A novel digital rectoscope for the triage of lower gastrointestinal symptoms in primary care: a prospective multicentre feasibility study

Study design and setting

LuCID is a prospective, multicentre observational feasibility study of a novel digital endo-rectal examination device called the LumenEye X1 in primary care in patients with lower GI symptoms fulfilling the criteria for a 2WW pathway referral¹² or IBD.

The study sites were NHS primary care practices located in Scotland (Forth Valley) and England (Blackburn, London [north and south], and Berkshire). Primary care practitioners were invited to take part in the study by letter of invitation. Endoscopic examinations were performed in a primary care setting by clinicians who had completed training on device usage, but not disease recognition. Colorectal surgeons delivered training using a synthetic simulated rectal model followed by at least five observed procedures. Training competency included the following: 1) safe operation of the device with appropriate sterilisation; 2) reproducible identification of key anatomical landmarks; and 3) quality assurance assessment including bowel preparation quality, complete recording of the anatomy, and a 360-degree inspection of the mucosa on withdrawal of the device. A ‘high-quality’ examination was defined as complete visualisation of the luminal mucosa from the dentate line up to and including identification of the rectosigmoid junction.

National and international bowel cancer charities were invited to comment on the study design before the start of the trial. An abstract describing the study was published on the website of the manufacturer SurgEase Innovations,¹³ and updates of its progress were regularly posted on social media including LinkedIn and Twitter.

Patients

Patients were recruited from general practice or from urgent suspected colorectal cancer referral waiting lists held in secondary care. The rationale for accepting patients from both routes was to increase the capture of patients who needed urgent assessment for their symptoms but faced lengthy delays for face-to-face assessment owing to the pandemic. Adults aged ≥18 years meeting any of the following criteria were included: symptoms fulfilling the 2WW criteria for suspected colorectal cancer;¹² known patients with IBD experiencing flare symptoms; patients with a suspected new diagnosis of IBD; patients with a positive faecal calprotectin, a raised faecal occult blood (FoB), or faecal immunochemical test (FIT) estimation; or any GI symptoms with a past history of colorectal polyps. Exclusion criteria included the following: the inability to provide informed consent or communicate effectively in English (to limit the number of people in small clinical rooms during the pandemic); pregnancy; allergy to plastics; the presence of an anal stricture; or the inability to take bowel preparation or adopt the left lateral position required for rectal examination. All suitable patients on 2WW lists in secondary care were contacted and could enrol in the study if they agreed. In primary care, clinicians could contact appropriate patients from their clinic lists. Patients were approached either by telephone or face-to-face a week before the examination visit, allowing sufficient time for participants to read the patient information sheet and provide informed consent. Patients were provided with a glycerine suppository to administer at home before attending for examination.

Each site was provisionally set a limit of 25 patients. However, owing to factors such as disparate patient footfall, medical personnel availability, and time taken to get relevant approvals, the range of patients recruited by site ranged from 10–34. Following national lockdown restrictions in January 2021, recruitment was paused for 3 months, and permission to extend the trial was granted by the funder and ethics review board. All patients were referred for further investigation at the primary care clinician’s discretion. Patients recruited from the 2WW referral pathway had the diagnostic tests they would have undergone had the LumenEye examination not been performed. The reports of any subsequent investigations were obtained from the patient’s electronic health record. Where appropriate, treatments for haemorrhoids, anal fissures, or IBD could be initiated based on the LumenEye findings in primary care. The results were reported in accordance with the Standards for Reporting of Diagnostic Accuracy (STARD) (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Flow diagram demonstrating the outcome of patients in the trial. Follow-up data are available for 110 of 114 patients. Three patients were unable to tolerate the procedure and one patient was lost to follow-up despite numerous efforts to contact him. CTVC = computed tomography virtual colonoscopy.

Tele-endoscopy platform (CHiP)

All clinicians were given training on the CHiP software (Figure 1B and C) and were allocated secure login credentials and passwords. Data were stored on a fully encrypted Azure cloud server, which adhered to all NHS security and governance standards including cyber security, data processing, information management, and penetration testing. The cloud server was hosted by the company that held ISO-27001, Cyber Security Essentials Plus, and NHS Data Security and Protection (DSP) Toolkit accreditation. The decision to use the CHiP platform was left to the primary clinician‘s discretion with simultaneous technical assistance, if needed. A secondary care clinician in the local hospital was available at the time of the examination with simultaneous access to the CHiP platform owing to the scheduling feature on CHiP. Specialists could also dial-in ad hoc, from their laptops or mobile phones, if needed. Videos and/or images were stored and reviewed retrospectively within one working day by an independent expert endoscopist for every patient recruited to the study to minimise the risk of missed pathology. Where pathology was missed by the examining primary care physician, the clinicians and patient were alerted and appropriate action taken.

Digital rectoscopy

The digital rectoscopy examination was performed with the patient in the left lateral position without sedation using the standard technique for rigid sigmoidoscopy. Manual air insufflation using the in-built bellows is required to achieve optimal rectal distension. Quality assessment of bowel preparation was made by each clinician using criteria based on the Boston Bowel Preparation Scale.¹⁴ Each bowel was assessed to be good (>80% mucosa visible), average (>50% mucosa visible), or poor (<30% mucosa visible). The duration of the examination and any discomfort experienced by the patient was recorded. Following the examination, a questionnaire developed from a published rigid sigmoidoscopy experience study¹⁵ was completed by the patient (see Supplementary Appendix S1). A case report form after each examination was completed recording details on bowel preparation, duration of the examination, distance of insertion, pathology detected, number of air insufflations required, and a rating of the ease of use of the LumenEye and CHiP technologies (see Supplementary Appendix S2).

Outcome measures

The primary outcome was patient and clinician acceptability. Secondary outcomes included safety and analysis of diagnostic accuracy, referral outcomes, and cost.

Variables

The main independent variables were qualitative assessment of patient comfort, anxiety, convenience, and level of reassurance of having digital rectoscopy assessment in the community. To avoid the risk of reporting bias, the entire cohort of patients was used in the user-feedback analysis.

For clinicians, the level of diagnostic confidence was the main variable. The procedures were performed by 10 primary care practitioners (eight GPs and two advanced nurse practitioners), with variable endoscopy experience ranging from novices to Joint Advisory Group accreditation for flexible endoscopy, which likely influenced clinician confidence and diagnostic ability. A confounding variable was the fact that patients were selected by clinicians, which could positively bias patient responses. There is also likely to be a high pick-up rate of pathology owing to all patients selected having fulfilled either 2WW criteria or having results indicative of having disease. Additionally, a degree of patient response bias is anticipated as they were asked to complete the questionnaire straight after the procedure. Ideally, an independent body would have requested patient feedback separate to the clinical event.

Statistical analysis

The conditions set by the funding body for this study stipulated trial completion to be achieved within 6 months of the first recruitment. After consultation with each primary care site and with consideration of the risk of national lockdown during the study, a pragmatic monthly recruitment target of one patient per week per site was deemed practical and not burdensome to primary care during times when face-to-face consultations were limited. With 1 month dedicated to training and recruiting NHS sites, a recruitment rate of four patients per month per site was set. A 10% attrition rate was calculated, setting the final recruitment target of 110 over 5 months. Because this study was designed to assess feasibility with qualitative end-points and determine the practicalities of trial delivery for a subsequent phase III study, this approach was deemed appropriate and of similar size to a comparable investigational point-of-care technology study.¹⁶ Analysis of quantitative data were performed on an as-treated basis and on an intention-to-treat basis in the qualitative domains. Variables were expressed as median and interquartile range (IQR), and the χ² and Kruskal-Wallis tests were used to compare categorical non-parametric data between groups. The Pearson correlation coefficient was used to identify significant relationships between variables. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the LumenEye against the gold standard (either CTVC, flexible sigmoidoscopy, or colonoscopy) was compared with a receiver operating characteristic (ROC) analysis. All statistical calculations were undertaken using Statistical Package for the Social Sciences (SPSS; version 27). A P-value of <0·05 was taken to be statistically significant.