Abstract #3250

Impact of Compressed Sensing on Volumetric Knee MRI

Shreyas S. Vasanawala¹, Peng Lai², Marcus T. Alley¹, Garry E. Gold¹, John M. Pauly³, Michael Lustig⁴

¹Radiology, Stanford University, Stanford, CA, United States; ²ASL West, GE Healthcare, Menlo Park, CA, United States; ³Electrical Engineering, Stanford University, Stanford, CA, United States; ⁴Electrical Engineering & CS, UC Berkeley, Berkeley, CA, United States

We investigate whether compressed sensing (CS) volumetric FSE joint MRI enables thinner slices with improved quality of reformatted images. A volumetric FSE sequence was modified for compatibility with both parallel imaging (ARC) L1-SPIRiT compressed sensing (CS) reconstructions. 24 consecutive routine knee MRI patients additionally underwent volumetric imaging: 12 each at low and high acceleration. For each subject, 12 anatomic structures were evaluated on both volumetric ARC and CS images and compared with conventional 2D FSE. Thinner slices afforded by the higher acceleration of CS improves volumetric knee MRI, particularly for delineation of structures primarily evaluated on reformatted images.