Introduction
The Turing health and medical science programme has long-term interests in research study design and evaluation methodology for AI-driven healthcare solutions. It is seeking to contribute to the development of standards for how to evaluate AI in clinical studies within the NHS and internationally.
We are working in collaboration with Kheiron Medical Technologies to undertake a two-year research programme that will specifically focus on optimising such standards for the breast cancer screening domain.
The project will undertake research into novel clinical trial methodologies and evaluation criteria that are adapted to the demands of AI-driven healthcare solutions. We will also examine the opportunities and challenges of using real-world observational data to provide evidence for the utility of medical AI solutions prior to deployment within large-scale prospective clinical trials.
Background
Further information about the award:
This collaboration is supported by a Phase 4 AI Award to Kheiron as part of the £140m Artificial Intelligence in Health and Care Award Programme run by the Accelerated Access Collaborative (AAC) in partnership with NHSX and the National Institute for Health Research (NIHR). This is an initiative to support AI solutions from initial feasibility to evaluation within NHS and social care settings. The award is a competitive process run by the AAC as part of the NHS AI Lab with NHSX and the NIHR. Phase 4 is intended to identify AI technologies that need more evidence to merit large-scale commissioning or deployment. The AAC will work with NHS sites to support their adoption of these technologies, and stress test and evaluate them within routine clinical or operational pathways to determine their efficacy and accuracy, and clinical and economic impact.
The project team will be seeking to develop novel methodologies and studies to assess the safety and efficacy of Mammography Intelligent Assessment (Mia), a novel AI breast screening solution, developed by Kheiron Medical on over three million images from different mammography modalities and different ethnic backgrounds.