FORE AI

The Future of Real Time Endoscopy Artificial Intelligence study aims to demonstrate the benefits of using AI in colonoscopy.

Background

Colonoscopy is an important tool for the identification and removal of pre-cancerous and cancerous polyps. However, the quality of the colonoscopy is highly dependent on the doctor/nurse and studies show significant miss rates (up to 25%) for polyps and cancers. Polyps that are found and removed are sent to histopathology for diagnosis which adds, on average, 3 weeks to patient waiting times. Odin Medical has developed a Computer Assisted Detection and Diagnosis system for colonoscopy.

Computer Assisted Detection and Diagnosis

The system acts like a second pair of eyes during the procedure using artificial intelligence (AI) to identify and diagnose polyps in images. It can support the doctor to find more polyps and provides instant diagnosis. This is strategically aligned to the NHS long term plan of saving 55,000 more lives a year by diagnosing more cancers earlier. The innovative AI solution deploys low cost, vendor agnostic, hardware into hospitals and uses secure cloud computing to process data. This makes it cost effective and accessible for all hospitals, small and large. It ensures we are developing AI in accordance with the first principal of the NHS constitution: a comprehensive service, available to all.

Study aims

The aim of FORE AI is to demonstrate the benefits of using AI in colonoscopy through a multi-centre randomised controlled trial. These benefits include, better patient outcomes by improving polyp/cancer detection rates, improved patient experience with instant diagnosis and increased operational efficiency by foregoing histopathology. The project will gather efficacy data and perform health economic analyses/cost saving evidence to support the adoption of AI in endoscopy. The project has three phases: 1) pre-trial, 2) trial and 3) post-trial. In the pre-trial phase the project consortium will prepare the paperwork and ethics, create training literature, develop software to capture trial data, perform site initiation visits and develop the healthcare economic model.

Study approach

During the study, the data will be captured and quality, regulatory and efficacy evidence will be collected. The AI will be evaluated, and models retrained to improve performance.

After the trial the data will be analysed to build the healthcare economic case to demonstrate cost effectiveness. A public and patient group has been involved throughout the technical development of the AI system and will continue to be involved in the trial evaluation through steering committee and focus groups.

Study outcomes

Study outcomes include:

• Clinical Evaluation Report including evidence of clinical safety and efficacy data
• Big datasets for AI evaluation
• Healthcare economic model
• Clinical publications.