Experimental assessment of clinical MRI-induced global SAR distributions in head phantoms.

OBJECTIVE: Accurate estimation of SAR is critical to safeguarding vulnerable patients who require an MRI procedure. The increased static field strength and RF duty cycle capabilities in modern MRI scanners mean that systems can easily exceed safe SAR levels for patients. Advisory protocols routinely used to establish quality assurance protocols are not required to advise on the testing of MRI SAR levels and is not routinely measured in annual medical physics quality assurance checks. This study aims to develop a head phantom and protocol that can independently verify global SAR for MRI clinical scanners. METHODS: A four-channel birdcage head coil was used for RF transmission and signal reception. Proton resonance shift thermometry was used to estimate SAR. The SAR estimates were verified by comparing results against two other independent measures, then applied to a further four scanners at field strengths of 1.5 T and 3 T. RESULTS: Scanner output SAR values ranged from 0.42 to 1.52 W/kg. Percentage SAR differences between independently estimated values and those calculated by the scanners differed by 0-2.3%. CONCLUSION: We have developed a quality assurance protocol to independently verify the SAR output of MRI scanners.