A new artificial intelligence tool is helping doctors fast-track breast cancer diagnosis for women most likely to have the disease, cutting typical wait times from several weeks down to roughly an hour. Researchers at UC San Francisco and UC Berkeley developed the system to triage patients with abnormal mammograms, allowing those at highest risk to move through imaging, evaluation, and sometimes even biopsy in a single day.

How Does AI Help Identify High-Risk Breast Cancer Patients?

The researchers used an open-source AI model called Mirai, developed by UC Berkeley data scientist Adam Yala, PhD. The system was trained on hundreds of thousands of mammograms linked to actual patient cancer outcomes, allowing it to recognize subtle patterns that predict cancer risk more accurately than a physician working alone.

When researchers applied the AI model to more than 4,100 screening mammograms at Zuckerberg San Francisco General Hospital and Trauma Center, it identified 525 women, about 12.7% of screened patients, as high risk. These patients could receive their mammogram interpretation immediately after imaging and have additional diagnostic imaging for any suspicious areas the same day.

"This is a really an exciting time. This moves us closer to personalized care, where we can tailor a plan so that each patient gets the right intervention at the right time," said Maggie Chung, MD, the first author of the study published in Nature Digital Medicine.

Maggie Chung, MD, UC San Francisco

What Are the Real-World Benefits for Patients?

The impact on patient experience is substantial. The AI model reduced the wait time for a diagnostic evaluation from several weeks to about an hour. For women who ultimately received a breast cancer diagnosis, the average wait for a biopsy dropped from more than two months to fewer than 10 days. This dramatic reduction in waiting periods addresses one of the most stressful aspects of cancer screening: the uncertainty and anxiety that comes with abnormal results.

It's important to understand that the AI system does not replace radiologists or make diagnoses on its own. Instead, it functions as a triage tool that helps physicians identify which patients would benefit most from accelerated care pathways.

How to Understand AI's Role in Breast Cancer Screening

• Risk Stratification: The AI model analyzes screening mammograms to predict which women have the highest likelihood of having breast cancer, allowing doctors to prioritize care for those most in need.
• Collaborative Tool: Rather than replacing radiologists, the AI works alongside physicians to enhance their decision-making and help them allocate resources more efficiently to high-risk patients.
• Personalized Screening: The system enables tailored screening schedules based on individual cancer risk rather than applying a one-size-fits-all approach to all women.

"This is a powerful example of how AI can be a collaborative partner for physicians. It shows how we can improve care when we bring clinicians and data scientists together to design these systems," explained Adam Yala, PhD.

Adam Yala, PhD, UC Berkeley

Before implementing the system in clinical practice, researchers first tested the AI model on more than 114,000 archival mammograms to ensure it would identify enough high-risk patients without overwhelming the clinic with excessive expedited evaluations. This careful validation step demonstrates the researchers' commitment to practical, sustainable implementation.

The broader vision behind this work addresses a fundamental challenge in breast cancer screening. Currently, many women follow the same screening schedule regardless of their individual risk profiles, which can vary dramatically from person to person. By using AI risk assessment, clinicians can identify women most likely to benefit from expedited care and provide them with the interventions they need at the right time.

"Right now, many women follow the same screening schedule, but their individual risk can be very different. AI risk assessment gives us the chance to identify the women most likely to benefit from expedited care and get them what they need," noted Maggie Chung, MD.

Maggie Chung, MD, UC San Francisco

The study, published May 19 in Nature Digital Medicine, represents a significant step toward more efficient and personalized breast cancer care. The research was funded in part by the National Institutes of Health, underscoring its importance to the broader medical research community. As this technology continues to develop and be tested in additional clinical settings, it has the potential to reduce anxiety for millions of women undergoing breast cancer screening while improving outcomes for those who do receive a diagnosis.