From the first X-rays to AI

Soon after arriving at Penn last year as the new chair of Radiology, Pari Pandharipande, MD, MPH, met the university archivist who had come to present two historic X-ray plates created at Penn that were newly returned to campus. Dating from 1890—before the first journal description of this radical new technology—the plates offered a glimpse at the beginning of the emerging discipline of radiology.

The early X-ray plates were a vivid reminder of the department’s history of innovation. That’s just one of the reasons Pandharipande already considered her new role a once-in-a-lifetime opportunity. “When you consider all our imaging modality subspecialties—X-ray, CT, MRI, PET—it's impossible to find one that hasn't been heavily influenced by one or more innovations here at Penn,” she said.

Today, medical imaging and radiology have reached another inflection point—and Pandharipande has a plan for how Penn Medicine will remain at the forefront, putting people first.

The field of radiology has become a primary test case for artificial intelligence (AI) use and integration in health care. Pandharipande is careful to position herself as neither an AI adversary nor an evangelist. “I think that you need to be both enthusiastic and a skeptic,” she said.

Currently, AI has been most thoroughly explored in the realm of image interpolation and interpretation. Pandharipande compares current advances in imaging AI to the “ChatGPT moment” of 2022-2023 when language models made a quantum leap forward. Yet Pandharipande does not believe AI will “take over” radiology as some pundits predict. Rather, she sees it as an opportunity to rethink the way radiologists function, such as easing the burden of their heavy workloads, or helping them formulate increasingly sophisticated interpretations and scientific questions.

“Down the line, a good AI model should be able to be at least as accurate as a radiologist for some routine tasks and more, but we have to figure out what is the irreplaceably human part of the equation,” she said. “To me, that’s decision-making at the patient level, because the patient is much more than a collection of data points. We’ll do best if we keep humans at the center of how we think about AI.”

Bringing it all together

Born into a family steeped in academia, Pandharipande was the first to pursue a medical career. After a childhood spent “obsessed” with basketball and other sports, she knew she wanted a career in medicine and was delighted to find radiology: “Looking at images was something I was excited about and came relatively easily to me. When I see an image, I remember it.”

Her original career plans never specifically involved becoming a department chair. “I wanted to be a good doctor, a good researcher, a good teacher to our residents—those were my priorities,” she said. She earned undergraduate and medical degrees at Cornell University and completed a residency at New York University Medical Center, followed by fellowships in cancer outcomes research and in abdominal imaging at Massachusetts General Hospital. She remained at Mass General for 17 years as a physician-scientist in increasingly senior clinical roles.

“But being an outcomes researcher and getting pulled into the enterprise integration work at Mass General led me to realize I could leverage my skills into something different,” she said.

Serving as executive director of Clinical Enterprise Integration for Mass General Brigham Radiology, Pandharipande combined work in patient care, research, and education, with unifying operations across a complex hospital system. She could bring together these different skillsets to shape the way people work together and improve outcomes for patients at the same time.

Next, as chair of Radiology at The Ohio State University College of Medicine and Wexner Medical Center, beginning in 2021, she oversaw the growth of her department’s imaging infrastructure, patient volumes, and service line. It was a good fit, and she envisioned finishing out her career there. Then Penn Medicine came calling, and the allure of the world’s first academic department of radiology proved too much to resist. She joined in June 2025.

A plan to shape the future of Penn Radiology

Today, Pandharipande is bringing to life a five-point departmental vision for Penn Radiology, with each point carefully interconnected to the others. As vital as she considers AI to radiology’s future, Pandharipande’s plan begins with, and always returns to, people.

 “The thing I love best is bringing teams together and watching people come into their own and do something that we never could have imagined,” Pandharipande said. “And Penn’s culture is perfect for that.”

The first point of her plan prioritizes activities with clear patient impact and improving the day-to-day experience of those working in radiology. Pandharipande expects AI to help streamline operations, shorten scan times, and optimize scanner selection—all areas of efficiency that can address some of the current strains in the field, which include long wait times, a shrinking workforce, and rising demand as an aging population experiences more complex medical issues.

For another key point of her plan, to advance research and innovation, Pandharipande emphasizes team science and high-risk, high-reward projects, noting Penn’s unusually collaborative culture, with basic and translational scientists working together, as key strengths.

“When you think about what it means to be patient-centered, I'm trying to extend that idea to think about making investments in science that will have profound impact,” she said.

Her plan also emphasizes complementary aims of sustaining Penn’s status as a leader in medical education, as well as supporting financial stability and the growth of radiology services. “When we do well, we are able to reinvest in the academic mission.”

Right place, right image, right care

Pandharipande’s experience in integrating operations of a large health care enterprise is central to one of her biggest aims: reimagining care delivery. Few of the basics have changed since her own training, she said. “Now AI and other tools give us the chance to deeply interrogate what we do and to make sure every patient has the right imaging test done at the right time and the right place.”

Under previous chair Mitchell Schnall, MD, PhD, Penn Medicine had already been working towards thinking about Penn Radiology system-wide as more than a collection individual hospitals and divisions, said S. William Stavropoulos, MD, executive vice chair of Radiology. It’s an initiative called One Penn Radiology, aiming to unify the radiology operations of Penn Medicine’s seven-hospital system to drive a consistent, excellent patient experience at all locations.

Pandharipande “has turbocharged that,” he said.

So far Pandharipande has achieved a respect for local hospital cultures in balance with standardizing systems and tools, Stravropoulos said. He gave the example of a hypothetical Lancaster-based patient involved in a clinical trial that requires a specific MRI protocol previously unavailable there. “Being able to stay on the trial and get that care locally is life-changing for that person.”

And not just for one person. The ultimate vision is for Penn Medicine radiologists to remain at the forefront of innovation in imaging technologies, while also providing this advanced care to patients in local communities, Pandharipande said. “This, to me, is what putting patients first means.”