DISCUSSION In order to recommence endoscopy services, practice statements from recognised societies provide thorough guides on how to implement and resume endoscopy services, however a strategy is also needed to manage the backlog of diagnostic and therapeutic endoscopic procedures for colorectal patients.35 Traditionally colonoscopy has been the first line diagnostic test and gold standard for the investigation of colorectal symptoms, including patients with minor and low risk symptoms. Unsurprisingly, the low threshold for performing colonoscopy has led to low diagnostic yields of significant disease, despite a high endoscopic workload. In a review of 1116 cases over a two-year period, the correlation of lower gastrointestinal symptoms with colonoscopy showed that only 14.5% of those patients had clinically significant disease and 1.7% had CRC.6 To clear the backlog of colonoscopies during the COVID-19 pandemic, a new paradigm for diagnostic testing is required, combining patient demographics, symptoms, and non-invasive testing to prioritize colonoscopies for patients at a high risk of cancer and to avoid unnecessary colonoscopies in those with a low cancer risk. Several non-invasive tests are now available, which may allow us to triage referrals into high and low risk categories and to reprioritize patients already on a waiting list for a colonoscopy. These include faecal immunochemical test (FIT), multitarget stool DNA test (FIT-DNA), faecal calprotectin, methylated Septin9 (mSEPT9) blood test, CT colonoscopy (CTC) and colon capsule endoscopy (CCE). FIT stool testing has now superseded faecal occult blood test (FOBT). FIT is specific for human haemoglobin, its easier for patients to perform, has a higher sensitivity, less false positive and negative rates than FOBT.7,8 In a study of 1000 symptomatic patients who also had a FIT with a threshold of ≥ 10μg Hb/g, 742 (74%) patients were FIT negative. There were 48 CRCs diagnosed and 7 FIT negative cancers; one was a lymphoma and the other 6 were caecal adenocarcinomas, of which 5 were anaemic. The specificity for FIT in this study was 86.9% (CI 96%) and the negative predictive value (NPV) 99.05%. The positive predictive value (PPV) for iron deficiency anaemia (IDA) was 34% compared with other symptoms 18%. Although a negative FIT test has a NPV of 99.05%, it is not as sensitive for right sided CRCs but when combined with iron deficiency anaemia, only one lymphoma and 1 CRC would have been missed.9 In another UK study, 612 patients who were urgently referred with symptoms of suspected CRC via the two-week wait referral pathway also had a FIT test with a cut off ≥ 10μg Hb/g. Of these, 477 (77.9%) patients were FIT negative, and there were 5 false negative FIT tests in this cohort of patients, who had a total of 35 cancers and 3 high-grade dysplasia, giving a FIT sensitivity for diagnosing CRC of 86.84%, specificity 82.23%, NPV 98.95% and PPV 22.96%.10 Of the 5 FIT negative cancers, 60% had an anaemia, 40% had a mass and 80% had lost weight, therefore combining these safety-net high-risk symptoms with FIT would have detected virtually all the cancers, while still significantly reducing the colonoscopy workload. Therefore, for patients who are worried about their symptoms in the COVID-19 era, when services are severely limited and doctors are not able to provide a timely review, a negative FIT with safety-netting will provide considerable reassurance. For patients with rectal bleeding, the FIT test is not appropriate and the mSEPT9 test may be a better option to rule out CRC with reasonable accuracy. In a recent systematic review and meta-analysis of diagnostic test accuracy of mSEPT9 for the detection of CRC based on 19 studies, the pooled estimates for mSEPT9 to detect CRC showed a sensitivity of 69%, specificity of 92%, positive predictive value (PPV) of 2.6% and NPV of 99.9% in an average risk population (0.3% CRC prevalence), and 9.5% PPV and 99.6% NPV in a high-risk population (1.2% CRC prevalence).11,12 Overall, mSEPT9 has a lower sensitivity and specificity to FIT therefore, on a purely performance-based comparison, mSEPT9 is not a replacement test for FIT but a complementary test. Furthermore, it is also more expensive at approximately £100 (approximately 125 US $) versus FIT £6 (approximately 8 US $) and is also not widely available outside of the US but has been approved by the US Preventative Services Task Force as a first-line colorectal cancer screening test.13 FIT, FIT-DNA and mSEPT9 alone are not sensitive tests for excluding precancerous colonic lesions such as advanced adenomas or polyps because of the lower sensitivity ranging from 5-42%.10,11,14 Patients with potential symptoms for CRC may also be offered a faecal immunochemical test (FIT) combined with a multitargeted stool DNA test (FIT-DNA) if they do not have rectal bleeding. This test combines a FIT with testing for altered DNA biomarkers in cells which shed into the stool. The FIT-DNA performance has an increased sensitivity for detecting advanced precancerous lesions compared with FIT alone when screening an average-risk CRC population. Its specificity is lower than that of FIT alone, which means it has a higher number of false-positive results and a higher likelihood of follow-up colonoscopy and an associated adverse event per screening test.13,15 Furthermore, FIT-DNA is not readily available outside of the US. CTC is also a good diagnostic test in the first line of investigations for symptomatic colorectal patients. In a study from the UK, CTC was offered to all patients over 60 with a change in bowel habit and iron deficiency anaemia.16 Over a study period of 12 months, 1,792 straight to test CTCs were performed and the CRC detection rate was 4.9% and polyp detection rate 13.5%. The investigators concluded that the results are comparable to colonoscopy in terms of diagnostic accuracy and similar to those of CTC in published multicentre trials. CTC also has the advantage of detecting extra-colonic pathology. In the future, CCE may also play a role in evaluating patients. The SCOTCAP Scottish Capsule Project trial which evaluated 455 symptomatic and surveillance patients with CCE ended March 2020 and the results are pending.17,18 No single test, including colonoscopy, has a 100% sensitivity or specificity for diagnosing or excluding CRC. In a retrospective, single-centre study of post-colonoscopy CRCs diagnosed within 636 months, the ratio of interval cancers (IC) to all new CRC over the decade between 2006-2015 was between 6-11.5%.19 Reasons for this include (1) missed diagnosis; (2) incomplete resection of advanced adenomas and (3) neoplasms that grow at faster rates than average sporadic CRC.2022 Missed diagnoses are likely in the first 36 months after a colonoscopy on the basis that it is too soon for a de novo cancer to have arisen.23 In this study a difficult examination was often associated with IC, and may be a risk factor for this problem.19 In a recent review of the quality in colorectal cancer screening with colonoscopy, interval CRC most commonly occurs within 24 months among individuals who had only 1-2 small adenomas on initial colonoscopy and this appeared to arise from flat, large (1 cm) adenomas or serrated polyps in the right side of the colon.24 For these reasons if the bowel preparation is not adequate, then colonoscopy should be repeated and increasing withdrawal time to 8-10 minutes will increase the adenoma detection rate.24 As it is generally accepted that CRCs typically grow over many months and years prior to presentation, one would expect that short delays in diagnosis and surgery should not have a negative impact on patient outcomes. There are no reported studies analysing the delay in diagnosis on survival outcomes but studies that analyse the effect of a delay in treatment from the diagnosis of CRC show that a deferral of surgery for up to 12 weeks is unlikely to have any effect on the survival outcome of CRC patients.2527 With the markedly reduced capacity of our endoscopy services during the COVID-19 pandemic, we need to consider radical solutions in the way we utilize colonoscopy as the first diagnostic test for patients with colorectal symptoms to try to identify patients early with CRC and prevent poor survival outcomes. Therefore, for patients already on diagnostic pathways and awaiting a colonoscopy, it will be essential to triage these patients into different risk categories for cancer. Both the FIT and mSEPT9 tests have proven reasonably good accuracy when combined with safety netting for excluding CRC and may be utilized to stratify patients into high and low risk groups , with the high risk groups then being re-prioritized for a diagnostic colonoscopy.9,13,28 The first line investigation should be the FIT test because of its higher sensitivity and specificity and lower cost than mSEPT9, which should be reserved for those patients with rectal bleeding where FIT is not appropriate. Both FIT and mSEPT9 tests have a high NPV and are therefore good tests to rule out cancers but they do not have a 100% sensitivity and a few cancers can be missed, i.e. false negatives, so it is important to have some safety netting based on patients symptoms and blood tests e.g. weight loss, mass, iron deficiency anaemia. Both Figures 1 and 2 shows a method of prioritization of symptomatic patients for colonoscopy using several factors which we propose during the COVID-19 pandemic. The ideal and most cost effective scenario will be for the availability of a FIT test and a mSEPT9 blood test to assist in risk stratification and prioritization of patients as these tests are relatively inexpensive and should help manage endoscopy demands until the burden of the COVID-19 pandemic subsides and endoscopy services can return to normal, which is not likely to be for many months, if not years, unless an effective vaccine becomes available. During the COVID-19 endoscopic recovery period, by using non-invasive tests such as the FIT and mSEPT9 tests and reducing the need for colonoscopy by up to 80%, this has huge cost saving implications. Furthermore, colonoscopies are also likely to have a higher yield of CRC if non-invasive tests are used to approve referrals as compared to 1.7% who had CRC in the study discussed above which was based on clinical symptoms only.6 However, rationing ones service to make it more efficient should not be at the sacrifice of small numbers of missed CRCs and efforts should be made to minimise this. Our recommendation is to use FIT with or without mSEPT9 and safety netting to risk stratify patients for significant pathology and for ruling out cancer thus requiring no further investigations. Those with positive FIT tests or with high risk symptoms such as weight loss, abdominal mass or iron deficiency anaemia should have an urgent investigation irrespective of their FIT or mSEPT9 test results. This would avoid missing the small number of cancers i.e. false negatives. A later clinical review of moderate and low risk patients should be made of FIT negative (if appropriate) or mSEPT9 negative. If symptoms persist despite appropriate initial clinical working diagnosis and treatment, and there is a clinical concern of colorectal cancer, then a colonoscopy could be performed at that stage. If there is still limited access to colonoscopy or CTC, another option is to repeat the FIT test if appropriate or the mSEPT9 test. Although not supported by clinical investigatory evidence a second negative test result is likely to be more reassuring than just one test result. Nevertheless, the importance of assessing each clinical case based on its own merits and not relying on just FIT or mSEPT9 cannot be overstated. CONCLUSION The COVID-19 pandemic has severely restricted our ability to provide timely diagnostic endoscopic tests to rule out colorectal cancer in symptomatic patients. Due to infection control measures, endoscopy waiting lists may still be incapacitated for several months due to COVID-19. In this manuscript we discuss strategic opportunities to drastically reimagine endoscopy services with non-invasive test and a safety netting approach to prioritize the use of colonoscopy. Ultimately, what may transpire from an overdue rethinking and reutilization of endoscopy services may be more efficient use in the future of our colonoscopy services post COVID-19. ACKNOWLEDGMENT Alison Walker English language proof reading |