Rapid MRI acquisition for pediatric low-grade gliomas

儿童低级别胶质瘤的快速 MRI 采集

• 批准号: 9231451
• 负责人: Kawin Setsompop
• 金额: $ 65.15万
• 依托单位: MASSACHUSETTS GENERAL HOSPITAL
• 依托单位国家: 美国
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-04-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9231451
• 关键词: Acceleration; Anatomy; Anesthesia procedures; Blinded; Brain; Brain imaging; Childhood; Clinical; Clinical Protocols; Data; Diagnosis; Diagnostic; Diffuse; Glioma; Goals; Image; Imaging technology; Joints; Magnetic Resonance Imaging; Measures; Methods; Modernization; Monitor; Motion; Optics; Pathology; Patients; Performance; Perfusion; Permeability; Phase; Physiological; Physiology; Population; Protocols documentation; Resolution; Sampling; Scanning; Scheme; Series; Services; Slice; System; Techniques; Technology; Testing; Thick; Three-Dimensional Imaging; Time; Tissues; Vascularization; base; compliance behavior; contrast imaging; cost; cost effectiveness; data acquisition; design; image reconstruction; improved; novel; outcome forecast; pediatric patients; public health relevance; reconstruction; temporal measurement; time use; tumor; two-dimensional

 DESCRIPTION (provided by applicant): We will develop technology that improves MRI's acquisition efficiency by 20× to create a comprehensive 9- minute brain exam that provides more detailed structural and functional information than current 40-minute clinical exams. Such exam should improve MRI's diagnostic power while greatly increase patient throughput and compliance. We first apply this technology to pediatric low-grade gliomas (LGG), where frequent tumor monitoring and anesthesia services are required. Anesthesia is detrimental to the developing brain and results in a nine-fold increase in cost. We will couple our short exam with a commercially available optical real-time motion tracking system to robustly remove the need for anesthesia. The proposed methods allow acquisition of more detailed quantitative physiology that could potentially improve diagnosis and prognosis of LGG. The slow image encoding in MRI has been its critical limiting factor. To provide whole-brain imaging quickly, clinical protocos use two-dimensional (2D) slice-by-slice imaging with high in-plane resolution but 4-5× thicker slices with a 20-40% gap. The gaps can result in missed information, while thick slices limit the ability to perform multi-planar reformats, which necessitates re-imaging if viewing in a different image plane is desired. On the other hand, quantitative measures of perfusion and vascularization require high temporal resolution single-shot EPI. Here, the slow encoding in the phase encode direction results in detrimental image distortion, compromised resolution/coverage, and limited physiological information. Moreover, with a wide variety of tissue contrast mechanisms-each sensitive to different aspects of pathology-patients are imaged with 5-8 scans with overlapping information. The result is exams of up to 40 minutes with suboptimal resolution/information. To overcome these issues and achieve rapid, high-quality and detailed imaging, we will develop "Wave-CAIPI" technology, a data acquisition/reconstruction scheme designed to optimally exploit available information in modern multi-channel receivers and in multi-contrast/time-series data for improved image encoding, to achieve high-quality 20× acceleration. With Wave-CAIPI, we will also replace standard 2D imaging with far more SNR-efficient Simultaneous Multi-Slice (SMS) and 3D imaging to achieve rapid imaging with high SNR. Initially, to achieve 10-15× acceleration, we will develop Wave-CAIPI, which applies efficient data sub- sampling concepts fully to all three spatial directions of an imaging volume. To achieve 20× acceleration, Wave-CAIPI will be augmented with "CS-Wave-CAIPI", which extends efficient sub-sampling across contrasts and time-series data. Joint Bayesian and temporal Compressed Sensing reconstructions will be developed, and a highly efficient "HSS" solver will be created to facilitate near/real-time reconstruction. Finally, we wil test the hypothesis that Wave-CAIPI can reduce MRI exam time for pediatric low-grade gliomas from 40 min to 9 min and eliminate anesthesia while providing more comprehensive diagnostic information.

 描述（由申请人提供）：我们将开发技术，将MRI的采集效率提高20倍，以创建一个全面的9分钟脑部检查，提供比当前40分钟临床检查更详细的结构和功能信息。这样的检查应该提高MRI的诊断能力，同时大大增加病人的吞吐量和依从性。我们首先将这项技术应用于儿科低级别胶质瘤（LGG），需要频繁的肿瘤监测和麻醉服务。麻醉对发育中的大脑有害，导致成本增加9倍。我们将结合我们的短期考试与商业可用的光学实时运动跟踪系统，以鲁棒地消除麻醉的需要。所提出的方法允许获得更详细的定量生理学，这可能会改善LGG的诊断和预后。 磁共振成像中的图像编码速度慢一直是其关键限制因素。为了快速提供全脑成像，临床方案使用具有高平面内分辨率的二维（2D）逐层成像，但切片厚4-5倍，间隙为20-40%。间隙可能导致信息丢失，而厚切片限制了执行多平面重新格式化的能力，如果需要在不同的图像平面中查看，则需要重新成像。另一方面，灌注和血管化的定量测量需要高时间分辨率的单次激发EPI。这里，相位编码方向上的缓慢编码导致有害的图像失真、受损的分辨率/覆盖范围以及有限的生理信息。此外，通过各种各样的组织对比机制-每种机制对病理学的不同方面敏感-患者通过具有重叠信息的5-8次扫描成像。结果是长达40分钟的检查，分辨率/信息不佳。为了克服这些问题并实现快速、高质量和详细的成像，我们将开发“Wave-CAIPI”技术，这是一种数据采集/重建方案，旨在最佳地利用现代多通道接收器和多对比度/时间序列数据中的可用信息，以改进图像编码，实现高质量的20倍加速。借助Wave-CAIPI，我们还将用SNR效率更高的同步多切片（SMS）和3D成像取代标准的2D成像，以实现高SNR的快速成像。 最初，为了实现10-15倍的加速，我们将开发Wave-CAIPI，它将高效的数据子采样概念完全应用于所有三个空间方向 成像体积。为了实现20倍的加速，Wave-CAIPI将增加“CS-Wave-CAIPI”，这将在对比度和时间序列数据中扩展有效的子采样。将开发联合贝叶斯和时间压缩感知重建，并将创建一个高效的“HSS”求解器，以促进近/实时重建。最后，我们将检验Wave-CAIPI可以将儿科低级别胶质瘤的MRI检查时间从40分钟缩短到9分钟，并消除麻醉，同时提供更全面的诊断信息的假设。