Development of Image Analysis Tools for Pulmonary Spatial and Temporal Mapping

点击翻译按钮获取中文摘要

Motivation: Medical imaging has enabled both the internal structure and function of the lung to be studied in vivo and in three-dimensions. Numerous advances in image acquisition technology have resulted in computed tomography (CT) emerging as the most effective imaging modality for assessing lung structure, and magnetic resonance imaging (MRI) with inhaled hyperpolarized noble gases emerging as a novel method for visualizing lung function. Although CT and MRI quantitative measurements have been developed, approaches for integrating and quantifying the inter-relationships between these imaging measurements are required. Research Problem: Methods have been developed to classify the lung on a voxel-by-voxel basis according to structural disease state for CT (normal, small airway disease, emphysema) and according to functional disease state for MRI (normal or abnormal gas exchange). However, multiparametric approaches are required that can investigate the spatial heterogeneity of these features throughout the lung on a regional basis, the temporal heterogeneity of how these features change or transition over time, and combine the measurements from both CT and MRI to investigate the interaction between structure and function. Research Objectives: The objective of my research program is to develop novel multiparametric imaging measurements that interrogate and quantify pulmonary spatial and temporal relationships. Specifically, over 5 years, my objectives are: 1) develop spatial heterogeneity measurements and spatial inter-relationship measurements using CT images; 2) develop voxelwise temporal trajectories using longitudinal CT images; and, 3) develop multiparametric measurements using structural and functional information from CT and hyperpolarized 129Xe MRI. Anticipated Outcomes: By the end of our proposed 5-year research plan, we will have developed image analysis tools that generate novel measurements that characterize spatial and temporal heterogeneity. Our long-term goals are to incorporate these imaging measurements into machine learning algorithms for predicting adverse health outcomes in patients with lung disease. This work will generate highly qualified personnel for the private and academic sectors in medical imaging, imaging physics/engineering and software development. Impact: This research program will provide imaging scientists within the natural science and engineering fields of academia and industry with multiparametric image analysis tools that can incorporate information from multiple imaging time points and from different imaging modalities to yield novel spatial and temporal heterogeneity measurements. Importantly, these image analysis tools will be develop to extend to other imaging modalities or organ systems, allowing for use in a broad range of applications.