Background This study investigated whether a quantitative faecal immunochemical test (FIT) could be used to select patients with either high- or low-risk symptoms of colorectal cancer for urgent investigation. Methods A double-blinded diagnostic accuracy study was conducted in 50 hospitals in England between October 2017 and December 2019. Patients were eligible for inclusion if they had been referred to secondary care with suspected colorectal cancer symptoms meeting national criteria for urgent referral and triaged to investigation with colonoscopy. Results The study included 9822 patients, of whom 7194 (73.2 per cent) had high-risk symptoms, 1994 (20.3 per cent) low-risk symptoms, and 634 (6.5 per cent) had other symptoms warranting urgent referral. In patients with high-risk symptoms, the sensitivity of FIT for colorectal cancer at cut-off values of 2 and 10 μg haemoglobin per g faeces was 97.7 (95 per cent c.i. 95.0 to 99.1) and 92.2 (88.2 to 95.2) per cent respectively, compared with 94.3 (84.3 to 98.8) and 86.8 (74.7 to 94.5) per cent in patients with low-risk symptoms at the same cut-off points. At cut-off values of 2, 10, and 150 μg/g, the positive predictive value for colorectal cancer was 8.9, 16.2, and 30.5 per cent respectively for those with high-risk symptoms, and 8.4, 16.9, and 35.5 per cent for those with low-risk symptoms. Conclusion FIT safely selects patients with high or low risk symptoms of colorectal cancer for investigation.
