Meeting Banner
Abstract #3703

Segmented Multi-Echo MPRAGE Acquisition for Accelerated T1-Weighted Brain Imaging

Pavel Falkovskiy1, 2, Tobias Kober, 12, Denise Reyes3, Kaely Steinert3, Matthias Seeger4, Matthew Bernstein3, Gunnar Krueger, 12

1Advanced Clinical Imaging Technology, Siemens Healthcare IM S AW, Lausanne, Switzerland; 2CIBM-AIT, cole Polytechnique Fdrale de Lausanne, Lausanne, Switzerland; 3Department of Radiology, Mayo Clinic, Rochester, MN, United States; 4Laboratory for Probabilistic Machine Learning, cole Polytechnique Fdrale de Lausanne, Lausanne, Switzerland

Patient comfort and short scan times are crucial in clinical settings. The MP-RAGE pulse sequence produces an excellent T1-weighted contrast but has relatively long acquisition times. In contrast to conventional parallel imaging, this work accelerates the acquisition by using an EPI-type multi-echo k-space segmentation scheme. Obtained images were rated to appear with clinically useful image quality at up to 4x scan time reduction (9:14 to 2:20 min). The concept is independent of the coil design and could be applied in situations when parallel imaging technology is limited, e.g. when there are restrictions in imaging dimensions and coil geometry.

Abstract PDF Presentation Video

Keywords

abundant accelerated accelerates acceleration acceptable achieved acquisition acquisitions adapted advanced affected allowed applied arising array arrays audience avoid bandwidth bars best blinded blurring brain caution channel clinic clinical clinically coil comfort comprised computed confirm dashed degradation described desirable deviation deviations diagnostic echoes error especially estimation even example excitation exhibited expected experienced exploiting explores feasibility fields fold foundations fully gains good graded grades gradient head healthy illustrated illustrates improve inconsistencies increasingly indicates inserted laboratory largely learning limits magnetization magnitude male materials mayo measures methodology mild mildly modified motion nevertheless noise observable observer opposed originating parallel participating patient pixel portions probabilistic prolonged protocol protocols pulse qualitative quality quantitative radiology rage rank readout readouts reduce reduced rely repeat requirements resolution respect samples scale scheme segmented segments session settings shortened space spatially square still structural subjects subsequently substantially summary supported system table taking technology train trains trajectory traversed trial turbo type useful volumes volumetric worst