Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Unified theory and experimental approach for measuring restricted diffusion and water exchange

Chakwizira, Arthur LU (2020) MSFT01 20201
Medical Physics Programme
Abstract
Fifteen million people lose their lives to stroke and cancer every year. This amounts to one in every four deaths worldwide, making these among the deadliest diseases affecting humankind. Fruitful attempts at containing stroke- and cancer-related mortality hinge on early detection. Diffusion MRI (dMRI) is an imaging modality that is sensitive to tissue microstructure, and therefore unveils tissue changes before they become visible on morpho- logical images. Maps of the apparent diffusion coefficient (ADC) find routine clinical use for cancer and stroke diagnosis. More detailed microstructural information is obtained by time-dependent dMRI, which probes the diffusion of water restricted within cells and ex- changing between cellular... (More)
Fifteen million people lose their lives to stroke and cancer every year. This amounts to one in every four deaths worldwide, making these among the deadliest diseases affecting humankind. Fruitful attempts at containing stroke- and cancer-related mortality hinge on early detection. Diffusion MRI (dMRI) is an imaging modality that is sensitive to tissue microstructure, and therefore unveils tissue changes before they become visible on morpho- logical images. Maps of the apparent diffusion coefficient (ADC) find routine clinical use for cancer and stroke diagnosis. More detailed microstructural information is obtained by time-dependent dMRI, which probes the diffusion of water restricted within cells and ex- changing between cellular environments. Measurements of water restriction and exchange yield estimates of cell size and permeability, respectively. Tracking these features may be important for reliable tumour characterisation and evaluation of treatment response. Estimating cell size and permeability is a challenge because the phenomena of restriction and exchange have opposing effects on the diffusion-weighted signal. With increasing dif- fusion time, restriction elevates the signal while exchange decreases it. Restriction and exchange can therefore not be disentangled by solely varying the diffusion time in a dMRI experiment. Typical models assume that restriction is prevalent at short time scales while exchange is dominant at long time scales. Estimates of size and exchange are made inde- pendently by doing experiments in different time regimes. Nevertheless, previous research has highlighted that this approach may induce inaccuracy in estimated diffusion metrics. In light of this, the objective of this thesis work was to develop a unified theoretical frame- work and experimental approach for measuring restriction and exchange at all time scales. Extending previous work, a commonly used model of exchange within dMRI - the Ka¨rger model - was generalised to accommodate arbitrary gradient waveforms. By incorporating the restriction information contained in the diffusion spectrum, this thesis work derived a general unified model of restriction and exchange that is valid for all time scales, gradient waveforms and b-values. In other words, the proposed model is an exact solution to the problem at hand. With the aim of deriving a complementary theory more informative for experimental design, an alternative approach employing the relation between the parti- cle velocity autocorrelation function and the diffusion spectrum was also explored. This yielded a second-order signal representation applicable to any gradient waveform, all time scales and moderate b-values. Monte-Carlo simulations were performed on a synthetic structure to validate the developed theory. Excellent agreement between simulated and estimated parameters was observed. Numerically optimised gradient waveforms improved precision in parameter estimates by a factor of 2 in relation to standard pulsed-gradient sequences. In a word, this thesis work presents a general, unified theoretical framework describing the effects of restricted diffusion and water exchange on the diffusion-weighted signal. The results pave the way for future research involving optimisation of experiments to minimise scan times, which is a crucial step towards clinical implementation. (Less)
Popular Abstract (Swedish)
Drygt femton miljoner människor avlider av cancer och stroke varje år, vilket gör dem till några av de svåraste sjukdomarna som vi drabbas av. Dödligheten hos dessa sjukdomar påverkas främst av när dem upptäcks. Ju tidigare sjukdomen diagnostiseras, desto högre blir sannolikheten för att den botas. Tidig upptäckt kräver känsliga diagnostiska metoder, särskilt de som kan detektera förändringar i vävnaden innan de visas på vanliga anatomiska bilder. Diffusions MR (dMRI) är en bildgivande metod som är känslig mot vävnadens mikroskopiska struktur och funktion. dMRI kan således detektera förändringar innan de syns på konventionella medicinska bilder. Mätningar av diffusionshastigheten hos vattenmolekyler används rutinmässigt inom kliniken för... (More)
Drygt femton miljoner människor avlider av cancer och stroke varje år, vilket gör dem till några av de svåraste sjukdomarna som vi drabbas av. Dödligheten hos dessa sjukdomar påverkas främst av när dem upptäcks. Ju tidigare sjukdomen diagnostiseras, desto högre blir sannolikheten för att den botas. Tidig upptäckt kräver känsliga diagnostiska metoder, särskilt de som kan detektera förändringar i vävnaden innan de visas på vanliga anatomiska bilder. Diffusions MR (dMRI) är en bildgivande metod som är känslig mot vävnadens mikroskopiska struktur och funktion. dMRI kan således detektera förändringar innan de syns på konventionella medicinska bilder. Mätningar av diffusionshastigheten hos vattenmolekyler används rutinmässigt inom kliniken för att diagnostisera cancer och stroke. Mer information om vävnadens mikrostruktur utöver diffusionshastigheten kan bidra till diagnos och karakterisering av tumörer, samt utvärdering av respons på behandling. Sådan information kan erhållas med hjälp av så kallad tidsberoende dMRI, vilket möjliggör mätning av cellernas storlek och genomsläpplighet. Dessa cellegenskaper är dock svåra att mäta på grund av att de har motsatta effekter på den insamlade MR signalen. En minskning av signalen kan antingen bero på att cellen har blivit större eller att genomsläppligheten har ökat, vilket gör det svårt att särskilja de två effekterna. Befintliga modeller inom dMRI antar att cellstorleken är viktig endast vid korta tidskalor, och genomsläpplighet är relevant endast vid långa tidskalor. Genom att utföra experiment på olika tidskalor kan man erhålla uppskattningar av cellstorlek och genomsläpplighet. Det har dock visats att denna metod kan införa osäkerheter i de uppmätta egenskaperna. Syftet med detta examenarbete var därmed att utveckla en teori och experimentell metod för att kunna mäta cellstorlek och genomsläpplighet utan att göra de ovan beskrivna antaganden. Detta gjordes genom att generalisera och kombinera befintliga teorier som beskriver effekterna av cellstorlek och genomsläpplighet på MR-signalen. Den generaliserade teorin gör det möjligt att designa, optimera och utföra experiment som ger pålitliga mätningar av cellstorlek och genomsläpplighet. Datorsimuleringar utfördes för att validera den utvecklade teorin. Teorin visade god överensstämmelse med simuleringarna. Detta arbete utgör ett viktigt tillägg i forskningsområdet dMRI. Resultaten kan användas som utgångspunkt för framtida forskning inom optimering av experiment för att minimera undersökningstider. På så sätt kan metoden eventuellt implementeras i kliniken. (Less)
Please use this url to cite or link to this publication:
author
Chakwizira, Arthur LU
supervisor
organization
course
MSFT01 20201
year
type
H2 - Master's Degree (Two Years)
subject
language
Swedish
id
9036091
date added to LUP
2021-01-17 12:57:29
date last changed
2021-01-17 12:57:29
@misc{9036091,
  abstract     = {{Fifteen million people lose their lives to stroke and cancer every year. This amounts to one in every four deaths worldwide, making these among the deadliest diseases affecting humankind. Fruitful attempts at containing stroke- and cancer-related mortality hinge on early detection. Diffusion MRI (dMRI) is an imaging modality that is sensitive to tissue microstructure, and therefore unveils tissue changes before they become visible on morpho- logical images. Maps of the apparent diffusion coefficient (ADC) find routine clinical use for cancer and stroke diagnosis. More detailed microstructural information is obtained by time-dependent dMRI, which probes the diffusion of water restricted within cells and ex- changing between cellular environments. Measurements of water restriction and exchange yield estimates of cell size and permeability, respectively. Tracking these features may be important for reliable tumour characterisation and evaluation of treatment response. Estimating cell size and permeability is a challenge because the phenomena of restriction and exchange have opposing effects on the diffusion-weighted signal. With increasing dif- fusion time, restriction elevates the signal while exchange decreases it. Restriction and exchange can therefore not be disentangled by solely varying the diffusion time in a dMRI experiment. Typical models assume that restriction is prevalent at short time scales while exchange is dominant at long time scales. Estimates of size and exchange are made inde- pendently by doing experiments in different time regimes. Nevertheless, previous research has highlighted that this approach may induce inaccuracy in estimated diffusion metrics. In light of this, the objective of this thesis work was to develop a unified theoretical frame- work and experimental approach for measuring restriction and exchange at all time scales. Extending previous work, a commonly used model of exchange within dMRI - the Ka¨rger model - was generalised to accommodate arbitrary gradient waveforms. By incorporating the restriction information contained in the diffusion spectrum, this thesis work derived a general unified model of restriction and exchange that is valid for all time scales, gradient waveforms and b-values. In other words, the proposed model is an exact solution to the problem at hand. With the aim of deriving a complementary theory more informative for experimental design, an alternative approach employing the relation between the parti- cle velocity autocorrelation function and the diffusion spectrum was also explored. This yielded a second-order signal representation applicable to any gradient waveform, all time scales and moderate b-values. Monte-Carlo simulations were performed on a synthetic structure to validate the developed theory. Excellent agreement between simulated and estimated parameters was observed. Numerically optimised gradient waveforms improved precision in parameter estimates by a factor of 2 in relation to standard pulsed-gradient sequences. In a word, this thesis work presents a general, unified theoretical framework describing the effects of restricted diffusion and water exchange on the diffusion-weighted signal. The results pave the way for future research involving optimisation of experiments to minimise scan times, which is a crucial step towards clinical implementation.}},
  author       = {{Chakwizira, Arthur}},
  language     = {{swe}},
  note         = {{Student Paper}},
  title        = {{Unified theory and experimental approach for measuring restricted diffusion and water exchange}},
  year         = {{2020}},
}