Exploring the Future of Digital Twins in Oncology: Highlights from the JCCO Retreat

What if we could develop digital twins of patients capable of predicting and optimizing their unique responses to cancer treatments in real time? At the fifth annual Joint Center for Computational Oncology (JCCO) retreat, held Nov. 18 at the Oden Institute for Computational Engineering and Sciences at The University of Texas at Austin, researchers and clinicians explored how future advancements in digital twin technology could make this vision a reality.

The collaborative retreat brings together experts from The University of Texas’ MD Anderson Cancer Center and the Oden Institute. This year’s event, Moving Digital Twins from the Server to the Clinic, centered around dynamic, data-driven models that simulate individual patient systems to revolutionize cancer treatment. Featuring computational science, oncology, and clinical care researchers, presentations included interactive discussions and thought-provoking debates, envisioning the potential of digital twins in clinical decision-making while reflecting on goals from the 2023 retreat.

A digital twin is a computational replica of a patient’s physiological system, capable of simulating predictive responses to treatments. That’s the promise of digital twins, which have the potential to balance and improve chemotherapy dosages and forecast tumor progression, offering a glimpse into a future where personalized medicine becomes a reality.

“It was very exciting to see how the projects have advanced since we started this effort five years ago,” said Thomas Yankeelov, Director of the Center for Computational Oncology at the Oden Institute and professor of biomedical engineering at UT. “In addition to the individual studies, it is very encouraging to see the level of interaction, cooperation, and collaboration we have achieved. With six more pilot projects getting off the ground and a primary goal of ramping up our inter-institute training program on the docket for 2025, our best times are still ahead of us.”

 

The keynote presentation by Guillermo Lorenzo, a postdoctoral fellow at the Oden Institute and a researcher affiliated with the University of A Coruña, talked about the potential mechanistic learning approach integrating physics-based growth models with data-driven insights. Lorenzo outlined how these frameworks could enhance early detection and improve personalized clinical management for patients undergoing active surveillance or post-radiotherapy care. His presentation concluded with a clear roadmap for future research, heavily emphasizing the importance of refining digital twin models, testing in virtual clinical trials, and assessing real-world applications.

During the retreat, presentations on digital twins ranged from theoretical underpinnings to practical applications. Researchers discussed the complexity of their models, which incorporate multiple parameters—such as heart rate, glucose levels, and blood pressure—to simulate an entire system accurately while emphasizing that modeling a complex disease like cancer presents immense challenges.

John Hazle, PhD, from The University of Texas MD Anderson Cancer Center’s Department of Imaging Physics, noted the value of the retreat in fostering collaboration. “These retreats allow MDACC scientists to see presentations from UT Austin/MDACC pilot project teams and to develop collaborations with UT faculty,” said Hazle. “The retreat is always exciting—seeing the good work coming out of the UT and MD Anderson faculty from the pilot studies—and a chance to discuss computational oncology science with our colleagues in Austin.”

The retreat’s “World Café” breakout sessions created engaging discussions on three central topics: 1) Identifying “low-hanging fruit” for digital twins in oncology, 2) Making digital twins practical for clinical oncology, and 3) Balancing efficacy and toxicity via digital twins. Each group delved into solutions, including the development of classification systems to assess the maturity of digital twin models—akin to NASA’s Technology Readiness Levels.

Participants explored key hurdles in advancing digital twins. One pressing issue is data scarcity. Models require vast datasets to achieve the predictive power needed for clinical use, yet acquiring such comprehensive data remains a significant barrier. Another challenge lies in balancing efficacy and toxicity. For instance, digital twins could help determine whether increasing a chemotherapy dose—despite its toxicity—would provide enough benefit to justify the risk.

Participants also debated whether digital twins should focus on treatment optimization or de-escalation. While optimization—improving the balance between toxicity and therapeutic outcomes—emerged as a priority, de-escalation was recognized as vital for minimizing unnecessary treatments.

Yankeelov reflected on the progress made since the center’s inception, highlighting new training initiatives. “Two years ago, we started an effort to train postdoctoral fellows at the interface of mathematical/computational modeling and oncology. The idea was to identify post-docs and have them jointly appointed at MDACC and the Oden Institute, spending significant time at both institutions. To date, I believe we have three post-doctoral fellows who are doing this, and we are looking to expand this program.” 

Hazle commented on the potential future of digital twins, saying, “We think digital twins are rapidly finding a place in translational and clinical cancer research. I expect these approaches to gain some clinical acceptance in the next 3–5 years.”

Digital twins represent a bold new frontier where the complexities of cancer and its treatment can be tackled in the coming years. Though many challenges remain, the discussions at this year’s retreat affirmed a collective commitment to turning these concepts into lifesaving realities.