Automated segmentation of the pulmonary artery in CT scans for the detection of pulmonary arterty hypertension

Heart disease is still the leading killer today in the United States, with almost 600,000 people dying of a heart related disease in 2010. Because of this, exploring new techniques of early heart disease detection and more advanced treatment options continues to be of paramount importance. Today, detection and diagnosis of heart disease can be done through both invasive and non-invasive measures, including lipid panels, catheterizations, and advanced imaging techniques amongst more traditional methods such as patient demographics and overall weight, blood pressure, and cholesterol levels. The problem we now face is that traditional methods of detection can only go so far, with 40-50% of heart attack victims exhibiting no signs of disease, and most of these people not willing to undergo a painful or expensive procedure without the signs of disease being present. With that being said, new imaging techniques could provide an avenue for doctors to diagnose heart related diseases, as imaging of the heart is relatively quick, non-invasive, and could possibly have costs covered by medical insurances. This research will attempt to develop an algorithm capable of detecting possible pulmonary artery hypertension from CT scans by segmenting the artery, collecting pertinent data, and comparing arteries from healthy and diseased patients. The resulting algorithm will be made into a software plugin package for the DICOM viewer OsiriX