Credit: Rui Vieira/PA
Using artificial intelligence to screen for breast cancer could trigger a surge in false diagnoses, technology experts have warned.
Scientists need to be "honest about the immaturity of most AI tools", the BCS – the UK’s chartered institute for IT – said.
Dr Philip Scott, chair of the BCS health and care executive, said: "While AI methods are promising, there is not yet enough scientific evidence to justify adoption in a cancer screening programme.
"Unfortunately, there is so much hype about AI that some people treat it like magic. Most AI in healthcare is early stage and not shown to work clinically. If you look at the scientific reviews, the experiments done with AI diagnostic tools are simply not good enough. Many studies are at risk of bias from selective use of patient data.
"If AI were adopted now in the screening of breast cancer, there is significant risk of overdiagnosis with all the anxiety that would cause."
Scientists are currently studying how the technology could be integrated into NHS systems, following earlier studies which found that AI was able to identify cancers from images "with a similar degree of accuracy to expert radiologists".
Earlier this summer, the Government funded a trial, run by researchers from Imperial College and Google Health, to assess how radiologists and clinicians would interact with the AI system – part of a push to start using more of the technology within the healthcare system.
Dr Scott made the comments in a formal response to a review by government advisory group the UK National Screening Committee, which has said that the current evidence remains "a long way from the quality and quantity required for implementation of AI into clinical practice of breast screening programmes".
In the UK, one in eight women will be diagnosed with breast cancer in their lifetime, with survival rates improving if the cancer is detected and treated earlier. However, doctors can find it difficult to accurately detect and diagnose breast cancer.
Scientists from the government-funded trial said AI tools should not replace clinicians, but instead "could help to reduce the current burden on radiologists and the NHS, improving outcomes for patients through earlier detection and treatment of cancer".
Lord Ara Darzi of Denham, chief investigator on the project, said in June: "Our early work in this area has shown that using algorithms to screen for breast cancer is feasible. This next step will be our first real life test of AI as part of a national screening programme."