Lung cancer is the deadliest type of known cancer in the United States, claiming hundreds of thousands of lives each year. However, despite the high mortality rate, the 5-year survival rate after resection of Stage 1A non–small cell lung cancer is currently in the range of 62%– 82% and in recent studies even 90%. Patient survival is highly correlated with early detection. Computed Tomography (CT) technology services the early detection of lung cancer tremendously by offering a minimally invasive medical diagnostic tool. Some early types of lung cancer begin with a small mass of tissue within the lung, less than 3 cm in diameter, called a nodule. Most nodules found in a lung are benign, but a small population of them becomes malignant over time. Expert analysis of CT scans is the first step in determining whether a nodule presents a possibility for malignancy but, due to such low spatial support, many potentially harmful nodules go undetected until other symptoms motivate a more thorough search. Computer Vision and Pattern Recognition techniques can play a significant role in aiding the process of detecting and diagnosing lung nodules. This thesis outlines the development of a CAD system which, given an input CT scan, provides a functional and fast, second-opinion diagnosis to physicians. The entire process of lung nodule screening has been cast as a system, which can be enhanced by modern computing technology, with the hopes of providing a feasible diagnostic tool for clinical use. It should be noted that the proposed CAD system is presented as a tool for experts—not a replacement for them. The primary motivation of this thesis is the design of a system that could act as a catalyst for reducing the mortality rate associated with lung cancer.
PDF
download
878987797
028-85220240
