Background

Cardiac MRI may greatly benefit from ultra high field (UHF) providing higher SNR and intrinsic contrast. But shorter RF wavelength yields transmit B1 (B1+) heteroge-neity and contrast variations through the heart. These can be addressed by parallel transmission (pTX) using multi-spoke RF pulses as previously shown in other organs. However, applying pTX in cardiac MRI at 7T is challen-ging and requires rapid, multi-channel ECG triggered B1+ calibration and ECG synchronized, motion-insensitive pTX acquisitions. In this initial work we investigate the impact of 2-spoke RF pulse design on cardiac imaging at 7T using a 16-channel pTX system Methods Healthy volunteers were imaged at 7T (Siemens, Erlangen) using a 16 channel pTX system with 16 channel transcei-ver body coil. All sequences were modified to enable both ECG triggering and full pTX capability. B1+ maps of each transmit (TX) channel were obtained using a fast estima-tion technique [1]. Each map was acquired within 484ms during diastole. 1- and 2-spoke pTX RF pulses (SINC shape, BWTP 4) were designed based on magnitude least squares (MLS) optimization [2] aiming at homogeneous excitation in manually drawn ROI. 3 excitations were compared: a) 1-spoke ( = standard excitation) with non-optimized initial RF phase setting (equal amplitude), b) 1-spoke with optimized RF magnitude/phase and c) 2-spoke RF pulse with magnitude/phase optimized for each spoke and channel. To keep a constant RF duration 1-spoke pulses were set to 1600us duration while 2-spoke subpulses were set to 800us. Coefficient of variation (CV, i.e. std/mean) in target ROI, total RF energy and peak energy per channel were computed for all RF pulses for same flip angle. Axial- and 4-chamber views were acquiried using a CINE GRE sequence (resolutio