Towards Patient-Specific and Reliable SAR Management for Parallel-Transmit Technology on Ultra-High Field MRI

人脑中的神经回路可能是任何电子工程师所能想到的最复杂的大规模电路：它包含大约1000亿个神经元，通过数万亿个通路瞬间连接起来，产生思想、记忆、行动或情感。当神经回路出现故障时，可能会出现不同的脑部疾病。英国有1000万人患有神经系统疾病，相关的经济负担估计为每年1160亿英镑。随着最近在欧洲和美国对大脑测绘项目的重大投资，神经科学家需要开发具有优越信噪比（SNR）的新成像技术，例如超高场磁共振成像，以显示复杂的神经回路如何相互作用，并最终产生预防和治疗神经系统疾病的方法。然而，由于磁激发场的不均匀性增加，在7T MR图像中出现的各种图像伪影提高了MRI测量的不确定性。解决这一问题的最新技术是利用一种新颖的射频技术：平行传输（pTX），通过多个具有独特空间轮廓的射频激励线圈来实现均匀磁化。为了将该工程解决方案应用于医学成像应用，迫切需要解决一个技术挑战：开发可靠的射频安全管理，以特定吸收率（SAR）测量，以确保该技术符合射频暴露法律限制。自pTX用于7T MRI以来，难以准确表征局部SAR，特别是针对特定受试者优化的pTX脉冲，极大地阻碍了其临床应用。目前世界范围内采用电磁（EM）模拟进行SAR估计的做法需要昂贵的安全裕度（至少高出40%）来适应不确定性。另一种SAR计算方法最近由我们的项目合作伙伴飞利浦医疗（Philips Healthcare）首创，基于b1图的超快速SAR计算，由于一些不切实际的默认假设，仍然存在精度不高的问题，并且尚未在超高场MRI的定量实验测量中得到验证。在这个项目中，我们将开发一种值得信赖的超快速SAR计算方法，通过首次将电磁数值解决方案与B1映射相结合，来解决这一工程挑战，从而实现从使用通用模型的数值模拟到主动管理特定研究对象的SAR的范式转变。特别是，我们将通过使用互补的EM知识来补充缺失的信息，解决与最新的基于MRI b1图的超快速SAR计算相关的准确性差的问题。这将是采用这种混合方法的第一次尝试，这种方法可以吸取每种技术的精华，以解决这个全球研究前沿领域的MRI SAR问题。该项目被认为是及时的，因为它将解决2012年EPSRC/MRC医学成像技术工作组报告中提出的两个挑战：“更安全、更低成本、更高吞吐量的系统和提高当前医学成像技术的价值”。该研究项目还将受益于与学术界（诺丁汉大学和昆士兰大学）和工业界（飞利浦医疗保健和Pulseteq有限公司）在MRI射频技术方面领先的国内和国际合作伙伴的密切合作，以确保其从一开始就对英国医疗保健行业产生直接影响。为了加速我们研究的影响，这里开发的SAR管理软件将向所有学术和临床MR研究人员提供免费许可证。申请人还将积极参与健康保护局咨询小组、英国放射学研究所和ISMRM安全委员会对我们的研究成果的审查。

Neural circuitry inside a human brain is probably the most complicated large scale circuit that any electronic engineer could think of: it contains around100 billion neurons, instantly linked through trillions of pathways to generate thought, memory, action or emotion. When faults rise within the neural circuitry, different brain disorders may arise. There are 10 million people in the UK living with a neurological condition, and the associated economic burden is estimated to be 116 billion pounds per year. With the recent significant investments in Brain Mapping projects in both Europe and the US, Neuroscientists need to develop new imaging technologies with superior signal-to-noise ratio (SNR), such as Ultrahigh field Magnetic Resonance Imaging, to show how complex neural circuits interact and eventually yield methods of preventing and treating neurological disorders. However various image artifacts presented at 7T MR images raise uncertainty of MRI measurements, because of increased inhomogeneity of the magnetic excitation field. The latest technology to solve this problem is to utilise a novel RF technology: Parallel Transmission (pTX), through multiple RF excitation coils with unique spatial profiles to achieve the homogeneous magnetization. To adopt this engineering solution for medical imaging application, there is a preeminent need to address one technical challenge: to develop a reliable RF safety management, measured by Specific Absorption Rate (SAR), to ensure this technique will meet the RF exposure legal constraint. Since the advent of pTX for 7T MRI, the difficulty in accurately characterising the local SAR, especially for the pTX pulse optimised for a specific subject, greatly hampers its clinical application. The current practice for SAR estimation in use worldwide by adopting electromagnetic (EM) simulation requires an expensive safety margin (at least 40% higher) to accommodate the uncertainties. The other SAR calculation approach recently pioneered by our project partner (Philips Healthcare), B1-map based ultrafast SAR calculation, still suffers from poor accuracy due to several unrealistic default assumptions, and has not been validated against quantitative experimental measurements at Ultrahigh field MRI yet. Within this project, we will develop a trustworthy and ultrafast SAR calculation approach to address this Engineering challenge by combining the electromagnetic numerical solution with B1 mapping for the first time, to enable a paradigm shift from using numerical simulation with generic models to proactively managing SAR for specific subjects under examination. In particular we will tackle the poor accuracy problem associated with the latest MRI B1-map based ultrafast SAR calculation by using complementary EM knowledge to supplement the missing information. It will be the first attempt to adopt such a hybrid approach, which could draw the essence of each technique, to tackle the MRI SAR issue in this global research forefront area.The project is seen as timely because it will address two challenges raised in the 2012 EPSRC/MRC Medical Imaging Technology Working Group Report: "Safer, lower cost, and higher throughput systems and Improving the value of current medical imaging technologies". The research project will also benefit from close collaboration with leading national and international partners in MRI RF technology from both academia (Nottingham University and Queensland University) and industry (Both Philips Healthcare and Pulseteq Ltd) to ensure its immediate impact on Healthcare Industry in the UK from the outset. To accelerate the impact of our research, the SAR management software developed here will provide free licences to all academic and clinical MR researchers. The applicant will also engage actively with the Health Protection Agency's Advisory Group, British Institute of Radiology and Safety Committee of the ISMRM for their reviews on our research output.