Faecal immunochemical tests versus colonoscopy for post-polypectomy surveillance: an accuracy, acceptability and economic study

Background: In the UK, patients with one or two adenomas, of which at least one is >= 10 mm in size, or three or four small adenomas, are deemed to be at intermediate risk of colorectal cancer (CRC) and referred for surveillance colonoscopy 3 years post polypectomy. However, colonoscopy is costly, can cause discomfort and carries a small risk of complications. Objects: To determine whether or not annual faecal immunochemical tests (FITs) are effective, acceptable and cost saving compared with colonoscopy surveillance for detecting CRC and advanced adenomas (AAs). Design: Diagnostic accuracy study with health psychology assessment and economic evaluation. Setting: Participants were recruited from 30 January 2012 to 30 December 2013 within the Bowel Cancer Screening Programme in England. Participant: Men and women, aged 60-72 years, deemed to be at intermediate risk of CRC following adenoma removal after a positive guaiac faecal occult blood test were invited to participate. Invitees who consented and returned an analysable FIT were included. Intervention: We offered participants quantitative FITs at 1, 2 and 3 years post polypectomy. Participants testing positive with any FIT were referred for colonoscopy and not offered further FITs. Participants testing negative were offered colonoscopy at 3 years post polypectomy. Acceptibility of FIT was assessed using discussion groups, questionnaires and interviews. Main outcome measures: The primary outcome was 3-year sensitivity of an annual FIT versus colonoscopy at 3 years for detecting advanced colorectal neoplasia (ACN) (CRC and/or AA). Secondary outcomes included participants' surveillance preferences, and the incremental costs and cost-effectiveness of FIT versus colonoscopy surveillance. Result: Of 8008 invitees, 5946 (74.3%) consented and returned a round 1 FIT. FIT uptake in rounds 2 and 3 was 97.2% and 96.9%, respectively. With a threshold of 40 pg of haemoglobin (Hb)/g faeces (hereafter referred to as mu g/g), positivity was 5.8% in round 1, declining to 4.1% in round 3. Over three rounds, 69.2% (18/26) of participants with CRC, 34.3% (152/443) with AAs and 35.6% (165/463) with ACN tested positive at 40 mu g/g. Sensitivity for CRC and AAs increased, whereas specificity decreased, with lower thresholds and multiple rounds. At 40 mu g/g, sensitivity and specificity of the first FIT for CRC were 30.8% and 93.9%, respectively. The programme sensitivity and specificity of three rounds at 10 mu g/g were 84.6% and 70.8%, respectively. Participants' preferred surveillance strategy was 3-yearly colonoscopy plus annual FITs (57.9%), followed by annual FITs with colonoscopy in positive cases (31.5%). FIT with colonoscopy in positive cases was cheaper than 3-yearly colonoscopy (2,633,382) pound, varying from 485,236 pound (40 mu g/g) to 956,602 pound (10 mu g/g). Over 3 years, FIT surveillance could miss 291 AAs and eight CRCs using a threshold of 40 mu g/g, or 189 AAs and four CRCs using a threshold of 10 mu g/g. Conclusion: Annual low-threshold FIT with colonoscopy in positive cases achieved high sensitivity for CRC and would be cost saving compared with 3-yearly colonoscopy. However, at higher thresholds, this strategy could miss 15-30% of CRCs and 40-70% of AAs. Most participants preferred annual FITs plus 3-yearly colonoscopy. Further research is needed to define a clear role for FITs in surveillance. Future work: Evaluate the impact of ACN missed by FITs on quality-adjusted life-years.