My name is Jason Lee, and I’m a rising senior pursuing a degree in Computational Biology with minors in Applied Statistics and Computer Science. When I’m not in lab or attending classes you’ll often find me hitting the heavy bags in the matted room at Trees as I am an avid fan of MMA. Some of my other interests include swimming, playing the guitar, traveling, and learning new languages.
This summer, I will be working at the Vascular Bioengineering Laboratory in the Center for Biotechnology and Bioengineering under my mentors Dr. David Vorp and Dr. Timothy Chung. My project aims to design a device capable of identifying patients at risk of a condition called Deep Vein Thrombosis (DVT) and diagnosing those that already have the condition. DVT, in short, involves the formation of blood clots, primarily in the deep veins of the thigh. These clots can dislodge and cause chest pain and breathing difficulties when they become lodged in the lungs (a pulmonary embolism for my medically-inclined readers). Additionally, DVT often progresses into another serious condition called Post-Thrombotic Syndrome (PTS). PTS occurs from the weakening of vein walls/valves, reducing their effectiveness in returning blood from the feet to the heart against gravity. The best approach to mitigate the impact of DVT is to therefore detect it as early as possible. The current clinical tests for DVT—D-dimer Blood Tests, MRI, Venography, and Ultrasound—are generally accurate, but have drawbacks such as being invasive (needing to enter or open the body), being highly expensive, and, having a limitation of diagnosing DVT only after symptoms or blood clots have appeared. My device hopes to improve upon all these points by being non-invasive, affordable, and capable of identifying the risk of DVT before the onset of symptoms or blood clot development.
My project consists of two integral components: hardware design and machine learning (ML). In the hardware section I will design a device incorporating an array of piezoelectric sensors within a pressure cuff to measure vein compliance, the ability of veins to stretch and expand in response to changes in blood volume or pressure, which I hypothesize differs significantly in patients with and without DVT. For the ML section I will use Python to train models with de-identified compliance data collected from a previous clinical study. In practice, the compliance waveforms collected from future patients will be processed and inputted to the final model to provide a prediction of DVT risk level.
I initially planned to attend medical school, but I became increasingly interested in computer science and engineering. I am considering pursing a graduate degree in an engineering discipline or doing work on medical devices. Medical school may be something I pursue in the future, but I decided to shelve that goal to explore my other interests. The Brackenridge Fellowship presents a valuable opportunity for me to share my research with a diverse audience and gain insights into the work of my peers. I believe that being a part of this interdisciplinary community will help me shape my career trajectory and expand my knowledge of various fields.
A fun fact about me: I was born in New York but raised in the Bay Area of California.
