That’s important, UT Dallas said, because breast cancer rates, particularly among younger women, have been rising since 2012 while the number of radiologists in the U.S. has continued to decline. Professors Mehmet Ayvaci and Radha Mookerjee, along with colleagues from other institutions, found that AI could reduce healthcare costs by up to 30 percent when compared to relying solely on human diagnosis.

However, Ayvaci said, removing humans from the diagnostic process entirely isn’t advisable because AI is still not as accurate as a radiologist.

“The workflow and most cost-effective strategy are the question,” Ayvaci said. “I’m talking about how to replace a task, not replace decision-making. Could we design a workflow where AI is playing a triage role?”

Both researchers are faculty members in the Naveen Jindal School of Management. Their findings were published in Nature Communications.

Studying mammography images with AI
The team built its models using real mammography images from the Digital Mammography DREAM Challenge, a publicly available dataset of more than 640,000 anonymous images. Gustavo Stolovitzky, a co-author of the study, is founder and chair emeritus of the DREAM Challenges, a crowdsourced initiative that applies collaborative data science to medical research.

“Our approach enabled us to precisely identify which patients could safely be evaluated using AI alone while determining which ones would need to be referred to a human,” Mookerjee said in a statement.

“Every patient’s data, including the mammography images, is input to an AI algorithm, which outputs a measure of risk,” he added. “You can think of this measure as a probability that the patient has breast cancer.”

That risk value is used to determine whether the patient should be referred to a human, Mookerjee says.

According to UT Dallas, integrating AI into mammogram screenings could help reduce false positives, which often lead to unnecessary procedures such as biopsies. A false negative from AI, however, raises important concerns.

“If AI makes a wrong decision, who should be held accountable? All of these are open questions,” Ayvaci said.

‘Brighter future’ of AI systems
Ayvaci said AI’s effectiveness in diagnostic settings hasn’t been fully tested in real-world healthcare environments. He is also researching how AI could reduce administrative burdens and improve cost efficiency, especially in rural clinics or underserved areas.

“The new generation of AI systems are capable of providing context and improving in a much faster timeline,” Ayvaci said. “All of this contributes to a brighter future of using these AI systems.”

Mookerjee, whose mother is a breast cancer survivor, said she was drawn to the project by the opportunity to apply AI to healthcare challenges.

“When Dr. Ayvaci told me about his idea to use AI in screening mammography images, it seemed like the perfect merger of these two areas,” she said.

Ayvaci and co-author Mehmet Eren Ahsen of the University of Illinois Urbana-Champaign are now exploring how this AI approach could help speed up patient care by supporting same-day procedures or faster follow-ups after diagnosis.

Source : UT Dallas Study Tests AI as a Triage Tool in Breast Cancer Screenings » Dallas Innovates