Model for quantification and analysis of pulmonary emphysema from low-dose radiation CT scans

This thesis presents techniques for the automated analysis of low-dose helical CT scans used in the detection of pulmonary emphysema. Our work focused on solving the following problems: (1) Determine if existing quantification methods can be successfully used in the detection and quantification of pulmonary emphysema from helical low-radiation CT scans. (2) Develop better models for quantifying emphysema based on additional spatial and densitometric information. The primary contributions of this work are:; The development of an automated segmentation algorithm for low-dose lung CT scans. The algorithm is modular, allowing for the removal of surrounding body tissue, vessels, and airways.; The development of the sliding window method for quantifying emphysema.; The development of a graphical display method of the emphysema index based on slice location. The graphical display will allow radiologists to not only quantify emphysema for the entire lung, but also to understand what regions of the lung are most affected by the disease.; The development of the spatial adaptive filtering method necessary for later quantification of pulmonary emphysema. The adaptive filtering method is customized to the noise characteristics present in low-dose chest CT scans.; The development of more accurate metrics capable of better quantifying the severity of emphysema.; The development of metrics capable of classifying different types of pulmonary emphysema.