A new artificial intelligence (AI) tool was tested by the UK National Health Service (NHS) and it showed promise in picking up signs of breast cancer missed by doctors in mammograms. Radiologists, oncologists and pathologists could benefit from this as it will reduce their workloads hence making them more productive.
During a recent trial at NHS Grampian Aberdeen in Scotland, Mammography Intelligent Assessment (MIA), an AI platform, detected early signs of breast cancer among 11 women out of 10,000 mammograms –– signals so small that human physicians had not seen them before.
Kheiron Medical Technologies and Imperial College London developed the MIA platform. It could greatly help doctors in the UK who have heavy workloads if the AI can identify these early signs of cancers.
The AI was trained using millions of mammograms from around the world. In a European healthcare setting, the AI increased early cancer detection by up to 13% during a study carried out at Imperial College London where 24 more cancers were discovered compared to standard readings by human doctors.
The AI platform can discern subtle patterns and irregularities within mammograms even when carcinomas are still at their earliest stages. A pilot study saw MIA look through mammograms taken from over 10,000 women thereby finding cancer among 11 females that were overlooked by physicians.
Physicians sometimes fail to notice signs of cancer because they may be tired or distracted. Unlike humans, AI does not get fatigued or diverted thus making it valuable for supporting doctors. The MIA platform acts like a safety net which ensures any incipient cancers are not overlooked.
Further research needs to be done before deploying MIA widely across clinical settings but current studies indicate that such artificial intelligence tools can greatly enhance breast cancer screening as well as other medical diagnostics and patient care areas.
Source: darkdaily
