LIBRA: Leeds Investigation of BReast screening AI
Research type
Research Study
Full title
A prospective study to assess the impact and benefits of an AI system in double reading for breast cancer screening
IRAS ID
316873
Contact name
Nisha Sharma
Contact email
Sponsor organisation
Kheiron Medical Technologies
ISRCTN Number
ISRCTN95571932
Duration of Study in the UK
0 years, 9 months, 8 days
Research summary
Breast cancer is the commonest cancer diagnosed worldwide. Breast cancer screening is considered to be the world-wide gold standard for early detection of breast cancer, with over 2.1 million women screened annually in the UK. Population-wide screening programmes like the one in the UK have large workloads as a huge number of mammograms (technically specific X-rays of the breast) are generated. Most hospitals simply cannot hire a sufficient number of radiologists to cope with workloads, resulting in overworked doctors and lowered job satisfaction.
Reading mammograms is one of the hardest cancer detection tasks. Interpreting clinicians must identify complex findings in the mammogram under intense time pressure. The interpreting clinicians are responsible to not miss evidence of cancer in the mammogram, while avoiding recalling women back unnecessarily. Given the circumstances, radiologists miss about every 5th (to 7th) cancer, while up to 90% of recalled patients may be called back unnecessarily.
This study assesses the medical device ‘Mia’, a software using artificial intelligence technology to interpret mammograms. In the study, eligible participants are randomised into one of two arms. All those randomised to Arm B receive standard of care breast screening, with their mammograms being ‘double read’ by at least two human radiologists. Participants in Arm A receive double reading plus a third read by Mia, and Mia’s opinion will be used alongside the two human reads when a clinician makes the final decision.
This research aims to assess how Mia affects arbitration, a process where cases are additionally reviewed when prior readers do not agree on the case to make a final decision. This study will also assess whether Mia can provide potential clinical benefit in breast screening, such as increasing the number of cancers found or reducing the number of women who need to be recalled for further tests.
REC name
South East Scotland REC 02
REC reference
22/SS/0077
Date of REC Opinion
1 Nov 2022
REC opinion
Further Information Favourable Opinion