Meeting Banner
Abstract #0288

MR Fingerprinting (MRF): A Novel Quantitative Approach to MRI

SUMMA25Dan Ma1, Vikas Gulani1, 2, Nicole Seiberlich1, Jeffrey Duerk1, Mark Griswold1, 2

1Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States; 2Dept. of Radiology, University Hospitals of Cleveland and Case Western Reserve University, Cleveland, OH, United States

MR Fingerprinting is a completely novel approach to quantitative MRI. Instead of collecting a series of weighted images and fitting signal time courses to an exponential model, MRF eschews collection of traditional images and a model based approach altogether. A compressed sensing based approach is adopted where information about each voxel is collected with signal timecourses generated using variable TRs, flip angles, and inversion pulses to encode relaxation information into unique signal evolution curves. Matching these curves to a comprehensive dictionary using pattern matching provides multiple parameter mappings (T1, T2, proton density and off-resonance) and synthesized reconstructed images simultaneously.

Keywords

able achieve addition address agreement analogous around assess attempts authors avoided basis best building call chosen clinically coefficient completed compressed concept constant constructed continuous course courses curves cylinders database degree degrees density depend design designed designs determine development dictionary diffusion directed echoes elements equilibrium errors every evolution evolutions example expected experimental explored extension fast fingerprinting fingerprints fitting funding generate generated generating goal gold good grants host identity increment individual infinite instead known like limited long maps matched matching measured medical motion much never next optimized orthogonal oscillating pattern personal phantom phone pixel presence probabilistic propose proton providing pulse pure pursuit quantification quantify quantitative quickly random randomly ranging rather reached recent recognition reconstructed reconstruction represent required retrieving scanning select sensing separately series sets sigma simulate simulation simulations simultaneously since solutions spin square squares theory though tissue tissues towards traditional trans type underlying unique variations varied varies various varying weightings wildly yield yielding