AI could transform breast cancer screening

The use of artificial intelligence in breast cancer screening is very promising, a new Swedish study suggests.

In a trial screening more than 80,000 Swedish women, researchers at Lund University determined that AI-driven tools could detect breast cancer at a “similar rate” to two radiologists, the BBC reported.

Researchers said AI’s potential cancer-screening abilities could address the shortage of radiologists and ease their workloads, while also cutting patients’ waiting times.

We’ve curated reporting and insights on the role AI could play in healthcare in the years to come.

• Physicians in the U.S. are burning out, and a shortage of up to 124,000 doctors is expected by 2034. David Chou, a health tech executive, wrote in Forbes that generative AI could reduce the workload of some physicians and help burnout. But AI is largely being used to “automate administrative tasks,” Chou wrote, as the industry is still skeptical about adopting AI for clinical decisions.
• Biotech company Moderna has turned to AI to identify areas with low vaccine uptake, and uses the data to roll out targeted messaging and campaigns to those communities. The company is also using the tech to hasten the development of new vaccines. AI “was what allowed us to figure out what mutations we needed to go after and expedite [the COVID-19 vaccine] to market,” Kate Cronin, chief brand officer at Moderna, said in June. — Euronews
• The Australian Medical Association has called for stronger oversight after revelations that doctors were using ChatGPT, OpenAI’s large-language model (LLM), to issue medical notes. Since there is no guarantee of patient privacy if medical concerns are shared with the LLM, the AMA asked doctors to halt the practice. The issue points to broader concerns over AI oversight. In a statement at the time, AMA president Steve Robson said, “We need to address the AI regulation gap ... especially in healthcare where there is the potential for patient injury from system errors, systemic bias embedded in algorithms and increased risk to patient privacy.” — The Guardian

In addition to breast cancer screenings, scientists are hopeful that AI will have broad applications across healthcare.

Studies in the U.K. have found that AI was able to accurately diagnose people with heart failure, and has predicted whether patients with a degenerative disease would experience disease progression. It may also have applications in managing overflowing hospital waiting rooms: One AI model was able to predict 8 in 10 times whether someone who came to the ER was in need of emergency treatment.