A TOF, DOI, MRI compatible PET detector to support sub-millimeter neuroPET imaging

点击翻译按钮获取中文摘要

Abstract The overall goal of this research is to develop next generation positron emission tomography (PET) detector technology to support non-invasive, quantitative brain imaging at spatial and temporal resolutions currently not achievable with human neuro-PET systems. The developed PET detector technology will also be compatible with operation in an MRI system. The proposed research is targeted to the NIH Brain Research through Advancing Innovative Neurotechnologies (BRAIN) initiative. Human brain imaging with PET/MRI will be an essential tool in neuroscience studies to "Develop innovative technologies to understand the human brain and treat its disorders; create and support integrated brain research networks." The key advancement that we introduce is a PET detector with <100 psec time-of-flight (TOF) PET coincidence resolution, <2 mm continuous depth of interaction (DOI) positioning and intrinsic detector spatial resolution to support <1 mm PET image resolution throughout the system imaging field of view (FOV). Detector modules have been designed that individually achieve these performance metrics; however, no detector module has been designed that supports all of them. To make disruptive advancements in neuro-imaging using PET, one must push the image spatial resolution (i.e., currently 2-3 mm image resolution) as well as coincidence timing resolution and also be MRI compatible. The impact of this project is that we will advance the state of the art in all of these critical performance areas. We will achieve these goals by first understanding the role that different PET detector performance parameters have on task-based figures of merit for neuroPET imaging. We will investigate how both TOF and image resolution impact figure of merit performance for estimation, detection and characterization imaging tasks. Monte Carlo simulation will be used along with both object-based (i.e., mathematical) and anthropomorphic digital phantoms. Next we will optimize SiPM device selection, electronics and detector geometry for <100 psec TOF coincidence timing, 1 mm intrinsic spatial resolution and <2 mm DOI positioning resolution. We will build and characterize a prototype PET detector module utilizing a novel dual- sided slat crystal detector design. To advance coincidence timing performance we will investigate the use of machine learning to estimate the arrival time of detected events. Finally, we will optimize the detector design for MRI compatibility. We will fully test and characterize performance of our prototype detector on the benchtop and in a MRI scanner while running clinical MRI pulse sequences. At the end of this developmental project we will be in position to build a state of the art, MRI compatible, TOF, DOI PET imaging system.