Bright Project Shines
Bright Project Shines
How Stanford Computer Science Students Used AI for Research and Care
Talk about killing two birds with one stone! Stanford’s Division of Blood and Marrow Transplantation and Cellular Therapy has figured out how to save thousands of work hours annually, all while giving undergrads in computer science an unforgettable learning experience.
Cancer registries play a foundational role in the advancement of medicine by facilitating the identification, understanding, treatment, and prevention of disease. But building and maintaining these databases is a massive endeavor. No one understands this better than the bone marrow transplant community, where participation in the Center for International Blood & Marrow Transplant Research (CIBMTR) registry is mandatory. Stanford alone dedicates a dozen full-time employees and about 23,000 work hours annually to meeting their CIBMTR registry requirements and maintaining their in-house database.
Vanessa Kennedy led an innovative project leveraging large language models and AI to help automate population of the Center for International Blood & Marrow Transplant Research (CIBMTR) registry.
The effort is worthwhile. “Nearly 100 publications have come up from [the CIBMTR registry], nearly all giant clinical trial ideas,” says Vanessa Kennedy, MD, assistant professor of blood and marrow transplantation and cellular therapy at Stanford. Living up to the Stanford tradition of innovation and multidisciplinary cooperation, Kennedy and her team asked, How can we leverage our expertise in information technology (IT) to do this smarter?
Through the Business and Research Innovations to Excellence (BRITE) program, Kennedy and colleagues enlisted undergraduate computer science students to come up with an artificial intelligence (AI) large language model solution that extracts data directly from the electronic medical record, including patient notes, to populate the Stanford database and meet CIBMTR requirements.
To say these students met their mandate is an understatement. At the 2025 Transplantion and Cellular Therapy Annual Meeting in Honolulu, their presentation won an award for best abstract. The transplant community understands what this kind of innovation could mean for the future. Facilitating the collection of crucial information that can guide care and direct innovative therapies with fewer work hours means faster access to more data and more resources to divert elsewhere.
“We have the right computer science students. We have the right amount of data. We have an amazing IT team that found a way to be completely HIPAA compliant and respect patient privacy. I think working in a place that doesn’t immediately shoot down innovative ideas is huge.” – Vanessa Kennedy, MD
This “could have only happened at Stanford,” says Kennedy. “We have the right computer science students. We have the right amount of data. We have an amazing IT team [that found] a way to [be] completely HIPAA compliant and respect patient privacy. I think working in a place that doesn’t immediately shoot down [innovative] ideas is huge.”
For their part, the students are eager to tackle a real-world problem rather than hypothetical issues limited to the classroom. Stanford student Edwin Pua worked on the BRITE project as part of his undergraduate computer science degree, describing it as “a life-changing experience” that offered a unique opportunity to “apply what I learned in university to help change the lives of real people. That meant a lot to me! Our North Star was really asking ourselves, How can we streamline data retrieval so that clinicians can focus on what really matters: caring for patients? I’m grateful to have been a part of something so meaningful.”
