By Karen Copeland
Lung cancer is the leading cause of cancer-related deaths in men and women. In addition, many patients with progressive lung cancer develop latent radiation-induced respiratory damage, even after they are declared cancer-free. The purpose of this investigation was to use three-dimensional modeling techniques using the java programming language to create an interactive environment through which researchers can better study the human lung structure and identify early markers of latent radiation-induced vascular injury such as pulmonary arterial hypertension (PAH). PAH is a condition that can be caused by radiation therapy treatment and can result in heart failure. In addition, it is very difficult to diagnose because its most-common symptoms are also found in other more well-known diseases such as asthma.
For this project, a program was written that reconstructs a virtual model of the human vascular system from CT scans using techniques such as automatic and snake-based segmentation, manually-selected seed points, pixel intensity thresholds, 3D skeletonization, the gatortail method (a novel, in-house algorithm), distance transformation, and recursive in-order tree traversal. Vascular characteristics that have not been studied in-depth since the 1970s such as bifurcation angle (the angle between two children vessels branching from a common parent), area and volume ratio of these children and parent vessels, vessel branch radius, and tortuosity (the degree of curvature of a vessel) were also calculated and graphed for analysis. With this data, we found that some traits, such as bifurcation angle, remain constant, while others, such as branch radius, decrease with distance from the center of the lung system. When comparing healthy lung systems to those diagnosed with PAH, we discovered that the tortuosity measurements in patients with PAH was slightly higher than the tortuosity measurements in healthy lungs, likely as a result of the increased pressure in the vessels caused by obstructions such as scar tissue and inflammation. Using this data, this research has the potential to optimize quicker diagnosis of subtle vascular damage and further the understanding of the human lungs.
This summer, I am working as an intern at Columbia University’s Earth Institute, under the supervision of Dr. Maria Tzortziou. There, I am working on a project that utilizes satellite data to study nutrient trends in the waters of the Long Island Sound (LIS). Using remote sensing techniques, the goal of this project is to analyze spatial, seasonal, and temporal trends of organic nutrients such as Chlorophyll-a and Dissolved Organic Carbon concentration, along with inorganic nutrients such as Nitrogen and Nitrate/Nitrite concentration. The presence and/or absence of these nutrients can serve as strong indicators of the health of the LIS system, giving us important information to better understand human’s influence on their surrounding environment, something that is especially evident in densely populated New York. Along with studying general trends, this research is also focused on specifically determining what changes can be seen in satellite nutrient data in Spring 2020 versus previous years, with the hopes of analyzing the effect of the current health situation and social distancing restrictions on the environment and pollution levels.