Radiology and AI Converge to Advance Medical Imaging

Applied Radiology Publisher Kieran Anderson recently spoke with Sonia Gupta, MD, an abdominal radiologist who is the Senior Medical Director of Rad AI, a startup based in Berkeley, California. Dr. Gupta has expertise in artificial intelligence (AI), diagnostic radiology, image-guided procedures, digital health, regulatory requirements for FDA and CE approval, and go-to-market strategies for AI R&D. She’s also a member of the Applied Radiology editorial advisory board. This article is based on their conversation.

 

Radiology is undergoing rapid advancements requiring skills that go beyond simply reviewing images on a screen. Indeed, Sonia Gupta, MD, says radiologists are being tasked with learning new digital skills, much like they did when transitioning from reading scans posted on light boxes to using PACS.

Dr. Gupta is the Senior Medical Director of Rad AI, a company born at the intersection of AI and radiology that builds products to maximize productivity, make health care more accessible, and improve patient outcomes. With her expertise in abdominal radiology and AI, she mentors residents and fellows in AI, and serves as a new member of the Applied Radiology editorial advisory board.  “It's an extremely exciting time to be a radiologist,” she says. “AI is coming; we don’t have a choice about it. I think this is a good thing, and we’ll be able to deliver better care to our patients as a result of all these advancements,” she said.

Rising to the challenge of COVID-19

The COVID-19 pandemic is accelerating this transition, Dr Gupta said, particularly in ultrasound and abdominal imaging. Dr. Gupta practiced in Boston during the first wave of coronavirus, where her facility was able to quickly deploy a remote workforce to keep patients and staff safe. Technology and AI advances also allowed her facility to care for patients while implementing new disinfection protocols and minimizing patient waiting times.

“It was definitely a very challenging time, but I was impressed by the dedication of the radiology department. The ability to adapt to an evolving situation and to be nimble during a time that was unprecedented impressed me to no end,” she said. “Radiology as a field really rose to the challenge.”  Dr. Gupta says AI is now playing an important role in the diagnosis and management of COVID-19 patients. For example, an AI-based model is being used in China to help diagnose COVID-19 cases based on chest CT scans. She recently wrote about the role of AI in the COVID crisis for Applied Radiology.

“The AI model could determine if there was a high or low risk for COVID, and then the patient could be isolated. At the beginning of the pandemic, tests for COVID were slower, and a CT scan was faster combined with an AI model,” she said. “It was impressive how quickly they developed a product that could be used clinically in multiple hospitals.” In New York City, the Mount Sinai Health System is using a data mining application that reviews patient demographics such as age, race, and comorbidities (including hypertension and high cholesterol), and then merges that information with patients’ COVID diagnoses to track and predict outcomes.

On-demand connections critical during COVID-19 and beyond

At the all-virtual RSNA 2020 annual meeting, Dr. Gupta co-moderated a workshop on the job market, AI, and effective mentorship. She was also able to attend more sessions than she would have been able to in person. “That’s the benefit of a virtual show,” she said. “You have access to on-demand lectures. You can work a clinical shift and then log in and listen to a lecture.”

She believes that publications like Applied Radiology will play a critical role in developing the next generation of imaging professionals. Trainees need to see clinical images to learn how to identify normal and abnormal anatomy, but many journals don’t include high-quality graphics and cases, she said. “As a trainee, all you want is examples of what a case looks like. And Applied Radiology has always done a fantastic job of having high-quality images,” she said.

Moving from print to online communication is another way media can help prepare the next generation of clinicians for careers in radiology, through online images and cases, educational webinars, video sessions and online content, especially as more training and medical programs focus on virtual learning.  “I think Applied Radiology is able to adapt to those kinds of changing learning styles because our residents and medical students now have a completely different learning style,” Dr. Gupta said, noting that the publication continues to adapt to meet the needs of current and future medical imaging professionals.

“Adapting to that changing learning style is something that Applied Radiology is doing. And it resonates with everyone that content is available to them when they need it, however they like it. I think that's something unique about Applied Radiology, making that connection to your readership and making sure that the content is available to them, in the way they would like it.”

The convergence of AI and radiology

At Rad AI, Dr Gupta is focused on bringing a clinical focus to AI development. She still performs abdominal imaging, CT, MRI, ultrasound, and oncologic imaging to stay close to the practical applications of AI in clinical practice.  “If I can't imagine working with it or thinking through where it helps me to take care of patients, then it doesn't really make sense.

“There shouldn't be a fear of AI. When radiologists get involved in AI development, the focus changes. Instead of looking at the image aspect of it, we start to look at workflow and alleviating burnout,” she said. “There are so many opportunities for AI to help us so that our attention can be directed more towards patient care. And we can do a better job and improve our accuracy.  That’s where Rad AI can help develop technologies to help radiologists improve their reporting and support their diagnoses.

“Rad AI makes sure that you don't forget anything important for patient care because it reads your findings and then generates your impression for you. So you already have a working draft that you can quickly edit, and then you're done. And that requires no image interpretation,” said Dr. Gupta, who is excited to be a part of the Rad AI team while continuing her clinical practice.  I'm still passionate about being a radiologist and mentoring residents and fellows, and the way to do that is to continue working clinically. So for me, it's the best of both worlds and I'm just so happy about it.”

Back To Top

Radiology and AI Converge to Advance Medical Imaging.  Appl Radiol. 

By McKenna Bryant| January 21, 2021

About the Author

McKenna Bryant

McKenna Bryant

McKenna Bryant is a freelance healthcare writer based in Nashotah, WI.



Copyright © Anderson Publishing 2021