Researchers at the Johns Hopkins Kimmel Cancer Center and Whiting School of Engineering — founding members of the Cancer AI Alliance (CAIA) — have announced two new projects that showcase how artificial intelligence can transform cancer research and patient care.
As the only research university and comprehensive health system in CAIA, Johns Hopkins brings together world-renowned expertise across medicine, engineering and public health. The Kimmel Cancer Center, Malone Center for Engineering in Healthcare, the Data Science and AI Institute and the inHealth Precision Medicine program all play key roles in the initiative.
CAIA, established in 2024, is built on a federated learning platform, which allows each of the participating cancer centers to keep their data secure while developing and sharing AI models.
"AI tools travel to the data, not the other way around," explains Vasan Yegnasubramanian, M.D., Ph.D., professor of oncology, pathology, and radiation oncology and molecular radiation sciences, and director of inHealth Precision Medicine at Johns Hopkins. "This preserves privacy while enabling broad collaboration."
In one project, led by Mathias Unberath, Ph.D., the John C. Malone associate professor of computer science, Jeff Weaver, Ph.D., director of research analytics, Yegnasubramanian and Alexis Battle, Ph.D., professor of biomedical engineering and computer science, director of the Malone Center for Engineering in Healthcare, and director of Research for Strategy and Partnerships, Data Science and AI Institute at the Whiting School of Engineering, the research team is fine-tuning a large language model using structured electronic health record data.
The model is being trained to follow patient trajectories over time, learning patterns that would allow it to predict later diagnoses, treatments or test results.
"It shows the potential of tailoring large AI models specifically for patients with cancer," says Battle. "This project, developed under an aggressive timeline, was a deeply collaborative effort that drew in faculty, students, research IT, and the Data Science and AI Institute's software engineering team."
The second project, led by Karisa Schreck, M.D., Ph.D., professor of neurology, Jessie Tong, Ph.D., assistant professor of biostatistics, Taxiarchis Botsis, M.Sc., Ph.D., associate professor of oncology, Weaver and Yegnasubramanian, is harnessing multicenter data to study IDH-mutant glioma and astrocytoma, rare subtypes of brain cancer. For the first time, Yegnasubramanian says, researchers are able to analyze practice patterns and outcomes across institutions from real-world use of new precision therapies for these tumors by leveraging the CAIA multi-cloud federated learning platform.
"The insights will help doctors individualize care, identifying patients most likely to benefit from targeted drugs and guiding if and when other treatments are needed," he says. "Ultimately, this work aims to provide patients and families with clearer, evidence-based guidance while protecting privacy."
The IDH gene was first discovered by Johns Hopkins Kimmel Cancer Center researchers in 2008. That discovery, along with a subsequent multicenter clinical trial, led to the 2024 FDA approval of the drug vorasidenib for IDH-mutant low-grade glioma.
These two Johns Hopkins-led studies are among eight projects launched under CAIA, which also includes Dana-Farber Cancer Institute, Fred Hutch Cancer Center and Memorial Sloan Kettering Cancer Center. Over the next year, CAIA leaders hope to expand to dozens of research models and add more cancer centers to the alliance, tackling challenges from predicting treatment response to understanding rare cancers.
"CAIA allows us to innovate across the full spectrum of cancer," says Yegnasubramanian. "On one end, we can develop foundational models trained on data from more than a million patients. On the other, we can study rare cancers and develop AI models that improve therapy for patients who previously had little guidance."
CAIA receives financial and technical support from technology partners Amazon Web Services (AWS), Deloitte, Ai2 (Allen Institute for AI), Google, Microsoft, NVIDIA and Slalom.
"Together, these projects demonstrate how Johns Hopkins and CAIA are laying the foundation for AI-driven advances that will help us make real progress against some of cancer's most complex problems," says Battle.
