In patients with aortic stenosis, accurate assessment of severity with echocardiography is central to surgical decision making. But, when image quality is poor or equivocal results obtained, another robust non-invasive technique would be invaluable. Cardiac magnetic resonance (CMR) may be a useful alternative. Phase contrast CMR can measure ow and velocity, therefore it is theoretically possible to estimate the main determinant of severity aortic valve area, using the continuity approach. However, it was found that the phase contrast estimate of stroke volume, sampled in the stenotic jet, systematically underestimated left ventricular stroke volume. This underestimation was greater with increasing aortic stenosis severity. Critical clinical treatment decisions depend on the ability to reliably differentiate between patients with moderate and severe aortic stenosis. To achieve accurate estimation of aortic valve areas the velocity and ow data obtained in these turbulent, high velocity jets must be accurate. In this thesis, non-stenotic and stenotic phantoms were designed and constructed to experimentally interrogate the error. It was determined that signal loss, due to intravoxel dephasing, decreased the reliability of the measured forward ow jet velocities. Extreme signal loss in the jet eventuated in salt and pepper noise, which, with a mean velocity of zero, resulted in the underestimation. Intravoxel dephasing signal loss due to higher order motions, turbulence and spin mixing could all be mitigated by reducing the duration of the velocity sensitivity gradients and shortening the overall echo time (TE). However, improvements in an optimised PC sequence (TE 1:5ms) were not satisfactory. Flow estimates remained variable and were underestimated beyond the aortic valve. To reduce the TE further, a new phase contrast pulse sequence based on an ultrashort TE readout trajectory and velocity dependent slice excitation with gradient inversion was designed and implemented. The new sequence's TE is approximately 25% (0:65ms) of what is currently clinically available (TE 2:8ms). Good agreement in the phantom was maintained up to very high ow rates with improved signal characteristics shown in-vivo. This new phase contrast pulse sequence is worthy of further investigation as an accurate evaluation of patients with aortic stenosis. / This work in this thesis was conducted at The Auckland Bioengineering Institute, The Centre for Advanced MRI and The Oxford Centre for Clinical Magnetic Resonance in collaboration with Siemens Health care.

In patients with aortic stenosis, accurate assessment of severity with echocardiography is central to surgical decision making. But, when image quality is poor or equivocal results obtained, another robust non-invasive technique would be invaluable. Cardiac magnetic resonance (CMR) may be a useful alternative. Phase contrast CMR can measure ow and velocity, therefore it is theoretically possible to estimate the main determinant of severity aortic valve area, using the continuity approach. However, it was found that the phase contrast estimate of stroke volume, sampled in the stenotic jet, systematically underestimated left ventricular stroke volume. This underestimation was greater with increasing aortic stenosis severity. Critical clinical treatment decisions depend on the ability to reliably differentiate between patients with moderate and severe aortic stenosis. To achieve accurate estimation of aortic valve areas the velocity and ow data obtained in these turbulent, high velocity jets must be accurate. In this thesis, non-stenotic and stenotic phantoms were designed and constructed to experimentally interrogate the error. It was determined that signal loss, due to intravoxel dephasing, decreased the reliability of the measured forward ow jet velocities. Extreme signal loss in the jet eventuated in salt and pepper noise, which, with a mean velocity of zero, resulted in the underestimation. Intravoxel dephasing signal loss due to higher order motions, turbulence and spin mixing could all be mitigated by reducing the duration of the velocity sensitivity gradients and shortening the overall echo time (TE). However, improvements in an optimised PC sequence (TE 1:5ms) were not satisfactory. Flow estimates remained variable and were underestimated beyond the aortic valve. To reduce the TE further, a new phase contrast pulse sequence based on an ultrashort TE readout trajectory and velocity dependent slice excitation with gradient inversion was designed and implemented. The new sequence's TE is approximately 25% (0:65ms) of what is currently clinically available (TE 2:8ms). Good agreement in the phantom was maintained up to very high ow rates with improved signal characteristics shown in-vivo. This new phase contrast pulse sequence is worthy of further investigation as an accurate evaluation of patients with aortic stenosis. / This work in this thesis was conducted at The Auckland Bioengineering Institute, The Centre for Advanced MRI and The Oxford Centre for Clinical Magnetic Resonance in collaboration with Siemens Health care.

In patients with aortic stenosis, accurate assessment of severity with echocardiography is central to surgical decision making. But, when image quality is poor or equivocal results obtained, another robust non-invasive technique would be invaluable. Cardiac magnetic resonance (CMR) may be a useful alternative. Phase contrast CMR can measure ow and velocity, therefore it is theoretically possible to estimate the main determinant of severity aortic valve area, using the continuity approach. However, it was found that the phase contrast estimate of stroke volume, sampled in the stenotic jet, systematically underestimated left ventricular stroke volume. This underestimation was greater with increasing aortic stenosis severity. Critical clinical treatment decisions depend on the ability to reliably differentiate between patients with moderate and severe aortic stenosis. To achieve accurate estimation of aortic valve areas the velocity and ow data obtained in these turbulent, high velocity jets must be accurate. In this thesis, non-stenotic and stenotic phantoms were designed and constructed to experimentally interrogate the error. It was determined that signal loss, due to intravoxel dephasing, decreased the reliability of the measured forward ow jet velocities. Extreme signal loss in the jet eventuated in salt and pepper noise, which, with a mean velocity of zero, resulted in the underestimation. Intravoxel dephasing signal loss due to higher order motions, turbulence and spin mixing could all be mitigated by reducing the duration of the velocity sensitivity gradients and shortening the overall echo time (TE). However, improvements in an optimised PC sequence (TE 1:5ms) were not satisfactory. Flow estimates remained variable and were underestimated beyond the aortic valve. To reduce the TE further, a new phase contrast pulse sequence based on an ultrashort TE readout trajectory and velocity dependent slice excitation with gradient inversion was designed and implemented. The new sequence's TE is approximately 25% (0:65ms) of what is currently clinically available (TE 2:8ms). Good agreement in the phantom was maintained up to very high ow rates with improved signal characteristics shown in-vivo. This new phase contrast pulse sequence is worthy of further investigation as an accurate evaluation of patients with aortic stenosis. / This work in this thesis was conducted at The Auckland Bioengineering Institute, The Centre for Advanced MRI and The Oxford Centre for Clinical Magnetic Resonance in collaboration with Siemens Health care.

In patients with aortic stenosis, accurate assessment of severity with echocardiography is central to surgical decision making. But, when image quality is poor or equivocal results obtained, another robust non-invasive technique would be invaluable. Cardiac magnetic resonance (CMR) may be a useful alternative. Phase contrast CMR can measure ow and velocity, therefore it is theoretically possible to estimate the main determinant of severity aortic valve area, using the continuity approach. However, it was found that the phase contrast estimate of stroke volume, sampled in the stenotic jet, systematically underestimated left ventricular stroke volume. This underestimation was greater with increasing aortic stenosis severity. Critical clinical treatment decisions depend on the ability to reliably differentiate between patients with moderate and severe aortic stenosis. To achieve accurate estimation of aortic valve areas the velocity and ow data obtained in these turbulent, high velocity jets must be accurate. In this thesis, non-stenotic and stenotic phantoms were designed and constructed to experimentally interrogate the error. It was determined that signal loss, due to intravoxel dephasing, decreased the reliability of the measured forward ow jet velocities. Extreme signal loss in the jet eventuated in salt and pepper noise, which, with a mean velocity of zero, resulted in the underestimation. Intravoxel dephasing signal loss due to higher order motions, turbulence and spin mixing could all be mitigated by reducing the duration of the velocity sensitivity gradients and shortening the overall echo time (TE). However, improvements in an optimised PC sequence (TE 1:5ms) were not satisfactory. Flow estimates remained variable and were underestimated beyond the aortic valve. To reduce the TE further, a new phase contrast pulse sequence based on an ultrashort TE readout trajectory and velocity dependent slice excitation with gradient inversion was designed and implemented. The new sequence's TE is approximately 25% (0:65ms) of what is currently clinically available (TE 2:8ms). Good agreement in the phantom was maintained up to very high ow rates with improved signal characteristics shown in-vivo. This new phase contrast pulse sequence is worthy of further investigation as an accurate evaluation of patients with aortic stenosis. / This work in this thesis was conducted at The Auckland Bioengineering Institute, The Centre for Advanced MRI and The Oxford Centre for Clinical Magnetic Resonance in collaboration with Siemens Health care.