Enabling Kinematic Joint Profiling Using MRI

项目总结

Project Summary We propose a technical feasibility study seeking to develop methods for quantitative kinematic profiling of moving joints using magnetic resonance imaging (MRI). In the context of this study, a kinematic profile is defined as a collection of joint characteristics computed and tracked during the course of movement. This project is motivated by the hypothesis that such profiling of moving joints can highlight dysfunction, treatment progress, and point towards favorable (or unfavorable) surgical interventions. At a high level, it is envisioned that the proposed kinematic profiles could fit into clinical management workflows much in the same way as blood biomarker panels. While kinematic imaging of joints can be performed using plain-film (PF) X-ray, computed tomography (CT), and ultrasound (US) methods, MRI is the gold-standard for advanced orthopedic assessment and is an appealing option for accessory kinematic analysis. A set of relatively fast kinematic profiling acquisitions could feasibly be added to routine orthopedic MRI exams, thereby providing optimal diagnostic imaging in both static and kinematic contexts within a single visit. Though several preliminary studies have hinted at the potential diagnostic value of kinematic imaging data, such data is difficult to interpret and cannot easily be quantified or captured in clinical records. In this study, we seek to establish fundamental methods that can provide simple and easily digestible kinematic imaging reports with data acquired in a short scan interval using conventional clinical MRI equipment. As a preliminary feasibility investigation of these methods, kinematic imaging of the wrist will be studied. Dysfunction of the scaphoid and lunate bones in the wrist is a well-studied kinematic problem of diagnostic significance. Novel 4D zero-echo-time MRI of the wrist will be used to capture the kinematic imaging using for profiling of the scaphoid-lunate mechanics during two established wrist movement patterns. The goal of this project is to establish and demonstrate methodological components required for MRI kinematic profiling. Data collection on a modest-sized cohort of 100 healthy control subjects is proposed for this purpose. Novel MRI pulse-sequence and post-processing development components are introduced and tasked for analysis of this normative data. Using the acquired MRI data, kinematic parameters for each dynamic dataset will be extracted and curated into a multi-parametric profile for each subject. Aim 2 of the study proposes the use of external sensor motion capture methods to validate the MRI-based kinematic parameter measurements on 50% of the study cohort. Finally, Aim 3 of the study seeks to use machine-learning clustering approaches to develop a kinematic profile normalization procedure using the acquired control dataset. Such normalization is a crucial milestone in the translation of kinematic profiling to the clinic and will establish a baseline for future translational studies of symptomatic cohorts.