Quantitative regional assessment of pulmonary parenchymal dynamics using image registration and analysis techniques
Abstract
The lung is a highly elastic organ, composed of an intricate network of blood vessels and airways that terminate in alveoli---the site of gas exchange. Normal alveolar walls are extremely thin and highly compliant, allowing the lung to deform easily as air passes through. Various pathological processes interfere with the normal respiratory deformation of the lung, either by disrupting the alveolar architecture or affecting the resistance and makeup of the airways. Characterization of normal lung motion patterns seeks to further the understanding of the changes brought about by these diseases. ^ In this dissertation, we present an image-based approach toward quantifying pulmonary kinematics. The methodologies detailed here comprise a thorough analysis of lung motion via image registration and other processing techniques. Using the steps described, we are able to estimate the motion of the lung parenchyma by establishing voxel-wise spatio-temporal correspondences between successive image frames, and infer not only global parameters, such as instantaneous lung volume, but regional parameters such as displacement and strain. Furthermore, we introduce methods and techniques that can be used to construct an atlas of normal lung motion during a single breath by performing physiologically appropriate temporal interpolations between acquired image data. These techniques are neither modality- nor species-dependent; the results presented are acquired from both magnetic resonance (MR) and computed tomography (CT) datasets in murine models, healthy volunteers and patients. ^
Subject Area
Engineering, Biomedical|Health Sciences, Radiology|Biophysics, Medical
Recommended Citation
Tessa A Sundaram,
"Quantitative regional assessment of pulmonary parenchymal dynamics using image registration and analysis techniques"
(January 1, 2007).
Dissertations available from ProQuest.
Paper AAI3260995.
http://repository.upenn.edu/dissertations/AAI3260995