Can Fit Rule Out Colorectal Cancer In Symptomatic Patients? Results From The Nice Fit Study

IntroductionThe faecal immunochemical test (FIT) is a non-invasive quantitative test which measures occult blood in faeces (faecal haemoglobin, FHb). FIT was introduced by NICE in 2017 to triage referral of patients with low risk symptoms (DG30). We report on the largest prospective and ethically approved diagnostic accuracy study to date of FIT in patients with high and low risk symptoms for colorectal cancer (CRC), meeting NICE NG12 and DG30 criteria.MethodsThis multicenter study was powered to establish the diagnostic accuracy of FIT for CRC. Patients were eligible for recruitment if they experienced bowel symptoms meeting NICE two-week (TW) referral criteria and had been triaged to investigation with colonoscopy. All eligible patients referred for colonoscopy on TW pathway were asked to complete a FIT test kit (HM-JACK) prior to their colonoscopy. Patients were excluded from analysis if they did not provide a valid FIT or did not undergo complete colonoscopy. Patients were classified as high or low risk as per NICE NG12 or DG30 criteria respectively. Colonoscopy results were compared to FIT measurements of FHb and the conduct of the tests was double-blinded. Quality assurance of endoscopy and clinical data was performed by senior clinicians and external statisticians analysed anonymised data. This trial is registered on isrctn.com, ISRCTN49676259.Results9822 patients from 50 sites across England participated in the study between October 2017 to March 2019, 329 cancers were detected (3.3% prevalence). Preliminary results for combined DG30 and NG12 groups show the sensitivity of FIT at FHb thresholds of 2, 10 and 150 μg/g was 97.0%, 90.9% and 70.8% respectively. The positive predictive value (PPV) of FIT in combined groups for CRC at thresholds of 2, 10 and 150 μg/g was 8.7%, 16.1% and 31.1% respectively and the negative predictive value (NPV) of FIT at these thresholds was 99.8%, 99.6% and 98.9% respectively. 6900 patients (71.3%) had high risk symptoms (NG12). The sensitivity of FIT in this group at thresholds of 2 and 10 μg/g was 98.4% and 92.8%, respectively. The PPV was 9.1% and 16.3%, respectively. In contrast the sensitivity of FIT was significantly lower for low risk symptoms (DG30) at 91.5% and 84.5% at cut-offs of 2 or 10 μg/g respectively (p<0.01). The PPV for low risk symptoms at these thresholds was 7.7% and 16.0% respectively.ConclusionsThis is the first study to report that at the lowest threshold of detectable blood (2 μg/g), FIT sensitivity is equivalent to the current gold standard investigation of colonoscopy. The results of this study support the use of FIT as an objective diagnostic tool to triage patients with both high and low risk CRC symptoms, reducing the number of unnecessary investigations.