In Britain, AI has helped detect breast cancer cases missed by doctors

In testing in the UK's National Health Service (NHS), an artificial intelligence tool called Mia has successfully identified signs of breast cancer in 11 women that were initially missed by doctors when analysing mammograms.

Here's What We Know

When examining over 10,000 breast scans, Mia correctly identified all of the cancers detected by medics, and additionally pointed out 11 patients with tiny tumours measuring around 6mm. At such an early stage, cancer is often difficult to spot with the naked eye.

One of these 11 women is Barbara, whose tumour was detected by AI. Thanks to a timely diagnosis, she only needed minimal treatment. Statistically, a tumour less than 15mm in size has a 90% chance of survival within five years.

According to Mia developer Kheiron, such AI tools can reduce the waiting time for mammography results from 2 weeks to 3 days. Eventually, one of the two radiologists who review each scan could be replaced by AI to reduce the workload of both doctors.

Gerald Lip, clinical director of screening at the NHS, said that out of nearly 11,000 participants in the study, only 81 women refused to have their scans analysed by artificial intelligence.

Mia, which runs on Microsoft cloud computing, took six years to build. It was trained on millions of anonymised mammograms from around the world. The creators emphasise the importance of data inclusivity when developing medical AI systems.

Experts welcome such technologies as a potential aid to making doctors' jobs easier, but say further research is needed to determine how best to use them.

Source: BBC