The symposium brought together national leaders in computational oncology with UNC faculty and learners, offering a deep dive into advances in genomics, learning health systems and clinical imaging.
With 18 scientific presentations, four external keynote speakers, and 20 posters, the event showcased UNC Lineberger’s growing role as a hub for cutting‑edge cancer data science.
Launching the day: building the future of cancer data science
Executive Director of UNC Lineberger, Robert Ferris, MD, PhD, opened the symposium by emphasizing the importance of data‑driven innovation for the cancer center’s research mission.
Session 1 – Genomics in Precision Oncology
Introduced by UNC Lineberger’s Charles Perou, PhD, the first session examined how genomics and computational analytics are transforming precision medicine.
Keynote speaker Rafael Irizarry, PhD, of Harvard T.H. Chan School of Public Health, discussed strategies for building impactful data science departments within research institutions. Presentations by UNC faculty highlighted the use of AI‑powered biomarker‑driven clinical trial designs, integrating precision oncology into clinical workflows, and leveraging large language models (LLMs) for hypothesis testing and next‑generation cancer omics analytics.
A panel discussion brought together all session speakers to explore reproducible analytics, cross‑disciplinary collaboration, and the infrastructure needed to bring precision oncology into routine practice.
Session 2 – Improving care quality through oncologic learning systems
Session 2, led by UNC Lineberger members Louise Henderson, PhD, MSPH, and Jennifer Elston Lafata, PhD, highlighted how learning health systems embedded within clinical workflows can improve cancer care quality.
Keynote presenter Sylvia Plevritis, PhD, of Stanford School of Medicine, detailed advances in multimodal AI for clinical decision support. UNC researchers demonstrated tools using reinforcement learning, EHR‑based risk prediction models, and AI systems designed to close cancer data gaps and support real‑world treatment decisions. Presentations underscored the importance of building systems responsive to both patients and providers, including approaches to predict and prevent high‑risk clinical events.
Session 3 – Image data and the next frontier of precision oncology
Led by Melissa Troester, PhD, MPH, and Katherine Hoadley, PhD, of UNC Lineberger, the final session focused on pathology and radiology innovations poised to accelerate image‑based precision oncology.
Keynote speaker Gregory Goldgof, MD, PhD, from Memorial Sloan Kettering Cancer Center, shared approaches for developing end‑to‑end AI systems in diagnostic hematology and pathology. UNC and industry presenters highlighted efforts to overcome barriers to clinical AI deployment—such as scalability and batch effects—and demonstrated multimodal imaging‑EHR models, foundation models for pathology, and AI‑driven prognostic tools.
“The talks today demonstrate the talent at UNC that will transform how cancer data shapes patient care quality and outcomes in North Carolina,” Troester said.
Poster session spotlights next‑generation talent
The day concluded with a poster session and reception featuring 20 projects from students and junior investigators across computational biology, population science and AI‑driven imaging analytics. Industry collaborators, including Tempus and Ataraxis AI, also contributed presentations.
A shared vision for integrated cancer data ecosystems
Across all sessions, a unifying theme emerged: the need to move beyond isolated algorithms toward integrated learning systems that connect genomics, imaging and clinical data to improve diagnosis, prognosis and patient outcomes.
Deputy Director for Research at UNC Lineberger Jeremy Rich, MD, MHS, MBA, closed the event by highlighting UNC Lineberger’s commitment to building an ecosystem where advanced analytics accelerate both discovery and care.
