How does the AI product work?
Mia has been designed to be the second reader in the workflow, so there is always a human in the loop. Should Mia and the first human reader disagree, a second human gets the final call, meaning radiologists will always have authority over the AI. Using AI technology for the second screening frees up clinicians to spend time with patients, reduces the pressure to find more radiologists and has the potential to screen greater numbers of women more quickly.
Developing AI-driven technologies products like Mia to provide accurate assessments for any patient, means using a huge amount of data. Mia has been built using convoluted neural networks, a form of deep learning that has involved being trained on over 3 million images from multiple sites. Using funding from the AI in Health and Care Award, Mia will go through multiple deployments and clinical studies across 15 hospital sites in the UK to ensure it is safe, effective and will benefit breast screening units, and the women who attend them.