084 High field (7 Tesla) cardiovascular MRI: a feasibility and comparison study

Introduction High field strength cardiovascular MRI is attractive as a clinical research tool because of its higher signal to noise ratio and its potential for increased image resolution. Technical challenges include B0 and B1 field in-homogeneity, high field ECG artefacts and specific absorption rate (SAR) limits1. This study was designed to determine the practicality and reproducibility of anatomical and functional cardiac imaging at 7 T compared to 3 T and the gold-standard 1.5T. Methods Subjects (n=3) were scanned within 1 month in a MAGNETOM 7T, MAGNETOM Tim Trio (3T) and MAGNETOM Avanto (1.5 T) system all running VB15A and with equivalent gradient systems (Siemens Healthcare, Erlangen, Germany). At 7 T, a 16 channel transceive stripline array, independently powered by 16 RF amplifiers (16×1 kW), was tuned and matched to each subject. Electrocardiographic gating used the standard 3-lead Bluetooth based Vector Cardiogram system. At 1.5 T and 3 T standard clinical 12 channel receive arrays were used with the volume transmit coil. At 7 T, complex transmit B1 field distributions were optimised with B1 shimming for a tradeoff between transmit efficiency and homogeneity based on a rapid single breath-hold calibration scan. B0 Shimming was performed with the shim volume localized to the heart. To further optimise image quality, especially at 3 and 7 T a ‘frequency scout’ scan was performed to enable use of a B0 offset to avoid off-resonance artefacts. On each scanner a short axis stack of images was acquired using both FLASH and SSFP. Scans were performed by the same experienced operator and independently analysed using Argus post-processing software to measure cardiac volumes, mass and ejection fraction. Results Reliable ECG triggering was obtained in all subjects, however careful skin preparation and more adjustment with respect to lead placement was required at 7 T than at lower field strengths due to a more pronounced magnetohydrodynamic effect. Images acquired using SSFP at 7 T had consistent B0 related banding artefacts over the anterior and inferolateral walls despite local B0 shimming and the selection of optimal offset frequencies to shift the artefacts outside of the region of interest. FLASH images were quantitatively superior at higher field strengths (SNR: 1.5 T:99, 3 T:127, 7 T:355; CNR:1.5 T:52, 3 T:58, 7 T:205). Measurement of cardiac volumes, ejection and mass showed minimal variation between the 1.5 T, 3 T and 7 T field strengths for both FLASH and SSFP sequences (all ranges less than the clinical standard of 10%) (abstract 084 figure 1). Conclusions Using optimised sequences to minimise high field artefacts, assessment of cardiac volumes and function is practical and reproducible at 7 T. Measurement of cardiac volumes, ejection fraction and mass at 7 T with FLASH or SSFP are consistent with measurements made at 3 T and the clinical gold standard of 1.5 T.