Photodocumentation in colonoscopy: the need to do better?

Introduction

There is wide variation in the recommendations for photodocumentation at colonoscopy. The British Society of Gastroenterology (BSG) quality assurance standard suggests ‘photographic proof of ileocaecal valve, terminal ileum, anastomosis or appendix orifice [is] required in all cases’.¹ Like the American Society for Gastrointestinal Endoscopy (ASGE)/American College of Gastroenterology (ACG) guidelines,² documentation of any other part of the colon is not currently mandated. Although caecal photographs serve to confirm that the colonoscope has been passed the full length of the colon, they do not document whether the colonic mucosa has been adequately visualised. Photodocumentation is a relatively simple ‘push button’ process, and most endoscopists now have access to reporting systems that can automatically integrate multiple images in real time. A pertinent question therefore is, should we perform more comprehensive photodocumentation and will this lead to improvements in the quality and reliability of the examination?

Background

One approach to overcoming the limitations of static photodocumentation would be to video record the entire procedure, and with improvements in data storage, this is now technically possible. Video-recording has been found to increase mean inspection time by 49% and quality of mucosal inspection technique (measured by blinded assessment) by up to 30%.³ However, there are still significant data storage costs and issues arise around sharing of information with colleagues and manipulation of large data files. Given that endoscopists are familiar with photodocumentation and files are smaller and easy to embed in current software records, enhancing current practice with specific emphasis on quality and training in photodocumentation may be a more acceptable and realistic way of enhancing quality.

An example of where enhanced photodocumentation has been recommended is the BSG standards for upper gastrointestinal (GI) endoscopy.⁴ Eight photos of anatomical landmarks are now advised. The impetus for this approach was a recognition that an improvement in quality is required to address the 11.3% failure rate in diagnosing cancer within 3 years of an upper GI endoscopy.⁵ The European Society of Gastrointestinal Endoscopy (ESGE) goes further recommending at least 10 images in total.⁶ Similarly for colonoscopy, the ESGE now recommends documentation of nine anatomical landmarks during colonoscopy (see landmarks 2-10 in figure 1).^{7 8}

Figure 1.

Recommendation of anatomical landmarks to photodocument during colonoscopy and examples of polyp photodocumentation. 1, terminal ileum; 2, caecum and appendiceal orifice; 3, caecum and ileocaecal valve; 4, ascending colon under the hepatic flexure; 5, transverse colon just distal to the hepatic flexure; 6, transverse colon just proximal to the splenic flexure; 7, descending colon below the splenic flexure; 8, middle part of the sigmoid; 9, lower part of the rectum in forward view; 10, lower part of the rectum in retroflexed view. (A) Sigmoid colon polyp photo under white light with magnification. (B) Sigmoid colon polyp photo under NBI. NBI, narrow band imaging.

This type of enhanced photodocumentation, particularly with a new emphasis on picture quality, will become increasingly important as we move towards more routine use of optical diagnosis.⁹ The use of a ‘resect and discard’ strategy has recently been endorsed by the ESGE.¹⁰ Such an approach relies on high-quality photodocumentation to replace histology as the sole ‘polyp record’. In addition, there is now growing recognition of postcolonoscopy colorectal cancer (PCCRC), a potent reminder of the need for high standards of examination and documentation at colonoscopy. In a population-based cohort study, Burr et al found a PCCRC-3yr rate of 7.4% (proportion of people having a negative colonoscopy and subsequently being diagnosed as having colorectal cancer 6–36 months later).¹¹ The World Endoscopy Organization has published consensus statements on PCCRC including minimum datasets to facilitate PCCRC analysis which include the recommendation to photodocument at least two of three caecal landmarks.¹² PCCRCs are known to occur most commonly in the rectum but can occur throughout the colon,¹³ so it would seem logical to extend photodocumentation to include a more comprehensive photo dataset as standard.

These factors underline the increasing importance of photodocumentation in colonoscopy. In this paper, we will explore the current standard of photodocumentation in clinical practice, barriers to high-quality photodocumentation and how these can be overcome, as well as future trends.

Current standard of photodocumentation

The national colonoscopy audit, which reviewed data from 20 085 colonoscopies performed in the UK in 2011, found that caecal intubation was confirmed by photography in only 50.2% of cases.¹⁴ A more recent study reviewed procedural data obtained in 2018 from 317 consecutive colonoscopies performed in Ireland.¹⁵ Here, four experts assessed the level of certainty of caecal intubation and found adequate photographic documentation in only 53.1% of cases. Interestingly, there was a statistically significant correlation between the number of images captured at colonoscopy and the rate of successful caecal photodocumentation.

In a much larger study, ESGE colonoscopy performance measures were assessed in 6445 colonoscopies across 12 European countries. Where it was the intended point of completion, caecal intubation was successfully recorded in just 77.5% of cases.

In a study examining the sensitivity and specificity of endoscopic photographs for determining completion of colonoscopy, 80 pairs of photographs were taken from completed colonoscopies (completion validated by video clips).¹⁶ These were combined with photos of non-caecal colonic sites and were sent for review by experienced endoscopists. Using just two endoscopic photographs to document colonoscopy completion produced a sensitivity of 51.4% and a specificity of 89.2%. These studies suggest that photodocumentation at colonoscopy is currently not adequate both in terms of quantity and quality.

Importance of photodocumentation

Photodocumentation during colonoscopy serves several important functions. Primarily, it provides an objective record of procedure completeness by documenting the extent reached. Where the endoscopist takes representative photographs, it may also allow assessment of other markers of procedure adequacy such as bowel preparation quality. In addition, significant pathological findings and details of therapy performed can be recorded.

The process of photodocumentation may also have secondary effects which could indirectly improve the quality of the procedure. For example, Park et al found in upper GI endoscopy that photodocumentation of the ampulla was significantly associated with the detection rate of small upper GI neoplasms.¹⁷ In addition, in colonoscopy, there is evidence that good caecal photodocumentation is a surrogate marker for polyp detection rate and therefore overall colonoscopy quality.¹⁸

At a broader level, the photographic record provides an objective account of the procedure which facilitates communication between clinicians and patients. This is particularly important where patients return for follow-up procedures to assess, for example, a postpolypectomy site. It also serves as an important medicolegal record.

Overcoming barriers to photodocumentation

There are several potential barriers to good-quality photodocumentation. Incorrect colonoscope set-up makes photodocumentation harder than it needs to be. To optimise colonoscope button set-up for the ‘photodocumentation process’, we suggest that buttons 1 and 4 be used, in any combination, to freeze the screen and take the photo (see figure 2). Although the set-up will ultimately come down to individual preference, this particular button assignment makes it ergonomically easy to take photos rapidly during a procedure. We have found on average an experienced endoscopist needed to freeze and unfreeze the image 1.9 times before a satisfactory picture was obtained; therefore, the freeze/unfreeze buttons should be easily accessible and used only with minimal shift in left-hand position.

Figure 2.

Authors’ preference for scope button set-up on a 290 series Olympus colonoscope and hand position relative to image capture buttons.

Inadequate time to take good-quality photographs and label images accurately during busy endoscopy lists may also hinder good-quality photodocumentation. However, artificial intelligence (AI) systems currently in development have the potential to automate this process in real time to minimise the manual work required. Furthermore, if linked to voice activation systems, this would remove the need to change hand position to take and record photos, freeing the endoscopist to concentrate completely on lining up an in-focus and clear image.

Poor-quality photographs taken during colonoscopy will not have the desired benefits associated with photodocumentation as they do not meaningfully add to the patient record and may actually be misleading to a reviewer. Common issues include out-of-focus polyps or where the polyp is too far away to see clearly (see figure 3).

Figure 3.

Common issues with photo quality (A–C) and high-quality photos (D–E). NBI, narrow band imaging.

Uncertainty about the site at which the photo is taken is an issue especially where images are not accurately labelled. This can be partially overcome by combining photodocumentation with a magnetic endoscope image which provides an objective marker of the site at which the photograph was taken (see figure 4).

Figure 4.

Splenic flexure polyp with accompanying magnetic endoscope image (ScopeGuide). NBI, narrow band imaging.

Lack of training is also a barrier to good-quality photodocumentation. Lu et al demonstrated that a performance review and an education session significantly improved the mean photodocumentation rate of the appendiceal orifice from 55% to 91% (p=0.03).¹⁹ In a previous audit of 184 gastroscopies, we examined the quality of upper GI endoscopy lesion reporting and found a training session significantly increased the number of anatomical sites documented from 3.8 to 6.9 photos per procedure (p<0.001). There was also a trend towards increasing number of lesions detected from 0.5 to 1.2 per procedure (p=0.230).²⁰ Simple training measures can have a significant impact on implementation of good-quality photodocumentation (see box 1).

Box 1. How to take excellent polyp photos: basic competencies.

1. Ensure a stable scope position and ensure the entire polyp is visible.

2. Wash and clean the polyp.

3. If an en face view is not possible or there is aortic/respiratory induced wall movement, use a biopsy forceps, snare catheter or open snare: push the instrument forward to flatten the fold, invert and stabilise the polyp position towards the lens.

4. Ensure the polyp surface is in focus.

5. Confirm that a high-quality photo has been achieved. Be patient; freeze the frame; take a white light and blue light image (eg, Narrow Band Imaging/Blue Light Imaging/i-Scan 2). Repetition of the process may be required to achieve a high-quality image.

6. Label images clearly.

7. Audit photodocumentation as a key quality metric.

Outside medicine, there are scenarios where clear guidance is provided in order to achieve good-quality photos such as for passport photos. In these contexts, it is important to take standardised photos. Similarly, within endoscopy, tailored guidance in photodocumentation of colonic mucosa and polyps may support training and could also be used as a research tool where images are retrospectively analysed. We propose the ‘3C Endoscopy Photo Quality Checklist’ which evaluates the three key requirements of a high-quality photo (see box 2).

Box 2. 3C Endoscopy Photo Quality Checklist to evaluate the quality of an endoscopic landmark/polyp photo.

1. Clean the mucosal surface.

2. Complete view (en face or panoramic view of the relevant area).

3. Correct focal distance (sharp).

In addition, we suggest all white light polyp photos should be accompanied by a blue light image which enhances views of vascular structures and thereby facilitates optical diagnosis. We also recommend a minimum of 10 photos of anatomical landmarks, including terminal ileum, are taken (see figure 1). Key learning points for photodocumentation are summarised in box 3.

Box 3. Key learning points and targets for research.

Key learning points

• Ensure the colonoscope buttons are correctly (ergonomically) set up to facilitate photodocumentation.

• Before taking photos, take time to optimise the picture (clean, complete, correct focal distance).

• Ensure a comprehensive photo set is recorded per colonoscopy—we recommend a minimum of 10 photos.

• Polyps should be photodocumented with white and blue light images.

• Label photos carefully at the end of the procedure.

Targets for research

• Real-time labelling of photos during colonoscopy using voice recognition.

• Autofocusing scopes to enhance picture quality.

• Use of artificial intelligence to assist in taking high-quality photos.

• Does enhanced photodocumentation reduce post-colonoscopy colorectal cancers?

Adopting an enhanced photodocumentation approach could potentially increase the length of an endoscopy report. In the context of growing awareness of our environmental impact in endoscopy (green endoscopy),²¹ some might raise concern over the increase in paper waste generated by printing more photos per case. This can be mitigated by accelerating the shift towards sharing reports digitally, which will help avoid this issue in the future.

Future of photodocumentation

AI has the potential to allow automatic capture of photos when the scope is at the correct anatomical site. In addition, it will be able to provide real-time feedback about the quality of a photo taken. Developments in voice recognition technology may allow real-time labelling of images with the potential to reduce recall error in labelling of multiple polyps and will allow increased automation of endoscopy reports with significant time saving.

In terms of the colonoscope itself, improvements in the characteristics of the camera such as resolution and focussing ability are constantly occurring. For example, prototypes of scopes that can autofocus during magnification have been developed, and a sharper view across the whole image field is now possible.

The use of AI technology for automated image labelling and analysis (eg, polyp characterisation) will influence decision-making during the procedure. Therefore, clear documentation of what technologies are used and their impact on the procedure will be required, particularly where this influences patient management.

Conclusion

Photodocumentation is a critical but slightly neglected part of colonoscopy practice today. We suggest a higher standard of photodocumentation is relatively easy to achieve and could positively impact cancer prevention and lead to easier implementation of an optical diagnosis, resect and discard strategy. Bodies that govern colonoscopy practice can mandate higher standards of photodocumentation but only in parallel with enhanced training and appreciation that some additional time is required during colonoscopy. High-quality photodocumentation has the potential to reduce the risk of PCCRCs and will unlock the potential of optical diagnosis to reduce the overly heavy reliability currently on histopathology. In the future, technological advances and automation will likely support the process, but the endoscopist will remain responsible for ensuring high quality and accuracy.

Ethics statements

Patient consent for publication

Not required.