UC San Diego Health researchers are helping lead a groundbreaking national study exploring how artificial intelligence (AI) could transform breast cancer screening and improve the patient experience.

Led by UCLA and UC Davis, the PRISM Trial (Pragmatic Randomized Trial of Artificial Intelligence for Screening Mammography) is the first large-scale, randomized clinical trial in the U.S. to evaluate AI’s role in mammogram interpretation. Supported by a $16 million award from the Patient-Centered Outcomes Research Institute (PCORI), the multi-institutional study brings together seven leading academic medical centers, including UC San Diego Health, to determine whether AI can help radiologists detect breast cancer more accurately while reducing unnecessary callbacks and anxiety for patients.

Leading the effort at UC San Diego Health are Haydee Ojeda-Fournier, MD and Professor of Radiology and site principal investigator, and Tara Retson, MD and Assistant Professor of Radiology and co-principal investigator. Both are also faculty affiliates at the Jacobs Center for Health Innovation (JCHI).

“This is an unprecedented opportunity to understand how AI may complement expert breast radiologists,” Dr. Ojeda-Fournier says. ​“Despite the enthusiasm for AI, we truly don’t yet know if it works in real-life clinical settings. PRISM will help us find out.”

The study uses Transpara Breast AI by ScreenPoint Medical, integrated into clinical workflows through the Aidoc aiOS platform. While AI assists in image interpretation, every mammogram in the trial is still read and finalized by a radiologist, ensuring that patient care remains guided by clinical expertise.

“This is an unprecedented opportunity to understand how AI may complement expert breast radiologists.”

“The biggest challenge is the initial heavy lift,” Ojeda-Fournier says. ​“Integrating AI tools like Transpara and Aidoc into existing workflows requires significant coordination across IT, radiologists, and operations to ensure data security, workflow compatibility, and user adoption. The biggest opportunity, though, is tremendous. Once the infrastructure is in place, these tools can enable automation and allow seamless data collection for research.” 

UC San Diego Health’s radiologists have already experience deploying multiple AI systems in real-world clinical environments, and the PRISM Trial will help them evaluate how these tools work and how they impact the physician workflow.

“There has been a lot of speculation and hype over the last few years, but we’re just starting to test it out for ourselves in the clinic,” Retson says. ​”We know from prior studies that when a strong AI algorithm is combined with a radiologist, the two have the potential to perform better as a team than either alone. There’s growing thought that computers and humans may pick up on different features in images, and that we compliment each other. If we can reduce unnecessary callbacks or make more confident interpretations, that has a direct impact on improving patient care and reducing anxiety, especially during what can already be a stressful experience.”

For both Ojeda-Fournier and Retson, collaboration is key in understanding how AI can support breast cancer screening,

“It truly takes a village,” Ojeda-Fournier says. ​“We’re grateful to our radiology department, hospital leadership, IT teams, and most importantly, to the women who will participate. Their involvement will provide the evidence needed to shape the future of breast cancer screening.”