Advanced

Patient-specific blood vessel modeling on data from MR scans of children with congenital heart disease

Söderlund, Andreas LU (2016) In Master's Theses in Mathematical Sciences NUMM11 20161
Mathematics (Faculty of Engineering)
Abstract
One percent of children have a congenital heart disease, often requiring surgical corrections. The last fifty years have seen a sharp increase in the use of evidence based medicine to identify and treat congenital heart diseases. This has been highly successful but treatments are not customized for individual patients. Because of this, patient specific medicine has emerged as an improvement where the general treatments are not suitable. At Lund University Hospital pioneering efforts are made to use Cardiac Magnetic Resonance (CMR) to scan children with a heart disease to do patient specific computational fluid dynamic modeling as a prediction of potential surgical outcomes. The current process of generating geometrical models for modeling... (More)
One percent of children have a congenital heart disease, often requiring surgical corrections. The last fifty years have seen a sharp increase in the use of evidence based medicine to identify and treat congenital heart diseases. This has been highly successful but treatments are not customized for individual patients. Because of this, patient specific medicine has emerged as an improvement where the general treatments are not suitable. At Lund University Hospital pioneering efforts are made to use Cardiac Magnetic Resonance (CMR) to scan children with a heart disease to do patient specific computational fluid dynamic modeling as a prediction of potential surgical outcomes. The current process of generating geometrical models for modeling used as input for computations is time consuming and is a bottleneck in the current work flow.

The aim of this thesis is to develop a tool that extracts the identified blood vessels from the CMR scans, finds their center lines and exports them to a CAD-program for further analysis. The tool is meant to tie into a work flow that uses the CMR scans to predict the blood flow outcomes of potential surgeries using fluid dynamic simulation. The tool has to be modular, extensible and usable to the Cardiac MR group at Lund University.

This thesis resulted in a plugin to Segment, a software package for medical image analysis freely available for research purposes developed at the Cardiac MR group which extracts the vessel outlines from the MR images and identifies the center lines of the vessels, modifies the outlines for easier 3D generation and exports the data to Creo, a commercially available CAD-program. The plugin is written in Matlab and is modular and extendible leading to easy modifications in the future as new requirements are added and old ones are updated. (Less)
Popular Abstract (Swedish)
En procent av alla barn i Sverige har ett medfött hjärtfel som ofta kräver en kirurgisk korrigering. De senaste femtio åren har vi sett en kraftig ökning i användandet av evidensbaserad medicin för att upptäcka och behandla medfödda hjärtfel. Detta har varit mycket lyckat men behandlingar är inte anpassade efter individuella behov. På grund av detta har patientspecifik medicin tillkommit som en förbäätring där generella behandlingar inte fungerar. Vid Lunds Universitetssjukhus utvecklas nya metoder för att använda magnetkamerabilder för att förutsäga vad kirurgiska ingrepp skulle resultera i genom användnig av blodflödessimuleringar. Den nuvarande metoden som används för att skapa de geometriska modellerna som används för simuleringarna är... (More)
En procent av alla barn i Sverige har ett medfött hjärtfel som ofta kräver en kirurgisk korrigering. De senaste femtio åren har vi sett en kraftig ökning i användandet av evidensbaserad medicin för att upptäcka och behandla medfödda hjärtfel. Detta har varit mycket lyckat men behandlingar är inte anpassade efter individuella behov. På grund av detta har patientspecifik medicin tillkommit som en förbäätring där generella behandlingar inte fungerar. Vid Lunds Universitetssjukhus utvecklas nya metoder för att använda magnetkamerabilder för att förutsäga vad kirurgiska ingrepp skulle resultera i genom användnig av blodflödessimuleringar. Den nuvarande metoden som används för att skapa de geometriska modellerna som används för simuleringarna är tidskrävande och en flaskhals i nuvarande arbetsflöde.

Målet med denna uppsats är att utveckla ett verktyg som extraherar identifierade blodkärl från magnetkamerabilder, hittar kärlens centrumlinjer och exporterar dem tiil ett CAD-program för vidare analys. Verktyget ska integreras i existerande arbetsflöde och ska vara modulär och användbar för Hjärtmagnetkameragruppen vid Lunds Universitet.

Uppsatsen resulterade i en applikation till Segment, en för forskare öppen mjukvara för medicinsk bildanalys utvecklad i Hjärtmagnetkameragruppen. Applikationen extraherar blodkärlsutlinjeringarna och identifierar kärlens centrumlinjer, modifierar dem efter användarens önskemål och exporterar dem till Creo, ett kommersiellt CAD-program. (Less)
Please use this url to cite or link to this publication:
author
Söderlund, Andreas LU
supervisor
organization
course
NUMM11 20161
year
type
H2 - Master's Degree (Two Years)
subject
keywords
MR, CMR, Cardiac, Magnetic, Resonance, Patient, Specific, Modeling, Matlab, Segment, Creo, B-splines, splines, smoothing, filters, interpolation, 3D
publication/series
Master's Theses in Mathematical Sciences
report number
LUNFNA-3022-2016
ISSN
1404-6342
other publication id
2016:E18
language
English
id
8876820
date added to LUP
2016-09-12 13:45:19
date last changed
2016-09-12 13:45:19
@misc{8876820,
  abstract     = {One percent of children have a congenital heart disease, often requiring surgical corrections. The last fifty years have seen a sharp increase in the use of evidence based medicine to identify and treat congenital heart diseases. This has been highly successful but treatments are not customized for individual patients. Because of this, patient specific medicine has emerged as an improvement where the general treatments are not suitable. At Lund University Hospital pioneering efforts are made to use Cardiac Magnetic Resonance (CMR) to scan children with a heart disease to do patient specific computational fluid dynamic modeling as a prediction of potential surgical outcomes. The current process of generating geometrical models for modeling used as input for computations is time consuming and is a bottleneck in the current work flow.

The aim of this thesis is to develop a tool that extracts the identified blood vessels from the CMR scans, finds their center lines and exports them to a CAD-program for further analysis. The tool is meant to tie into a work flow that uses the CMR scans to predict the blood flow outcomes of potential surgeries using fluid dynamic simulation. The tool has to be modular, extensible and usable to the Cardiac MR group at Lund University.

This thesis resulted in a plugin to Segment, a software package for medical image analysis freely available for research purposes developed at the Cardiac MR group which extracts the vessel outlines from the MR images and identifies the center lines of the vessels, modifies the outlines for easier 3D generation and exports the data to Creo, a commercially available CAD-program. The plugin is written in Matlab and is modular and extendible leading to easy modifications in the future as new requirements are added and old ones are updated.},
  author       = {Söderlund, Andreas},
  issn         = {1404-6342},
  keyword      = {MR,CMR,Cardiac,Magnetic,Resonance,Patient,Specific,Modeling,Matlab,Segment,Creo,B-splines,splines,smoothing,filters,interpolation,3D},
  language     = {eng},
  note         = {Student Paper},
  series       = {Master's Theses in Mathematical Sciences},
  title        = {Patient-specific blood vessel modeling on data from MR scans of children with congenital heart disease},
  year         = {2016},
}