New AI Tool Detects Up To 13% More Breast Cancers than Humans Alone – Radiography

Image: The AI tool, called Mia, could significantly increase early detection of breast cancers (Photo courtesy of Imperial College London)

In 2020, the World Health Organization reported 2.3 million new breast cancer cases globally, with 685,000 fatalities. Early detection through screening is crucial in reducing the intensity of treatment and mortality rates. Yet, about 20% of breast cancers are potentially missed during early screening stages. A groundbreaking artificial intelligence (AI) tool now promises to significantly enhance the early detection of breast cancers.

Developed by Kheiron Medical Technologies (London, UK) in collaboration with Imperial College London (London, UK), the AI tool named Mia has demonstrated its ability to detect up to 13% more breast cancers than human screeners. Mia, holding a CE Mark class IIa, has shown promise in enhancing breast cancer screenings by identifying cancerous tissues that may be overlooked by human radiologists. In Hungary, where the study was conducted and mammograms undergo double-reading by two radiologists, Mia was introduced as an additional ‘reader’.

During the research, spanning April 2021 to January 2023, Mia was tested on mammograms from 25,065 women across four Hungarian screening sites. After initial evaluation by two radiologists, the mammograms were analyzed by Mia, which highlighted potential false negatives for a third human evaluator to review. This process identified mammograms initially cleared by human readers but showing subtle indications of cancer through Mathew’s study, divided into three phases (two pilot phases and a live roll-out), revealed that Mia detected 24 more cancers than the standard human readings, marking a 7% relative increase. This led to 70 additional recalls (0.28% relative increase). In the three phases, Mia identified an additional 6, 13, and 11 cancers respectively, enhancing the relative cancer detection rate by 13%, 10%, and 5%.

Significantly, 83% of the extra cancers detected through Mia in real clinical settings were invasive, underlining Mia’s effectiveness in identifying cancers that greatly benefit from early detection. The researchers have emphasized the need to validate these findings in other nations with different screening approaches and populations. They have also noted that the follow-up period of two to nine months requires an extension to fully evaluate Mia’s impact on cancer detection and mortality reduction. Concurrently, a comprehensive study in the UK is nearing completion to affirm the benefits of the AI-assisted reader workflow, with plans to initiate the roll-out in the US.

“Our prospective real-world usage data in Hungary provides evidence for a significant, measurable increase of early breast cancer detection when Mia is used in clinical practice. The key question now is how we can justify not using Mia in breast screening when there is such a dramatic improvement in cancer detection,” said Dr. Peter Kecskemethy, CEO of Kheiron.

“These results have exceeded our expectations. Our study shows that using AI can act as an effective safety net – a tool to prevent subtler signs of cancer falling through the cracks. Seeing first hand that the use of AI could substantially reduce the rate of missed cancers in breast screening is massive, and a major boost for our mission to transform cancer care with AI technology,” said Dr. Ben Glocker from Imperial’s Department of Computing, who leads machine learning research teams at Imperial and Kheiron.

Related Links:

Kheiron Medical Technologies

Imperial College London