ChatGPT excels at picking the correct imaging test for patients with breast pain or needing breast cancer screening, researchers at Somerville, Mass.-based Mass General Brigham found.
The researchers asked OpenAI's ChatGPT 3.5 and 4 (its latest update) what imaging tests to order for 21 made-up patients. GPT-4 answered 98.4 percent of breast cancer screening prompts correctly, compared to 88.9 percent for GPT-3.5, according to the June 21 study in the Journal of the American College of Radiology.
"In this scenario, ChatGPT's abilities were impressive," said corresponding author Marc Succi, MD, associate chair of innovation and commercialization for radiology at Mass General Brigham and executive director of the MESH Incubator, in a June 22 news release. "I see it acting like a bridge between the referring healthcare professional and the expert radiologist — stepping in as a trained consultant to recommend the right imaging test at the point of care, without delay."
The researchers noted that before generative AI is used in medical decision-making, it will have to be tested for bias and privacy concerns, while regulations for healthcare AI are still being worked out.