The Sylvester Compleensive Cancer Center, part of Miami Miller University School of Medicine, will be leading newly funded multicenter clinical trials to assess whether artificial intelligence (AI) can help interpret mammograms more accurately, with the goal of improving breast cancer screening and reducing unnecessary callbacks for patients.
Known as the PRISM exam (practical randomized trial of artificial intelligence for screening mammography), the study is supported by a $16 million award from the Patient-Centered Results Institute (PCORI). This study includes hundreds of thousands of mammograms interpreted at academic medical centers and breast imaging facilities in California, Florida, Massachusetts, Washington and Wisconsin. The AI support tool investigated will be Transpara through Screenpoint Medical with clinical workflow integration provided by the AIDOC AIOS platform.
As the first major randomized trial of AI in breast cancer screening in the US, this study represents a pivotal step. Our goal is to assess the impact of AI rigorously and objectively. The radiologist continues to have solid control over all diagnostic decisions. ”
MD, Director of Breast Imaging Services at Jose Net, Sylvester, Associate Professor of Clinical Radiation at Miller School, Research Co-Investigator
Emergency questions about breast cancer treatment
Breast cancer continues to be one of the leading causes of cancer deaths in women in the US, with routine mammography screening reducing mortality through early detection. It also has drawbacks including unnecessary testing, anxiety and costs, and false positives that can lead to fugitives.
“AI offers a big promise, but also encourages important questions,” Net said. “Our focus is understanding whether AI truly enhances cancer detection by radiologists or simply leads to more misinformation.”
Patient-centered design
What makes prisms different is their focus on patient-centered research. The trial was developed in close partnerships with patient advocates, clinicians, health system leaders and policy makers.
Each participating facility will continue routine screening as usual, with no changes to patient experience. Mammograms are randomly assigned to be interpreted randomly by their own radiologist or with the assistance of FDA-cleared AI support tools. In all cases, the radiologist reads the exam and makes a final decision.
“A trial of this scale focusing on AI in breast cancer screening has never been conducted in the US,” Net said. “The findings affect not only clinical practices, but also insurance contracts, technology integration and patient interactions. NET shows that there is important optimism that AI will strengthen health care, but few rigorous studies have assessed the actual impact in real-world settings.
“This trial presents an opportunity to create credible evidence with an emphasis on the patient's perspective,” he said.
In addition to analyzing cancer detection and recall rates, this study includes focus groups and investigations to understand how patients and radiologists perceive and trust AI-assisted care.
Joint initiatives across six states
The Prism Trial brings together seven major academic medical centers.
- UCLA (Dual Princal Investigator, Management Coordination Site, Cooperative Investigator and Site PI, Dr. Hanna Millch)
- UC Davis (Dual Princal Investigator, Data Coordination Center led by Dr. Diana Miglioretti)
- Boston Medical Center (President Site Researcher, Dr. Claire B. Pointeon)
- UC San Diego (Site Principal Researcher, Dr. Haydee Ojeda-Fournier)
- Sylvester Comprehensive Cancer Center (Chief Site Researcher, Dr. Jose M. Net) is part of the University of Miami.
- University of Washington – Fred Hutchinson Cancer Center (President of Site, Dr. Johnny M. Lee)
- University of Wisconsin, Madison (Co-investigator, Dr. Christophe I. Lee; Former Site Investigator, Dr. My Elezaby, Dr. Ryan Woods)
Putting the right balance
The Net emphasized that the goal is not to replace human expertise, but to understand how AI complements it. “Our skilled radiologists have always had the final say. AI serves as a useful assistant, but ultimately they are the radiologists who control it.”
The trial is expected to inform future policy decisions, best screening practices, and how to most effectively integrate new technologies into patient care.
sauce:
Miami Mirror University School of Medicine
