A UK hospital aims to enhance breast cancer diagnostics and improve surgical outcomes using AI.
Researchers and surgeons are collaborating to implement a unique AI-enabled system at Nottingham City Hospital that combines optical coherent tomography (OCT) with Raman spectroscopy.
The Medical Research Council has awarded a research team from the University of Nottingham £1.8 million to lead a project in collaboration with the University of Kent and surgeons at Nottingham University Hospitals NHS Trust.
The aim is to install the combined instrument later this year.
Raman spectroscopy can detect molecular differences between normal tissue and cancer, but it requires hours to scan a whole specimen at high resolution.
OCT can measure structural images of tissue at high speeds but lacks molecular specificity.
AI algorithms specially developed for processing medical images will triage the OCT images before they are processed further in the system.
The combined instrument maximises the speed and resolution of OCT and the molecular specificity of Raman spectroscopy. This means that surgeons could analyse lumpectomy specimens while a patient is in an operating theatre, allowing for the potential excision of breast cancer in a single operation.
Further surgery is often linked to poorer patient outcomes, long recovery times, and higher healthcare costs.
Professor Ioan Notingher from the University of Nottingham’s School of Physics and Astronomy is leading the project in collaboration with Professor Emad Rakha, Professor Adrian Podoleanu, and Professor Philippe de Wilde from the University of Kent and Hazem Khout from Nottingham University Hospitals NHS Trust.
An initial OCT system designed by the University of Kent has already been set up at Nottingham City Hospital, which a. It collects images of breast lumpectomy specimens which were then processed by postgraduates using AI.
For the next phase of the project, this new OCT-Raman system is being assembled in Kent and will be established at Nottingham City Hospital, allowing further research and evaluation to begin.
Postgraduate researcher, Dr Radu Boitor, from the University of Nottingham, recently brought the Raman half of the system to Kent to be integrated with the OCT system and software assembled by postgraduate Kent researchers.
Once the combined system is ready and tested in Kent, it will be transported to Nottingham. This requires demo testing before characterising the system on various samples to perfect the software and make all processes more self-operating.
A tunable laser, a galvo scanner, and two large translation stages are under the OCT control designed in the Applied Optics Group. This enables the system to communicate with the Raman system, which has a sensitive camera and proprietary software developed at the University of Nottingham. The AI software functions as the central component of this set-up.
Professor Podoleanu, is leading the assembly of the OCT part in Kent
Professor Philippe De Wilde, supervising the AI algorithms, pointed out that AI algorithms for computer vision have great potential in new medical devices that can be produced in the UK, saying: ‘Our algorithms do not cost millions to train and run, such as GPT, but run on a single processor. This is both economical and environmentally sustainable.’
