Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI
<p>Arterial spin labelling (ASL) is a non-invasive magnetic resonance imaging (MRI) technique that utilises arterial blood water as an endogenous tracer to measure cerebral blood flow (CBF). While the ASL white paper recommends the clinical use of a single post-labelling delay (PLD) pseudo-con...
| מחבר ראשי: | |
|---|---|
| מחברים אחרים: | |
| פורמט: | Thesis |
| שפה: | English |
| יצא לאור: |
2022
|
| נושאים: |
| _version_ | 1826309813068365824 |
|---|---|
| author | Zhang, LX |
| author2 | Michael, C |
| author_facet | Michael, C Zhang, LX |
| author_sort | Zhang, LX |
| collection | OXFORD |
| description | <p>Arterial spin labelling (ASL) is a non-invasive magnetic resonance imaging (MRI) technique that utilises arterial blood water as an endogenous tracer to measure cerebral blood flow (CBF). While the ASL white paper recommends the clinical use of a single post-labelling delay (PLD) pseudo-continuous ASL (pCASL) method to robustly quantify CBF, there has been a growing body of research on multi-delay ASL which provides the ability to estimate for parameters other than CBF.</p>
<p>A framework applying optimal sampling strategy (OSS) to multi-PLD pCASL sequences has been developed to obtain optimised protocols by minimising the estimation uncertainty on both CBF and arterial transit time (ATT) or CBF only. Although this framework is flexible for designing ASL protocols with a variable number of target parameters, it adopted an idealised kinetic model and used a narrow range of prior values in the optimisation, rendering the optimised protocols susceptible to various sources of error during estimation.</p>
<p>This thesis seeks to systematically investigate the estimation sensitivity of existing optimised protocols to a number of physiological effects that are not accounted for in the idealised model. Additionally, it aims to extend the current framework and develop novel protocols optimised for both delay and echo times to improve the estimation of blood-brain barrier (BBB) water exchange.</p>
<p>Firstly, four pCASL protocols were examined for CBF and ATT estimation performance with the effects of flow dispersion, macrovascular signal contamination (MVC), and prolonged ATT. It was found that the protocol optimised for CBF estimation had the lowest CBF sensitivity to MVC, and extended kinetic models were beneficial to control for the presence of these effects. Results from simulation were validated using a previously acquired in vivo dataset.</p>
<p>Next, parameter sensitivity to MVC were evaluated for a comprehensive set of protocols optimised for CBF estimation. A wide range of tissue and macrovascular transit time ground truths was considered, and different assumptions regarding the relationship between the two transit times were compared and discussed. The resulting sensitivity maps further illustrated the benefit of adding an extra compartment in the analysis to account for MVC and could be used for data interpretation.</p>
<p>Finally, the OSS framework was extended to enable the combined optimisation of PLD and TE in multi-TE pCASL to measure water exchange time, Texch, across the BBB. Protocols were optimised with different optimality criteria under two multi-TE schemes and evaluated in terms of their theoretical estimation root-mean-square errors (RMSEs) using simulation. While the optimised protocols resulted in lower RMSEs of corresponding parameter estimation, reliably detecting a pathological level of Texch change in an individual using the optimised protocols within five minutes remains challenging.</p> |
| first_indexed | 2024-03-07T07:41:19Z |
| format | Thesis |
| id | oxford-uuid:aa3bd6ec-46e5-48de-81f9-14cb762b8448 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:41:19Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | oxford-uuid:aa3bd6ec-46e5-48de-81f9-14cb762b84482023-04-19T11:41:25ZImproving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRIThesishttp://purl.org/coar/resource_type/c_db06uuid:aa3bd6ec-46e5-48de-81f9-14cb762b8448Biomedical engineeringPerfusion (Physiology)Magnetic resonance imagingEnglishHyrax Deposit2022Zhang, LXMichael, CBulte, D<p>Arterial spin labelling (ASL) is a non-invasive magnetic resonance imaging (MRI) technique that utilises arterial blood water as an endogenous tracer to measure cerebral blood flow (CBF). While the ASL white paper recommends the clinical use of a single post-labelling delay (PLD) pseudo-continuous ASL (pCASL) method to robustly quantify CBF, there has been a growing body of research on multi-delay ASL which provides the ability to estimate for parameters other than CBF.</p> <p>A framework applying optimal sampling strategy (OSS) to multi-PLD pCASL sequences has been developed to obtain optimised protocols by minimising the estimation uncertainty on both CBF and arterial transit time (ATT) or CBF only. Although this framework is flexible for designing ASL protocols with a variable number of target parameters, it adopted an idealised kinetic model and used a narrow range of prior values in the optimisation, rendering the optimised protocols susceptible to various sources of error during estimation.</p> <p>This thesis seeks to systematically investigate the estimation sensitivity of existing optimised protocols to a number of physiological effects that are not accounted for in the idealised model. Additionally, it aims to extend the current framework and develop novel protocols optimised for both delay and echo times to improve the estimation of blood-brain barrier (BBB) water exchange.</p> <p>Firstly, four pCASL protocols were examined for CBF and ATT estimation performance with the effects of flow dispersion, macrovascular signal contamination (MVC), and prolonged ATT. It was found that the protocol optimised for CBF estimation had the lowest CBF sensitivity to MVC, and extended kinetic models were beneficial to control for the presence of these effects. Results from simulation were validated using a previously acquired in vivo dataset.</p> <p>Next, parameter sensitivity to MVC were evaluated for a comprehensive set of protocols optimised for CBF estimation. A wide range of tissue and macrovascular transit time ground truths was considered, and different assumptions regarding the relationship between the two transit times were compared and discussed. The resulting sensitivity maps further illustrated the benefit of adding an extra compartment in the analysis to account for MVC and could be used for data interpretation.</p> <p>Finally, the OSS framework was extended to enable the combined optimisation of PLD and TE in multi-TE pCASL to measure water exchange time, Texch, across the BBB. Protocols were optimised with different optimality criteria under two multi-TE schemes and evaluated in terms of their theoretical estimation root-mean-square errors (RMSEs) using simulation. While the optimised protocols resulted in lower RMSEs of corresponding parameter estimation, reliably detecting a pathological level of Texch change in an individual using the optimised protocols within five minutes remains challenging.</p> |
| spellingShingle | Biomedical engineering Perfusion (Physiology) Magnetic resonance imaging Zhang, LX Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title | Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title_full | Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title_fullStr | Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title_full_unstemmed | Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title_short | Improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling MRI |
| title_sort | improving the quantification of cerebral haemodynamics using optimally sampled arterial spin labelling mri |
| topic | Biomedical engineering Perfusion (Physiology) Magnetic resonance imaging |
| work_keys_str_mv | AT zhanglx improvingthequantificationofcerebralhaemodynamicsusingoptimallysampledarterialspinlabellingmri |