Radiologists can benefit from AI support
Amelie Benoist/BSIP/Universal Images Group via Getty
Artificial intelligence models could help detect cancer and reduce doctors’ workloads, according to the largest study of its kind. Radiologists who chose to use AI were able to identify one in 1000 cases of breast cancer.
Alexander Katalinic of the University of Lübeck, Germany, and his colleagues worked with nearly 200 certified radiologists to test an AI trained to identify signs of breast cancer in mammograms. Radiologists examined 461,818 women at 12 breast cancer screening sites in Germany between July 2021 and February 2023, and each person could choose to use AI or not. As a result, 260,739 were checked by AI plus a radiologist, while the remaining 201,079 patients were checked by only one radiologist.
Those who chose to use AI detected breast cancer at a rate of 6.7 instances per 1,000 scans – 17.6 percent higher than 5.7 per 1,000 scans among those who chose not to use AI. Also, when women underwent biopsies after a suspected cancer diagnosis, those diagnosed with AI were 64.5% more likely to have cancer cells found in a biopsy, compared to 59.2% of women who did not use AI.
The scale at which AI improved breast cancer detection “was very positive and exceeded our expectations,” Katalinic said in a statement. “We can now demonstrate that AI significantly improves cancer detection rates in breast cancer screening.”
“The goal was to show non-inferiority,” says Stefan Bunk of Vara, an AI company also involved in the study. “If we can show AI less than radiologists, that’s an interesting workload-saving scenario. We were surprised that we were able to show superiority.”
There has been an overreliance on AI in medicine some were worried because of the risk, it could miss some signs of a condition, or it could lead to a dual system of treatment, where those who can afford it are afforded the luxury of human interaction. There was some evidence that radiologists spent less time reviewing scans that the AI already suggested were “normal,” meaning cancer was not likely, averaging 16 seconds, compared to 30 seconds the AI could not. do not classify But these latest findings have been welcomed by those specializing in the safe deployment of AI in medicine.
“The study provides further evidence of the benefits of AI in breast screening and should be another call for policymakers to accelerate the adoption of AI,” he says. Ben Glocker at Imperial College London. “Its results confirm what we’ve been seeing time and time again: with the right integration strategy, the use of AI is safe and effective.”
The study helps radiologists make decisions about whether to use AI, and would like to see more AI trials conducted in a similar fashion. “We can’t easily assess this in the lab or through simulations and instead have to learn from real-world experience,” says Glocker. “The technology is ready; we need policies to follow now.’
Topics:
