithms that rely upon knowledge from larger human population data. AI systems are also known to have the capacity to provide a prognosis with overallaccuracy levels that are better than those provided by trained professionals. When these two independent and robust methods are combined, the concept of human digital twins arise. A Digital Twin is a digital replica of any given system or process. They combine knowledge from general data with subject oriented knowledge for past, current and future analyses and predictions. Assumptions made during numerical modelling are compensated using knowledge from general data. For humans, they can provide an accurate current diagnosis as well as possible future outcomes. This allows forprecautions to be taken so as to avoid further degradation of patient’s health.In this thesis, we explore primary forms of human digital twins for the cardiovascular system, that are capable of replicating various aspects of the cardiovascular system using different types of data. Since different types of medical data are available, such as images, videos and waveforms, and the kinds of analysis required may be offline or online in nature, digital twin systems should be uniquely designed to capture each type of data for different kinds of analysis. Therefore, passive, active and semi-active digital twins, as the three primary forms of digital twins, for different kinds of applications are proposed in this thesis. By the virtue of applications and the kind of data involved ineach of these applications, the performance and importance of human digital twins for the cardiovascular system are demonstrated. The idea behind these twins is to allow for the application of the digital twin concept for online analysis, offline analysis or a combination of the two in healthcare. In active digital twins active data, such as signals, is analysed online in real-time; in semi-active digital twin some of the components being analysed are active but the analysis itself is carried out offline; and finally, passive digital twins perform offline analysis of data that involves no active component.For passive digital twin, an automatic workflow to calculate Fractional Flow Reserve (FFR) is proposed and tested on a cohort of 25 patients with acceptable results. For semi-active digital twin, detection of carotid stenosis and its severity using face videos is proposed and tested with satisfactory results from one carotid stenosis patient and a small cohort of healthy adults. Finally, for the active digital twin, an enabling model is proposed using inverse analysis and its application in the detection of Abdominal Aortic Aneurysm (AAA) and its severity, with the help of a virtual patient database. This enabling model detected artificially generated AAA with an accuracy as high as 99.91% and classified its severity with acceptable accuracy of 97.79%. Further, for active digital twin, a truly active model is proposed for continuous cardiovascular state monitoring. It is tested on a small cohort of five patients from a publicly available database for three 10-minute periods, wherein this model satisfactorily replicated and forecasted patients’ cardiovascular state. In addition to the three forms of human digital twins for the cardiovascular system, an additional work on patient prioritisation in pneumonia patients for ITU care using data-driven digital twin is also proposed. The severity indices calculated by these models are assessed using the standard benchmark of Area Under Receiving Operating Characteristic Curve (AUROC). The results indicate that using these models, the ITU and mechanical ventilation can be prioritised correctly to an AUROC value as high as 0.89.</abstract><type>E-Thesis</type><journal/><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication>Swansea, Wales, UK</placeOfPublication><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords>Digital Twin, Computational Mechanics, Systemic circulation, Blood flow, Artificial Intelligence, Deep Learning, Computer Vision, CFD, Cardiovascular system.</keywords><publishedDay>10</publishedDay><publishedMonth>11</publishedMonth><publishedYear>2021</publishedYear><publishedDate>2021-11-10</publishedDate><doi>10.23889/SUthesis.63221</doi><url/><notes/><college>COLLEGE NANME</college><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><supervisor>Sazonov, Igor. and Nithiarasu, Perumal.</supervisor><degreelevel>Doctoral</degreelevel><degreename>Ph.D</degreename><degreesponsorsfunders>College of Engineering (College of Engineering Centenary Scholarship)</degreesponsorsfunders><apcterm/><funders/><projectreference/><lastEdited>2024-04-22T15:26:49.3828638</lastEdited><Created>2023-04-20T11:27:40.2396869</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering</level></path><authors><author><firstname>NEERAJ</firstname><surname>CHAKSHU</surname><order>1</order></author></authors><documents><document><filename>63221__27238__dfd073e00c87401cb0d80baf30b34ecb.pdf</filename><originalFilename>2023_Chakshu_NK.final.63221.pdf</originalFilename><uploaded>2023-04-27T11:46:04.4180602</uploaded><type>Output</type><contentLength>8133454</contentLength><contentType>application/pdf</contentType><version>E-Thesis - restricted access</version><cronfaStatus>true</cronfaStatus><embargoDate>2022-11-11T00:00:00.0000000</embargoDate><documentNotes>Copyright: The Author, Neeraj K. Chakshu, 2023. Distributed under the terms of a Creative Commons Attribution-ShareAlike 4.0 International License (CC-BY-SA).</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by-sa/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling v2 63221 2023-04-20 Digital Twin of Cardiovascular Systems bc2dc7d21001bbc11a29dc6dc1d16813 NEERAJ CHAKSHU NEERAJ CHAKSHU true false 2023-04-20 Patient specific modelling using numerical methods is widely used in understanding diseases and disorders. It produces medical analysis based on the current state of patient’s health. Concurrently, as a parallel development, emerging data driven Artificial Intelligence (AI) has accelerated patient care. It provides medical analysis using algorithms that rely upon knowledge from larger human population data. AI systems are also known to have the capacity to provide a prognosis with overallaccuracy levels that are better than those provided by trained professionals. When these two independent and robust methods are combined, the concept of human digital twins arise. A Digital Twin is a digital replica of any given system or process. They combine knowledge from general data with subject oriented knowledge for past, current and future analyses and predictions. Assumptions made during numerical modelling are compensated using knowledge from general data. For humans, they can provide an accurate current diagnosis as well as possible future outcomes. This allows forprecautions to be taken so as to avoid further degradation of patient’s health.In this thesis, we explore primary forms of human digital twins for the cardiovascular system, that are capable of replicating various aspects of the cardiovascular system using different types of data. Since different types of medical data are available, such as images, videos and waveforms, and the kinds of analysis required may be offline or online in nature, digital twin systems should be uniquely designed to capture each type of data for different kinds of analysis. Therefore, passive, active and semi-active digital twins, as the three primary forms of digital twins, for different kinds of applications are proposed in this thesis. By the virtue of applications and the kind of data involved ineach of these applications, the performance and importance of human digital twins for the cardiovascular system are demonstrated. The idea behind these twins is to allow for the application of the digital twin concept for online analysis, offline analysis or a combination of the two in healthcare. In active digital twins active data, such as signals, is analysed online in real-time; in semi-active digital twin some of the components being analysed are active but the analysis itself is carried out offline; and finally, passive digital twins perform offline analysis of data that involves no active component.For passive digital twin, an automatic workflow to calculate Fractional Flow Reserve (FFR) is proposed and tested on a cohort of 25 patients with acceptable results. For semi-active digital twin, detection of carotid stenosis and its severity using face videos is proposed and tested with satisfactory results from one carotid stenosis patient and a small cohort of healthy adults. Finally, for the active digital twin, an enabling model is proposed using inverse analysis and its application in the detection of Abdominal Aortic Aneurysm (AAA) and its severity, with the help of a virtual patient database. This enabling model detected artificially generated AAA with an accuracy as high as 99.91% and classified its severity with acceptable accuracy of 97.79%. Further, for active digital twin, a truly active model is proposed for continuous cardiovascular state monitoring. It is tested on a small cohort of five patients from a publicly available database for three 10-minute periods, wherein this model satisfactorily replicated and forecasted patients’ cardiovascular state. In addition to the three forms of human digital twins for the cardiovascular system, an additional work on patient prioritisation in pneumonia patients for ITU care using data-driven digital twin is also proposed. The severity indices calculated by these models are assessed using the standard benchmark of Area Under Receiving Operating Characteristic Curve (AUROC). The results indicate that using these models, the ITU and mechanical ventilation can be prioritised correctly to an AUROC value as high as 0.89. E-Thesis Swansea, Wales, UK Digital Twin, Computational Mechanics, Systemic circulation, Blood flow, Artificial Intelligence, Deep Learning, Computer Vision, CFD, Cardiovascular system. 10 11 2021 2021-11-10 10.23889/SUthesis.63221 COLLEGE NANME COLLEGE CODE Swansea University Sazonov, Igor. and Nithiarasu, Perumal. Doctoral Ph.D College of Engineering (College of Engineering Centenary Scholarship) 2024-04-22T15:26:49.3828638 2023-04-20T11:27:40.2396869 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering NEERAJ CHAKSHU 1 63221__27238__dfd073e00c87401cb0d80baf30b34ecb.pdf 2023_Chakshu_NK.final.63221.pdf 2023-04-27T11:46:04.4180602 Output 8133454 application/pdf E-Thesis - restricted access true 2022-11-11T00:00:00.0000000 Copyright: The Author, Neeraj K. Chakshu, 2023. Distributed under the terms of a Creative Commons Attribution-ShareAlike 4.0 International License (CC-BY-SA). true eng https://creativecommons.org/licenses/by-sa/4.0/
title Digital Twin of Cardiovascular Systems
spellingShingle Digital Twin of Cardiovascular Systems
NEERAJ CHAKSHU
title_short Digital Twin of Cardiovascular Systems
title_full Digital Twin of Cardiovascular Systems
title_fullStr Digital Twin of Cardiovascular Systems
title_full_unstemmed Digital Twin of Cardiovascular Systems
title_sort Digital Twin of Cardiovascular Systems
author_id_str_mv bc2dc7d21001bbc11a29dc6dc1d16813
author_id_fullname_str_mv bc2dc7d21001bbc11a29dc6dc1d16813_***_NEERAJ CHAKSHU
author NEERAJ CHAKSHU
author2 NEERAJ CHAKSHU
format E-Thesis
publishDate 2021
institution Swansea University
doi_str_mv 10.23889/SUthesis.63221
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering
document_store_str 1
active_str 0
description Patient specific modelling using numerical methods is widely used in understanding diseases and disorders. It produces medical analysis based on the current state of patient’s health. Concurrently, as a parallel development, emerging data driven Artificial Intelligence (AI) has accelerated patient care. It provides medical analysis using algorithms that rely upon knowledge from larger human population data. AI systems are also known to have the capacity to provide a prognosis with overallaccuracy levels that are better than those provided by trained professionals. When these two independent and robust methods are combined, the concept of human digital twins arise. A Digital Twin is a digital replica of any given system or process. They combine knowledge from general data with subject oriented knowledge for past, current and future analyses and predictions. Assumptions made during numerical modelling are compensated using knowledge from general data. For humans, they can provide an accurate current diagnosis as well as possible future outcomes. This allows forprecautions to be taken so as to avoid further degradation of patient’s health.In this thesis, we explore primary forms of human digital twins for the cardiovascular system, that are capable of replicating various aspects of the cardiovascular system using different types of data. Since different types of medical data are available, such as images, videos and waveforms, and the kinds of analysis required may be offline or online in nature, digital twin systems should be uniquely designed to capture each type of data for different kinds of analysis. Therefore, passive, active and semi-active digital twins, as the three primary forms of digital twins, for different kinds of applications are proposed in this thesis. By the virtue of applications and the kind of data involved ineach of these applications, the performance and importance of human digital twins for the cardiovascular system are demonstrated. The idea behind these twins is to allow for the application of the digital twin concept for online analysis, offline analysis or a combination of the two in healthcare. In active digital twins active data, such as signals, is analysed online in real-time; in semi-active digital twin some of the components being analysed are active but the analysis itself is carried out offline; and finally, passive digital twins perform offline analysis of data that involves no active component.For passive digital twin, an automatic workflow to calculate Fractional Flow Reserve (FFR) is proposed and tested on a cohort of 25 patients with acceptable results. For semi-active digital twin, detection of carotid stenosis and its severity using face videos is proposed and tested with satisfactory results from one carotid stenosis patient and a small cohort of healthy adults. Finally, for the active digital twin, an enabling model is proposed using inverse analysis and its application in the detection of Abdominal Aortic Aneurysm (AAA) and its severity, with the help of a virtual patient database. This enabling model detected artificially generated AAA with an accuracy as high as 99.91% and classified its severity with acceptable accuracy of 97.79%. Further, for active digital twin, a truly active model is proposed for continuous cardiovascular state monitoring. It is tested on a small cohort of five patients from a publicly available database for three 10-minute periods, wherein this model satisfactorily replicated and forecasted patients’ cardiovascular state. In addition to the three forms of human digital twins for the cardiovascular system, an additional work on patient prioritisation in pneumonia patients for ITU care using data-driven digital twin is also proposed. The severity indices calculated by these models are assessed using the standard benchmark of Area Under Receiving Operating Characteristic Curve (AUROC). The results indicate that using these models, the ITU and mechanical ventilation can be prioritised correctly to an AUROC value as high as 0.89.
published_date 2021-11-10T15:26:45Z
_version_ 1797045359946498048
score 11.03559

Similar Items