DeepHealth, an AI company owned by the Radiology group RadNet, presented its research on breast cancer screening using AI software at the annual meeting of the Radiological Society of North America (RSNA).
The study followed 747,604 women who underwent screening over a 12-month period. The results showed that the cancer detection rate in the AI-assisted group was 43% higher than in the group that did not use the technology.
Among these, some women have chosen to use AI software that meets the standards of the US Food and Drug Administration (FDA).
The AI acts as a “second set of eyes,” helping radiologists spot abnormalities on scans. The researchers found that the 22% increase in overall cancer detection was due to the fact that higher-risk patients were more likely to sign up for the program in the first place.
However, the remaining 21% were identified with direct support from AI technology, which increased the rate of image "recall" for more thorough examination of suspected cases.
However, experts say more randomized controlled trials are needed to assess the real-world accuracy of AI in breast cancer screening and detection, so that the research can be highly reliable when deployed on a large scale.
