Endocuff Vision for assisting visualisation during colonoscopy

The committee noted that 3 of the 4 studies showed that Endocuff Vision improved the adenoma detection rate in people having colonoscopies as part of bowel cancer screening. The ADENOMA trial was the only study powered to detect a difference in adenoma detection rates and the results showed a statistically significant improvement with Endocuff Vision of almost 11%. The external assessment centre (EAC) concluded that ADENOMA was a high-quality study with a low risk of bias. The clinical experts confirmed that the trial accurately represented NHS clinical practice. The committee considered the E‑Cap study which was the only trial that showed no improvement in adenoma detection rates with Endocuff Vision. It understood that this study included only 4 colonoscopists who already had baseline adenoma detection rates for standard colonoscopy that were much higher than the reported average of 46.5% (Lee et al. 2012). The committee noted that the benefits of Endocuff Vision may have been less apparent in this study because of the colonoscopists' relatively high level of expertise. Having considered all the clinical evidence, the committee concluded that using Endocuff Vision increases the adenoma detection rate in people having colonoscopies as part of bowel cancer screening and noted that the UK National Screening Committee is responsible for deciding if new technologies, such as Endocuff Vision, should be included in relevant screening pathway service specifications.

The committee noted that secondary outcomes from ADENOMA included a statistically significant increase in the number of detected cancers with Endocuff Vison. The study was not powered to detect differences in cancer detection rates, but the committee considered this secondary outcome to be promising. The committee concluded that more evidence on the detection of cancers during bowel cancer screening with Endocuff Vision would be helpful.

Only 2 of the 4 studies (E-Cap and ADENOMA) included patients having colonoscopies unrelated to bowel cancer screening (n=1,135). E‑Cap also included patients who were under surveillance after having polyps removed, but ADENOMA was the only study that involved both symptomatic and surveillance populations. The clinical experts explained that the baseline adenoma detection rate was much lower in people having colonoscopies unrelated to bowel cancer screening, because this population is more heterogeneous in terms of patient age and presenting symptoms than the bowel cancer screening population. Results from the ADENOMA trial showed that use of Endocuff Vision in a non-screening population (n=975) was associated with a marginal (0.4%) improvement in adenoma detection rate from a baseline rate of 23.9% in the standard colonoscopy group. The committee therefore concluded that the benefits of Endocuff Vision are less certain in these patients.

The committee noted that adenoma detection rate is the most widely used quality indicator for diagnostic colonoscopy: it is used by the Joint Advisory Group on Gastrointestinal Endoscopy (JAG) and the NHS bowel cancer screening programme to monitor colonoscopy standards. The clinical experts, including a representative from the bowel cancer screening programme, explained that the minimum rate is 40% but that centres aspire to 50%. The clinical experts agreed that the average adenoma detection rate in a screening population was around 50%, and that anything above 65% would be considered exceptional. However, they noted that adenoma detection rates vary in both screening and non-screening populations. The clinical experts also explained that the adenoma detection rate is influenced by a number of factors, including how long the colonoscopist spends inspecting the walls of the bowel, how well the bowel is cleansed, levels of colonoscopist fatigue and the resolution of the colonoscope.

The clinical experts explained that people having bowel cancer screening may develop subsequent cancers that are referred to as 'interval cancers' (that is, cancers that are diagnosed in the intervals between tests). The clinical experts noted that the incidence of interval cancers varies across the UK from 2% to 13%. Detecting and removing more adenomas may lead to a subsequent reduction in the risk of interval cancers (because some adenomas may progress to cancerous lesions). The EAC and clinical experts explained that increases in adenoma detection and resection rates are associated with a reduced risk of interval cancers. They referred to 2 studies that evaluated this relationship:

• In Kaminski et al. (2010), people who had colonoscopies done by colonoscopists with an adenoma detection rate of less than 20% were at a higher risk of developing interval cancers than people who had colonoscopies done by colonoscopists with an adenoma detection rate of over 20%.

• In Corley et al. (2014), an increase in the adenoma detection rate of 1% resulted in a decrease in the risk of developing interval cancers of 3%.
  
  The clinical experts cautioned that a plateau effect is likely for the adenoma detection rate, above which there would be diminishing effects on interval cancer risk. Because these studies only evaluated baseline adenoma detection rates of up to 50%, it is uncertain if and at what level this plateau would become apparent. The clinical experts were uncertain as to what effect the increase in adenoma detection rate from 50% to 60% (as seen with Endocuff Vison in the ADENOMA trial) would have on interval cancer risk. However, they were not aware of any evidence to suggest that the increase would not be clinically useful. Consequently, the committee concluded that the adenoma detection rate could be considered a predictor of interval cancer risk, and so it is plausible that Endocuff Vision may reduce the incidence of interval cancers.

The committee noted that some rare adverse events have been reported with the predecessor device, Endocuff, including colon perforation. The clinical experts explained that Endocuff Vision has been designed to avoid these problems. The committee noted that the incidence of adverse events with Endocuff Vision was comparable to that of standard colonoscopy.

The clinical experts explained that the priority in colonoscopy is to achieve a complete procedure (with a high caecal intubation rate) with minimal discomfort to the patient. They noted that the use of Endocuff Vision may cause a small amount of pain and discomfort when inserting the colonoscope, but this is usually temporary and no additional sedation is needed. The clinical experts emphasised that the decision to use Endocuff Vision should be based on clinical judgement taking into consideration patients' comorbidities, physical presentation, symptoms and age.

The clinical experts described certain circumstances when Endocuff Vision is unlikely to provide benefits. These included: investigation in patients with a history of significant anal pathology, moderate to severe diverticulosis or a fixed sigmoid colon; for colonoscopies in young, non-sedated patients because of possible anal trauma; and when inflammatory bowel disease is suspected or active colitis is present. Endocuff Vision should be used with caution in patients aged over 85 years because of the higher risk of diverticular disease. The clinical experts also identified circumstances when using Endocuff Vision may be technically helpful: for example, as an insertion aid in patients with long and mobile colons, or when detecting and removing large polyps or those in difficult areas of the colon.

All 4 of the included studies were done at UK centres. The ADENOMA trial recruited patients across 7 NHS centres and involved a total of 48 colonoscopists with varying skill levels. Each site was allowed to have no more than 4 bowel cancer screening programme colonoscopists. The committee was aware of the role of JAG accreditation for colonoscopists in the bowel cancer screening programme. The bowel cancer screening programme establishes quality standards for screening colonoscopy, provides training resources for staff and records a national audit of the screening service. A representative from the screening programme confirmed that the use of Endocuff Vision is currently an option for screening colonoscopists. The committee recalled that the results of the single-centre studies were potentially biased by local expertise (see section 3.3) but it understood that the expertise of colonoscopists varied across the NHS. The committee concluded that overall the evidence was broadly generalisable to the NHS.

Endocuff Vision is compatible with most colonoscopes and is available in 4 different sizes, depending on the diameter of the colonoscope used for the procedure. The clinical experts emphasised the importance of selecting the correct size and ensuring that Endocuff Vision is securely fitted to avoid detachment during the procedure. The clinical experts clarified that using Endocuff Vision is associated with a learning curve: in the ADENOMA trial, colonoscopists needed specific training and had to have done at least 20 procedures with Endocuff Vision before starting the study. The clinical experts considered that 20 to 30 procedures with Endocuff Vision would be sufficient to ensure competency. In its cost model, to represent the learning curve with Endocuff Vision, the company assumed that there would be no improvement in efficacy for the first 20 procedures with Endocuff Vision (for full details see section 4.2 of the assessment report).

The committee recalled the EAC's conclusion that the level of improvement in the adenoma detection rate with Endocuff Vision was inversely related to the expertise of the colonoscopists involved; that is, more benefits were seen with Endocuff Vision when used by colonoscopists who had lower baseline adenoma detection rates (see section 3.3). The EAC identified a recent meta-analysis (Williet et al. 2018), which showed that Endocuff Vision (or its predecessor device Endocuff) improved adenoma detection rates in colonoscopists with low-to-moderate baseline rates (less than 35 %) but not in colonoscopists with high baseline rates (over 45 %). The clinical experts broadly agreed with this observation, but noted that statistically significant improvements were still seen in the ADENOMA trial even when baseline adenoma detection rates were above 45%.

The results of the ADENOMA trial showed that the use of Endocuff Vision does not increase overall procedure time. The clinical experts also noted that in the trial, the withdrawal time for Endocuff Vision was comparable to that of standard colonoscopy despite more polyps being removed: this suggested that Endocuff Vision may increase efficiency. The clinical experts confirmed that this accurately reflected their own clinical experience; they added that Endocuff Vision may help with polyp removal by anchoring the scope and providing a degree of device stability. In the ADENOMA trial, inserting Endocuff Vision took 1 minute less than inserting a standard colonoscope. Although the clinical experts considered this to be a notable time saving, they understood that it would be unlikely to lead to an increase in patient throughput. However, they clarified that any improvements in procedure efficiency may lead to lower levels of colonoscopist fatigue and may be of particular value in complex cases.

The committee considered the EAC's changes to the company's cost model (see section 3.5) and agreed that the updated parameters better reflected cost and resource use in a UK screening population. The committee acknowledged the unusually large number of assumptions in the cost modelling, but considered that this approach was necessary to fully capture Endocuff Vision's effect on adenoma detection rates and the likely incidence of interval cancers.

The EAC's sensitivity analyses showed that Endocuff Vision was cost saving when it improved the adenoma detection rate by over 3%. The clinical experts explained that this level of improvement was feasible in the context of a UK screening programme. Furthermore, the committee noted that the results of the ADENOMA trial implied that a 3% improvement is plausible even within a service delivered by colonoscopists with high baseline adenoma detection rates (and therefore high levels of expertise).

The EAC's revised cost model showed that over 10 years, the use of Endocuff Vision was associated with a cost saving of around £53 per patient if used by a colonoscopist with a baseline adenoma detection rate of 51%. An EAC scenario analysis showed that when initial screening is done using faecal immunochemical testing instead of faecal occult blood testing, the estimated cost saving per patient with Endocuff Vision increased to around £58. The committee noted that, although the clinical evidence for Endocuff Vision is in people with a positive faecal occult blood test result, the technology is also likely to benefit people who have a positive faecal immunochemical test result (because the test does not fundamentally change the colonoscopy procedure itself). The EAC also adapted the model to a non-screening population (see section 3.6), and results showed that using Endocuff Vision in this population is likely to incur additional costs. The committee concluded that using Endocuff Vision in people having colonoscopies as part of bowel cancer screening is likely to be associated with a small cost saving, largely through a reduced incidence of interval cancers and by allowing the diagnosis of cancers at an earlier stage. The committee concluded that on the basis of the evidence presented, the case for adopting Endocuff Vision in a screening population was supported.