Home > Members Area > BSG Webinars > BSG Webinar: Endoscopy – where we are now and where we might be going

BSG Webinar: Endoscopy – where we are now and where we might be going

This webinar, recorded on the 24th August 2020, featured talks on the use of AI in GI diseases, sustainability in endoscopy, and endoscopy training. The sessions were chaired by BSG VP Endoscopy Ian Penman and Consultant Gastroenterologist Srisha Hebbar.

Professor Prateek Sharma – Artificial Intelligence in Barrett’s Oesophagus (an update from the USA)

Dr Manu Nayar – Endoscopy training during COVID-19 and shape of training – How can the BSG help?

Dr Bu Hayee – Sustainability in endoscopy – green for go?


Questions from the webinar


Endoscopy training during COVID-19 and shape of training

1) Should endoscopy training be limited to a small cohort of trainees? May be controversial, but probably practical!

Ian Penman: This unlikely to be realistic given the shape of the workforce and the way care is delivered. UK trainees will still need to train in endoscopy.

2) With limited endoscopy training, can a trainee become a consultant in future without doing an advanced endoscopy fellowship?

Ian Penman: There is no reason why trainees need to undertake an advanced fellowship before becoming a consultant: such fellowships are aimed at those wishing to pursue a subspecialist interest in advanced techniques. For most, the challenge will be to maximise training opportunities to ensure they are competent and confident by the time they become consultants.

3) Should trainees be limited to only one mode of endoscopy training?

Ian Penman: This is unlikely to be attractive to the majority of trainees; or to employers, especially in smaller hospitals with fewer staff.

4) What are the future plans for the JAG training courses? When will it be commenced?

Ian Penman:  JAG are working hard with training centres on plans to resume patient-focused training courses. More information can be found on the JAG website.

5) How will surgical, paediatric and nurse Endoscopy trainees fit into the new training structure?

Ian Penman:  The Shape of Training reforms apply only to medical specialties and not other disciplines. Advanced Fellowships are locally organised and in many cases will be open to applicants from all training backgrounds.

6) Do we know roughly when this new shape of training will be implemented?

Ian Penman:  Shape of Training is intended to start in autumn 2021.

7) I presume that to set up a fellowship, the host unit has to look for funding internally rather than apply for something regionally. Is this right?

Ian Penman: There are no hard and fast rules – funding arrangements for these posts will differ from centre to centre.

8) If we lose our SpR to the fellowship programme, then we will struggle to provide other parts of medical and gastro service. I fear this will mean that small DGHs will lose (permanently) their training posts. Will it not affect the quality of care in small hospitals?

Ian Penman: Not necessarily. Local arrangements will vary from post to post, but some are post-CCT while in others, duties are split between advanced training and service provision and may include a contribution to out of hours cover etc.

9) Would the endoscopy training fellowships be open to LED / SAS doctors who want to pursue endoscopy as a career, i.e. those opting for the CESR route?

Ian Penman: This will be decided locally for each post, but in theory, there is no reason why these posts should not be available to LED/SAS doctors.

10) How will the new “shape of training” affect clinical endoscopist training?

Ian Penman: It should not have any direct effect, but there is a need for wider discussions about how we can train ALL endoscopists to the very highest levels and these discussions are underway.

Artificial Intelligence in Barrett’s Oesophagus

1) Does AI have the potential to analyse non dysplastic mucosa and predict the chance of future Cancer, and as a result reduce the need for surveillance?

Prateek Sharma: Yes the potential is there but not being tested at this time. This would require a combination of CAD assisting with the diagnosis of NDBE (computer vision) and using the patients demographic and endoscopic information (age, smoking, BE length etc) to predict future risk of cancer.

2) Do you think there will be over reliance on AI by endoscopists in future rather than focusing on improving their detection skills?

Prateek Sharma: Definitely a possibility. Our goal should be allow AI to ‘assist’ us rather than ‘replace’ our own skill set.

3) How much more time will AI add to Barrett’s surveillance compared to the present practice?

Prateek Sharma: At the start, probably 2-10 minutes depending on the lesion/length of BE. Once the system and the endoscopist are ‘trained’, it should reduce the time since random biopsies will not be required.

4) Is the supervised learning training set likely to be significantly affected by the kappa limitation (inter-observer variation) even between pathologists and are you accounting for this with independent path reviews

Prateek Sharma: Good point, the ground truth is the reading of expert pathologist (s) and more the better. It will be critical to have this in the published literature and algorithms – they use the reading of experts who agree on diagnosis rather than individual readings. Specifically for a diagnosis of LGD, it would be very problematic.

5) Should AI be introduced at an early stage of endoscopy training?

Prateek Sharma: Probably. The GI societies will have to introduce this into the endoscopy teaching curriculum. As with any new technology, there should be cognitive aspects taught as well, rather than just the endoscopy portion (which may be the easy one).

6) Will there be a post OGD UGI Cancer metric as a QA marker in Barrett’s Oesophagus?

Prateek Sharma: Ideally yes, this would be the most rigorous outcome very similar to interval colon cancer. Given the relatively lower numbers of esophageal cancer, this will be difficult to measure at an individual level. Surrogate markers such as post EGD dysplasia may be easier to measure.

7) Are these AI systems available for prime time use?

Prateek Sharma: Not yet for Barrett’s, they’re still being studied in research trials.