Mia, an Artificial Intelligence (AI) tool, demonstrates its potential by identifying undetected cancerous indications in mammograms. It is known that early detection is crucial in treating all types of cancer, especially those with a high spread rate. Mia has the ability to identify tiny tumors that are often unnoticeable to the human eye but can potentially grow and spread at a fast pace.
In a test carried out by the NHS, Mia analysed over 10,000 mammogram scans, not only highlighting known cancer cases but also uncovering cancer in 11 women that had been missed by doctors. Microsoft, a participant in the Mia project, forecasts the technology could lessen radiography workloads by 30%.
Out of the 10,889 women involved in the trial, a minimal 81 opted not to have the AI review, indicating a majority trust in the technology. Dr Gerald Lip, project leader at NHS Grampian, pointed out that AI tools usually excel at identifying certain disease symptoms if they are trained with substantial data.
Evidence of Mia’s potential life-saving impacts has already emerged. Barbara, one patient reported by the BBC, benefitted from the precision of Mia’s AI, which detected her 6mm cancer early. As a result, she could receive a less invasive treatment with a higher success rate, making her appreciative of the tool.
The development of Mia has come with its own challenges. “It took six years to build and train Mia,” shared Sarah Kerruish, Chief Strategy Officer at Kheiron Medical, the company behind Mia.
The NHS has been vocal in its support for AI medical technologies, including a national deployment of AI-supported lung cancer detection that has shown up to 40 times greater accuracy than conventional methods. AI also has significant promise in breast cancer screening, and supports anti-cancer drug development.
Expressing optimism about Mia, Dr. Katharine Halliday of the Royal College of Radiologists said that the results underline the exciting potential AI presents for diagnostics. However, while Mia’s results were independently evaluated by the University of Aberdeen, they have yet to be peer-reviewed and documented in official research.
