Cardiac Magnetic Resonance (CMR) is a promising non-invasive imaging technique capable of assessing cardiovascular function and anatomy for both adult and pediatric patients without the need for ionization energy. Even with the stated benefits, CMR utilization in the clinic is low (~1-2%) due in large part to long acquisition time necessary for magnetic resonance imaging (MRI), the need for gadolinium-based contrast agent (GBCA) administration, and the need for robust respiratory and cardiac motion suppression. The purpose of this dissertation was to describe the development and validation of an improved MR protocol designed to increase clinical CMR utilization by addressing current limitations. This dissertation was primarily focused on improving functional and anatomical CMR by developing rapid real-time free-breathing cine acquisition for assessing heart function, rapid real-time free-breathing phase contrast acquisition for assessing valvular hemodynamics, and free-breathing non-contrast magnetic resonance angiography (NC-MRA) for assessing thoracic angiography. 12-fold accelerated real-time cine imaging was produced using tiny golden angle radial k-space sampling combined with golden-angle radial sparse parallel MRI (GRASP) compressed sensing (CS) reconstruction. 38.4-fold accelerated real-time phase contrast imaging was produced using golden angle radial sampling combined with Kaiser Bessel Gridded GRASP (KGB-GRASP) reconstruction. 5.5-fold accelerated NC-MRA was produced using stack-of-star sampling, self-navigation, and extra motion-state GRASP (XD-GRASP) reconstruction. NC-MRA imaging was further improved using convolutional neural network (CNN) to rapidly reconstruct (~3 min) XD-GRASP images. Upon development, real-time cine, real-time phase contrast, and non-contrast MRA sequences were validated in cardiovascular disease patients.

Creator

• Haji-valizadeh, Hassan

DOI

• https://doi.org/10.21985/n2-yb3e-7m55

Subject

• Medical imaging

Language

• en

Alternate Identifier

• etdadmin_upload_682286
• http://dissertations.umi.com/northwestern:14782

Keyword

• Cardiovascular MRI
• compressed sensing
• real-time
• adult and pediatric
• deep learning
• free-breathing

Date created

• 2019-01-01

Resource type

• Dissertation

Rights statement

• In Copyright