Artificial intelligence can "outperform" doctors in diagnosing breast cancer from mammograms.
According to a study in the journal Nature, an international team - which includes researchers from both Google Health and Imperial College London - has designed and trained a computer model on X-rays from almost 29,000 women.
It's said the AI was as good as a pair of doctors working together, while the algorithm outperformed six radiologists - and experts think it could help boost detection.
As it stands, the NHS uses two radiologists to look at each woman's X-rays, with a third doctor used to assess the images if they disagree.
For the study, the AI was given anonymised images - to ensure the women couldn't be identified - which meant AI only had the mammograms to go on, rather than having access to patient history.
The results show that the model matched the current double-reading system, and it was even better at spotting cancer than just one doctor.
Google Health's Dominic King said: "Our team is really proud of these research findings, which suggest that we are on our way to developing a tool that can help clinicians spot breast cancer with greater accuracy."
While X-ray reading is vital work, the process is time consuming, and the UK has an estimated shortage of 1,000 radiologists.