The Cardiac Magnetic Resonance (CMR) Group develops discovery-oriented and translational cardiac and cardiovascular MRI techniques to deepen our understanding of cardiac performance and improve clinical care. 

The CMR Group is very interested in further developing MRI acquisition, reconstruction, and analysis methods that improve the speed, accuracy, and precision with which we can measure cardiac and cardiovascular structure, function, flow and remodeling.

The CMR Group has been interested in the challenge and opportunities of encoding the self-diffusion of water in the beating heart as a means to estimate microstructural organization of the heart in vivo. Our team develops novel MRI pulse sequences for cardiac diffusion tensor imaging (cDTI) and we also refine the reconstruction and analysis methods that are necessary for studying the heart’s microstructure.

The CMR Group also develops quantitative methods to measure cardiac performance. We principally do this by developing acquisition, reconstruction, and analysis methods for MRI techniques that encoding for cardiac displacement information (e.g. MRI tagging and cine DENSE). Our method enables estimating cardiac strain and can be coupled with cDTI methods to measure “myofiber” strain in vivo.

The CMR group also develops quantitative methods to measure cardiovascular flow using phase-contrast MRI. We are interested in ultra-fast approaches to acquiring and reconstructing 2D time resolved phase contrast MRI (2D-PC) and rely on machine-learning based approaches to enable high rates of acquisition acceleration and fast reconstruction (inference) times. Additional efforts are focused on flow measurement accuracy and precision for which we combine our research in gradient impulse response function (GIRF) with our GrOpt approach to image as fast-as-possible while mitigating encoding errors (e.g. background phase). Our efforts are also focused on fast and accurate methods for 4D-Flow encoding.

For several years our lab has worked on developing the gradient optimization (GrOpt) toolbox. GrOpt is an open-source technology that enables gradient waveform design subject to constraints. With GrOpt we can design MRI pulse sequences to maximize encoding efficiency, minimize encoding errors, and enable encoding of novel forms of MRI-based information.

The CMR Group also works carefully to rigorously validate our MRI methods. To facilitate this we work with a range of equipment that furthers this goal. For example, we use 3D-printing to create rigid and compliant structures that represent digitally defined test objects or patient-specific anatomic structures (e.g. pathologic aortas). These phantoms can be instrumented to measure pressures, while connected to computer controlled pumps that provide cardiovascular-like flows. We build these phantoms to be MRI compatible thereby permitting rigorous testing on the bench and in the MRI scanner.

The CMR group is part of the Radiological Sciences Lab in the Department of Radiology at Stanford University and also the Division of Radiology at the Veterans Administration Palo Alto Health Care System.