Surgeon’s innovative approach attracts federal investment for potential to shorten ICU stays by as much as 25%
What if doctors could test a customized treatment on a virtual version of a critically ill patient before administering it in real life?
That’s the promise of medical “digital twins.” Gary An, M.D., a University of Vermont (UVM) trauma surgeon and researcher at UVM’s Larner College of Medicine, has received a research and development contract worth up to $38 million to develop artificial intelligence-powered digital twins to help clinicians personalize treatment for critically ill patients, a UVM-Larner statement said. The research award is the largest in UVM history.
Funded through the Advanced Research Projects Agency for Health (ARPA-H)—a federal agency within the U.S. Department of Health and Human Services that supports pioneering, high-impact medical research—the anticipated five-year initiative aims to transform how clinicians diagnose and treat critical illness by modeling a patient’s unique immune response in real time. The project is part of ARPA-H’s Critical Illness Immunological Reprogramming and Control Point Learning Engine (CIRCLE) program.
If successful, researchers believe the technology could reduce intensive care unit (ICU) length of stay by at least 25 percent. Currently, 4.6 million Americans are treated in ICUs annually, costing up to $70 billion each year.
“UVM research has given rise to a number of truly transformative advancements in health care, and Dr. An’s project is the latest chapter in that legacy of vitally important work,” UVM President Marlene Tromp said. “The tremendous impact that digital twins could have on potentially life-saving medical diagnoses and treatment is evidenced by this historic level of research funding, and we as a university could not be more proud to be pioneering this incredible project.”
Dr. An’s project, known as ReSCUED, or Reprogramming Severe Critical Illness Using Extensible Digital Twins, seeks to address one of medicine’s most persistent challenges and correct the disordered immune response that drives diseases of the ICU, such as severe trauma, burns, or sepsis. For example, sepsis, a life-threatening reaction to an infection, occurs when the body’s immune system triggers widespread toxic inflammation that damages healthy tissues and can rapidly progress to organ failure. Sepsis requires urgent hospital care in an ICU with intravenous fluids, medications to support blood pressure, antibiotics and often machines to support failing organs, such as ventilators or dialysis.
“Sepsis is a huge health care problem, and one that will only get bigger as the population gets older and we get better at keeping people alive,” said An, the project’s principal investigator. “In the past, before ICUs, people simply died. Today, with advances in organ support technologies, we can keep people alive longer in the ICU, but often their bodies can’t get out of the immune dysfunction hole, and we don’t know how to help them out of it. The multidimensional dynamics of immune dysfunction is too complex for a person, even an expert, to comprehend. But we can train a computational model to do that.”
Under the ReSCUED approach, researchers will measure a critically ill patient’s immune and inflammatory state through frequent blood samples and physiological monitoring, create a digital twin that models the patient’s condition, and evaluate potential treatment strategies using FDA-approved medications before those treatments are administered. The goal is to provide clinicians with an individualized tool that can help them understand how a patient’s illness is evolving and use that information to guide care.
What is a Digital Twin?
A digital twin is a mathematical model of a real-world object, process, or system that uses real-time data to predict performance, identify problems, and guide decision-making. A familiar example may be the digital twins used in the aerospace sector. Sensors on an aircraft engine continually provide information about its performance, allowing a computational model to forecast how that specific engine may function over time and providing data to optimize aircraft operations.
Health care researchers envision digital twins as personalized models that can forecast how a disease may progress in an individual patient. By continuously adding and analyzing molecular and clinical data, the AI-powered tool could recommend the most effective treatment and help clinicians determine when and what medical intervention is likely to have the greatest impact.
When ReSCUED is implemented, a patient’s blood will be collected every six hours and measurements are made of the critical cells, proteins and molecules present. This data are applied to the digital twin to provide real-time information about the patient’s immune response, allowing the digital twin to refine its forecast as the patient’s condition changes. Researchers will use the digital twin to train an AI-based “virtual consultant” capable of analyzing highly complex biological interactions to aid a clinician by suggesting potential intervention strategies tailored to a specific patient and testing those treatment options virtually, before applying them to the patient, creating a new level of precision in critical care medicine.
“The complexity of critical illness exceeds what any individual can interpret in real time. Artificial intelligence gives us a way to evaluate those complex biological dynamics and determine how existing treatments might be used more effectively for a particular patient,” said Dr. An. “Our goal is to give clinicians a more precise understanding of what is happening within an individual patient and provide information that could help them select the right treatment at the right time.”
Multi-Site Project
UVM’s Larner College of Medicine will serve as the lead institution in this multidisciplinary initiative, which brings together clinical, computational, and biotechnology expertise from universities and companies across the country. The participants’ data will be collected at three clinical sites: at Wake Forest University School of Medicine and the University of Alabama at Birmingham Heersking School of Medicine through their participation in the Quantum Leap Healthcare Collaborative, and at Washington University School of Medicine. An’s team at Larner will focus on developing the computational models and digital twin platform. The project is expected to create new computational modeling jobs at Larner and strengthen Vermont’s growing biotechnology and AI research ecosystem.
Two private-sector partners will contribute specialized technologies. The DNA Medicine Institute of Cambridge, Massachusetts, will provide a bedside molecular testing platform originally developed for the International Space Station. InflammaSense, a California-based medical technology company, will provide wearable monitoring technology that measures activity in the vagus nerve, a key regulator of inflammation.
The project is structured as an up to five-year, milestone-based initiative. During the first three years, researchers will develop and validate the digital twin and demonstrate its ability to predict patient outcomes computationally. If these milestones are achieved, subsequent phases will test the technology in additional experimental settings before progressing to clinical trials involving critically ill patients.
By the end of the project, researchers hope to create an integrated platform that combines physiology, molecular testing, computational modeling, and AI to support clinical decision-making. The ultimate goal is to improve treatment selection, reduce ICU length of stay times, and improve health outcomes for critically ill patients.