Sep 30, 2026
Dr. Gary An received a record breaking award for his research on AI “twins.” Courtesy photo by Jeff Paxton BURLINGTON — For doctors struggling to treat patients with severe burns or infections, researchers at the University of Vermont are developing a not-so-human solution.  They’re working to create an artificial intelligence “twin” of a patient, allowing doctors to test individualized treatments before they’re used on humans. The university received its largest research award ever— up to $38 million — to fund the development of the AI model for clinical trials.  “It reinforces UVM’s role as a leading research university dedicated to improving lives,” university President Marlene Tromp said at the school’s event announcing the federal grant Wednesday.  The work is important because patients with severe infections, burns or injuries can have very disordered immune systems, according to trauma surgeon and researcher Dr. Gary An, who is leading the project. Patients can fall “into a pit” in which their immune system can’t recover, and doctors have an incredibly difficult time treating them, he said.  “This is extraordinarily frustrating to me,” An said.  Enter the AI “twin.”  The twin is a computer model that mimics the brain’s neural networks, representing a human body responding to an illness or injury, An said. Researchers are training an AI program that treats those conditions like a game, with the AI working to win the game and cure the patient.  A doctor would take a sample of a patient’s blood, give it to the AI, then receive a recommendation from the AI about how to treat that patient, An said. That way, doctors might use more effective treatments and shorten patient stays in intensive care.  An’s research will develop the model before it goes through clinical trials. It could take seven to 10 years before the model might be approved for widespread use on patients, said Kirk Dombrowski, vice president of research at UVM, during an interview Wednesday. Six researchers currently work alongside An in the lab researching the AI twins, according to the lab’s website.  The grant comes via the Advanced Research Projects Agency for Health, an agency within the national Department of Health and Human Services that invests in medical technology research. The funds are expected to last across the project’s planned five years of work. UVM’s announcement of the award comes just a week after it received a $19 million grant to build a national neuroscience observatory. And it comes nearly two years after the university earned R1 status in 2025, the highest designation a research university can receive.  “I think we’ve just shown federal agencies and colleagues and partners that we’re capable of working at a different scale,” said Dombrowski, referring to the record award.   While researchers at UVM will be leading work on building computational models and training the AI program, the research effort is coordinated among multiple hospitals and companies across the country.  UVM researchers are collecting data from patients who opt in at three sites: Wake Forest University School of Medicine in North Carolina, WashU Medicine in Missouri, and University of Alabama at Birmingham Heerskink School of Medicine.  The universities were chosen because UVM has existing research relationships with the institutions, according to Dombrowski. UVM’s hospital isn’t included in the data collection because it doesn’t treat a large enough number of intensive care patients.  Dombrowski said that funding for An’s project comes as the federal government has signaled that it’s willing to invest in AI development.  “I think that the scale of what we’re capable of in research has just grown so dramatically,” Dombrowski said. Read the story on VTDigger here: UVM awarded record $38M to create ‘AI twin’ of sick patients. ...read more read less
Respond, make new discussions, see other discussions and customize your news...

To add this website to your home screen:

1. Tap tutorialsPoint

2. Select 'Add to Home screen' or 'Install app'.

3. Follow the on-scrren instructions.

Feedback
FAQ
Privacy Policy
Terms of Service