1Croydon University Hospital. 2Basingstoke and North Hampshire Hospitals. 3Royal Marsden Partners. 4Royal Surrey County Hospital Background:

Bowel symptoms have low specificity for colorectal cancer. The faecal immunochemical test (FIT), can measure occult blood in stool and could exclude bowel cancer without need for a colonoscopy. This study investigated the diagnostic accuracy of FIT in symptomatic patients.

Methods:

This was a powered, multicenter, ethics approved, double-blinded NIHR-badged diagnostic accuracy study of FIT for colorectal cancer or other findings at colonoscopy. Patients were eligible for recruitment if they bowel symptoms meeting NICE referral criteria, and had been triaged to investigation with colonoscopy. Patients were excluded if they did not provide an analyzable FIT, or did not undergo a complete colonoscopy. External statisticans analysed FIT diagnostic accuracy at different faecal haemoglobin (FHb) cut-offs for referral.

Findings:

9822 patients from 50 sites in England participated in the study. The diagnoses found at colonoscopy included no disease (31%), colorectal cancer (3.3%), HRA (4.2%) and IBD (6.1%). If patients were only referred for investigation if their FHb was above cut-offs of 2, 10 and 150 ug/g, the reduction in referrals would be 63%, 80% and 92% respectively. The cancer detection rate at cut-offs of 2, 10 and 150 ug/g significantly decreased from 98% to 90% and 75% respectively (p < 0.01). The PPV of FIT at cut-offs of 2, 10 and 150 ug/g was 8%, 16% and 31% respectively.

Interpretation:

FIT is a diagnostic tool that can guide referral of patients with suspected colorectal cancer with greater positive predictive value than symptoms.