As the use of artificial intelligence across healthcare increases nationwide, The Daily Pennsylvanian spoke to professors, physicians, and researchers at Perelman School of Medicine about how they are integrating AI and machine learning into their research.
Across the academic health system, scientists are now using AI to improve our understanding of biological systems and modern medicine. University of Pennsylvania faculty members say the technology is being used in a variety of ways, from conducting translational research and analyzing data to optimizing health care delivery.
pattern detection and Risk prediction
In an interview with DP, Christos Davatsikos, professor of radiology and electrical systems engineering, explained how AI can be used to identify early signs of disease and inform preventive treatment plans.
His lab was one of the first AI-guided radiation therapy projects in the field, developing a method to use AI to analyze brain MRI scans and predict the future progression of tumors.
Davatsikos works on treating patients suffering from glioblastoma, a deadly form of brain cancer, and told DP that AI pattern recognition has led to “significantly increased survival times” for those affected.
He explained that AI is a “fundamental” research tool for detecting patterns, as the increasing number of biomarkers used to track diseases makes it “difficult” for a single person to visualize how the brain is constantly changing.
“We look at the MRI, but we don’t know where the tumor has invaded,” Dabatsikos says. “But AI can look at different MRI contrasts and create signatures that predict or detect very subtle changes that later cause tumor currents.”
Birkan Tunci, a professor of psychiatry who is also a researcher at Children’s Hospital of Philadelphia, told DP that he uses AI to study “almost every mental illness,” including depression, anxiety, ADHD, and autism.
Tunch described how the AI will “capture cues and signals” from the team’s video and audio recordings, allowing them to compare observations of individuals with autism and other conditions. He added that identifying these “signatures” can help with early diagnosis and can be “very beneficial to families.”
Brian Litt, a professor in the Department of Neurology at the Perelman School of Medicine, is working on a project that incorporates AI to test implantable brain devices that communicate with their hosts to report fluctuating levels of risk.
“Your medical device might text you and say, ‘What did you do? You just increased your chances of having a seizure by 60%,'” Litt told DP. “And you might have had a beer or taken a new antibiotic or something.”
Large-scale data analysis
Li Shen, professor of informatics in biostatistics and epidemiology, discussed one of his major research projects, how he integrated AI with machine learning and informatics techniques to analyze data and identify specific Alzheimer’s disease biomarkers.
“We want to identify not only genetic risk factors, but also protective factors, because that can be used to understand disease mechanisms and develop drugs to accelerate the discovery of treatments,” Shen told DP.
Li-san Wang, associate dean of computing at the School of Medicine, is co-director of the Center for Artificial Intelligence and Data Science for Integrated Diagnostics, which also leverages AI in Alzheimer’s disease research.
In a statement to DP, Wang wrote that in his research, AI helps integrate genetic knowledge with other types of biological data (such as RNA, protein, and epigenomic data) and could prove “essential” to understanding the biological mechanisms behind Alzheimer’s disease.
Greg Bowman, professor of bioengineering, biochemistry, and biophysics, spoke with DP about using AI in various aspects of his work.
Bowman explained how AI is being used during the data analysis stage to design peptides that bind to proteins that were previously not considered “viable targets.”
“Given the large amount of data we generate over time, it can be very helpful in finding patterns that are difficult to tease out visually,” Bowman says.
Streamline your research workflow
Davatzikos said that while AI is fundamental to his research, general improvements in computational and mathematical methods can also evolve the research process.
“The analysis of brain scans, for example, was an art through visual inspection,” Davatsikos told DP. “Now we have tools to accurately measure certain things.”
“My work was primarily about making imaging more quantitative and more scientifically grounded,” he said.
Shen said using AI to reduce “tedious computing time” had a “huge impact” on his research.
“You actually have all the knowledge rights available on the Internet and any knowledge base you have access to, so you can provide additional thoughts and new sales insights to your experts,” Shen said.
Litt told DP about the Pennsylvania Healthcare Transformation Innovation Center, which is using AI to help coordinate operating room scheduling across Pennsylvania hospitals.
“You have to be very smart about how you schedule your time in the operating room. Any downtime costs a lot of money,” Litt says. “We have Kevin Shea, a distinguished technical fellow at Penn Health, who has an algorithm that could potentially save $500,000 a month in operating room downtime by using AI to set the schedule.”
Wang echoed Litt’s opinion that AI could be beneficial in the day-to-day tasks of his job, such as organizing, annotating, and managing data, which he said remains a “big challenge in large-scale collaborative research efforts.” But he clarified that AI remains “resource-intensive and costly,” limiting its potential uses.
Tunci also acknowledged that current AI is a tool to “improve clinical workflows” and not a replacement for human labor.
“We never think of it as a replacement, like trying to replace a human construct, because even the concept of psychology is a human construct,” Tunch said. “We need people at the center.”
Staff reporter Addison Saji covers Penn Medicine. He can be reached at saji@thedp.com. I am studying English at Penn. Follow @addisonsaji on X.
Staff reporter Sameeksha Panda covers Penn Medicine. He can be reached at panda@thedp.com. I am studying chemistry at Penn University.
