Breast screening technology developed using artificial intelligence can detect abnormalities that would have been missed using current screening procedures.
From research to delivery in clinical settings, this groundbreaking work undertaken by the University of Aberdeen, NHS Grampian and Kheiron Medical Technologies has included the analysis of 220 thousand mammograms from more than 55 thousand people in determining how well an AI tool could detect breast cancers.
The AI software, named ‘Mia’, created by Kheiron Medical Technologies, was assessed by a research team led by Professor Lesley Anderson, chair in Health Data Science at the University of Aberdeen, as part of the Industrial Centre for Artificial Intelligence Research in Digital Diagnostics (iCAIRD) programme. Dr Clarisse de Vries, a radiology imaging researcher at the University of Aberdeen, led the data analysis.
The analysis found that Mia successfully identified potentially missed, known as ‘interval’ cancers, detected between screening visits. The team found that Mia would have suggested recalling 34.1% of the women who went on to develop cancer in between screenings. Using current screening measures, these cancers remained undetected until the women developed symptoms.
Dr De Vries explains: “Currently, two experts examine each mammogram and decide whether the person should be invited back for additional investigations. If the two experts disagree, a third expert makes the final decision.
“Like a human expert, Mia can examine a screening mammogram and give an opinion as to whether that person should be invited back for additional investigations.
“Mia has previously been developed and tested on some groups, but until now had not yet been used on data from the NHS in Scotland.
“Our finding is a massive step forward in using AI technology in diagnostic medicine – we showed that once ‘tuned’ to the local environment, AI can greatly benefit clinicians and, importantly, people at risk of developing cancer. ”
Professor Anderson highlights: “The research also identified that ongoing quality assurance monitoring is essential due to changes in the performance of Mia with updates to the mammography machine software.”
Consultant radiologist Dr Gerald Lip has been instrumental in embedding Mia in clinical settings: “When originally applied to the data from NHS Grampian, Mia was too sensitive – recommending the recall of women for further investigation when it may not be necessary. However, Mia’s performance markedly improved when adjusted to suit the local conditions and technology, recalling a minimum number of women possible while maintaining a high cancer detection rate.
“The GEMINI (Grampian’s Evaluation of Mia in an Innovative National breast screening Initiative) project was born as an innovative collective to understand the impact of introducing AI into the breast screening programme within NHS Grampian. GEMINI combines all the strands of research, software, the tech industry and the patient’s healthcare outcomes.”
Dr De Vries explains: “Our results show that AI, and in this case Mia, offers huge potential for detecting cancers that may otherwise be missed.
“Fundamentally, however, our study shows that AI tools must be tested first and tuned for the local population and conditions, and we have been fortunate to have been able to do just that here in Grampian.
“Previously, it was unclear whether AI tools developed elsewhere could be used in different settings and screening centres. Now we know there are risks in just taking an AI tool developed elsewhere and implementing it locally. You must first test the tool on the local data to ensure it will work as expected.”
Kheiron’s team is delighted that the software program is now being evaluated in Grampian’s Aberdeen Royal Infirmary. Peter Kecskemethy, CEO of Kheiron, said: “Working with NHS Grampian and the University of Aberdeen teams has been an immense privilege. They are the UK pioneers of how AI technology should be evaluated and implemented in radiology, and we could not be more excited about collaborating on the next stage of the journey, which is assessing how best to use Mia in a live clinical setting.”
Professor Roger Staff, head of imaging Physics at NHS Grampian, added: “This is a critical study, identifying the steps required to get this technology into service. Although the results indicate that the technology is not quite ‘plug and play,’ it has the potential for major health and operational gains for the service.”
The three-year-long project was funded as part of the Artificial Intelligence Research in Digital Diagnostics (iCAIRD) by Innovate UK on behalf of UK Research and Innovation (UKRI).