Automating faecal immunochemical test result interpretation using artificial intelligence and machine learning

"Bowel cancer screening services invite asymptomatic persons aged \>60 years to conduct guaiac-based Faecal Occult Blood testing (gFOBt) every two years. From Autumn 2018, NHS England's Bowel Cancer Screening Programme (BCSP; commissioned by Public Health England; PHE) will roll-out Faecal Immunochemical Test (FIT) to replace gFOBt. FIT is simpler, more sensitive, and improves neoplasia detection However, FIT is less specific than gFOBT, increasing the number of 'false positive' patients being referred for colonoscopy.

FIT results will be interpreted manually by a Bowel Cancer Specialist Screening Practitioner. To improve the overall efficiency of the BCSP, and to optimise the interpretation of the FIT result, Advanced Expert Systems Ltd (AES), experts in Artificial Intelligence (AI) and Machine Learning (ML) software, propose a feasibility assessment of a new system (using similar principles to an AI-based system they have developed for the MoD) that

1. Automates the interpretation of FIT-results, considering them within a framework of other data variables within the electronic patient record;
2. Automates the wider BCSP as much as possible, from invitation of patients to delivery of results and onward forwarding of patients, where required."