Computational Flow Dynamics Predicting Pulmonary Blood Flow for Pediatric Congenital Heart Disease; How Food Deserts in Georgia Are Kept Dry

Congenital heart defects (CHD) are a group of conditions that affect the functionality of a patient’s heart beginning at birth, and it affects about 40,000 births in the United States each year. Two of the key tools cardiologists use to diagnose and assess the severity of these conditions are computed tomography (CT) scans, and nuclear lung perfusion (NLP) scans. CT scans give the 3-D structure of the heart with a series of X-ray images, and NLP scans give the percentage of blood flow to different sections of the lungs. This project aims to find an alternative to NLP scans by utilizing computational flow dynamics to predict blood flow to each part of the lungs using a CT scan. The goal was to convert the patients’ CT scans into 3-D models, and run flow simulations on this model that produce the same output as a NLP scan. A pipeline to convert CT scans into models compatible with flow simulations was successfully created, and initial flow simulations were performed. The simulations were not able to run to 100% completion due to computational abnormalities discussed in the results, but the success of this project lies in the fact that it established a pipeline to go from CT data to a flow dynamics output, paving the way for future work.

School of Engineering and Applied Science

Bachelor of Science in Biomedical Engineering

Technical Advisor: Michael Shorofsky, Haibo Dong

STS Advisor: Bryn Seabrook

Technical Team Members: Christian Marinaro, Ritik Verma