Development of perivascular space mapping toolset as a diagnostic aid for Alzheimer's disease

点击翻译按钮获取中文摘要

Project Summary Alzheimer’s disease (AD) is a devastating disease that affects millions of Americans and imposes a huge socio- economic burden. AD-related cognitive decline is associated with the accumulation of Aβ (Amyloid beta) plaques and neurofibrillary tangles of hyperphosphorylated tau protein, which are insufficiently cleared and degraded by mechanisms such as the glia-lymphatic system or by transport across the blood-brain barrier (BBB). Animal research has shown that the glia-lymphatic pathway plays a substantial role in the net clearance of Aβ. In addition, reduction of interstitial fluid efflux to the cerebrospinal fluid, or reduction of transport across the BBB, could lessen the Aβ clearance. Human magnetic resonance imaging (MRI) studies have also shown that the integrity of the perivascular space (PVS), the pathway of glia-lymphatic system, is a marker of glia-lymphatic brain health and its alteration is an important feature of AD pathology. PVS alteration in AD has been shown to be independent of Amyloid uptake, implicating astroglial involvement specific to AD. Information about PVS integrity could assist clinicians with making specific diagnosis about patients AD status and mechanism. Therefore, a tool that allows non-invasive in-vivo mapping of PVS from clinical MRI is of high significance to aid AD diagnosis and disease monitoring but does not yet exist. Current clinical routine to investigate PVS is a relatively crude approach based on counting total number of observed PVS in MRI by neuro-radiologists, which is non-specific, rater-dependent, and does not capture the distribution, whole extent of PVS change nor volumetric features of the PVS. From a practical point of view, this technique is time consuming and laborious. These limitations in part have prevented neurologists from adopting PVS quantification into their clinical routine. We have developed computational techniques for mapping and quantifying PVS morphology from MRI, which enables automated and accurate quantification of PVS across the brain without the need for time consuming manual intervention. These techniques were developed on research data, which had slightly higher resolution than clinical MRI. The goal of this proposal is to validate and optimize this technology on clinical MRI data, which can be variable in quality and resolution. We also aim to develop a scanner- and hardware-agnostic deployable analytical solution for automated PVS quantification, so that PVS mapping can be performed independent of the radiology/imaging IT infrastructure. If successful, we will have a fully developed and validated backend software that enables automated assessment of PVS and aids specific AD diagnosis and disease monitoring. Upon the completion of this goal, our Phase II will focus on translation of the technology via implementation and testing of the technology in collaboration with our strategic partners in clinic (Neuroradiology division of Keck School of Medicine), industry (BioGen Inc.) and research (MarkVCID and NASA). Our ultimate goal will be to integrate PVS measures into clinical diagnosis, disease monitoring and intervention efficacy assessment, as a specific non-invasive and in-vivo imaging marker of AD pathology that can be automatically and reliably measured.