Endoscopy 2021; 53(S 01): S76
DOI: 10.1055/s-0041-1724441
Abstracts | ESGE Days
ESGE Days 2021 Oral presentations
Saturday, 27 March 2021 11:00 – 11:45 Resect and discard for diminutive polyps: Where do we stand? Room 6

Learning Curve Of Optical Diagnosis With A Resect And Discard Strategy For Screening Colonoscopy: Preliminary Results From The Discard3 Study

A Ahmad
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
A Wilson
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
S Thomas-Gibson
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
N Suzuki
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
A Humphries
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
A Haycock
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
K Monahan
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
M Vance
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
A Dhillon
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
,
B Saunders
1   Wolfson Unit for Endoscopy, Imperial College London, London, United Kingdom
› Author Affiliations
 
 

    Aims Advanced endoscopic imaging allows optical diagnosis of diminutive polyps. Performance thresholds for optical diagnosis have been set to allow a “resect and discard strategy” which may streamline patient care and reduce histopathology need. Aim: to assess early implementation and quality assurance of this strategy in a bowel cancer screening unit setting (1.1 million population).

    Methods In this prospective feasibility study, 8 bowel cancer screening colonoscopists completed a validated optical diagnosis training module and performed procedures at St Mark’s Hospital, London (Feb-Oct 2020). All ≤5mm polyps had white-light and narrow-band images. Endoscopists classified the optical diagnosis as high or low confidence. All polyps had standard histopathology.

    Performance was analysed in 2 month time periods with active feedback given. There was repeat histology review of inaccurate high confidence optical diagnoses.

    Results In the 8 months to date, 639 polyps were optically diagnosed (468 high confidence and 171 low confidence) from 283 patients who were divided into 4 time periods.

    The proportion of low confidence diagnoses increased from 15 % in periods 1-2 to 29 % in period 4.

    In periods 1-2, 1/8 endoscopists achieved the 90 % accuracy threshold. In periods 3 and 4, 2/8 and 4/8 endoscopists achieved the threshold respectively. Overall, accurate diagnoses increased from 81 % to 86 % (period 1 to 4). Of those who reached the threshold by period 4 the mean number of high confidence calls was 60 (range 41-99).

    For periods 1-4, the positive and negative predictive values for optical diagnosis of adenomas were 89.3 % and 64.4 % respectively:

    Tab. 1

    Histology

    Total

    Adenoma

    Non-adenoma

    Optical diagnosis

    Adenoma

    275

    57

    332

    Non-adenoma

    33

    103

    136

    Total

    308

    160

    468

    Conclusions The learning curve for optical diagnosis in a bowel cancer screening setting varies between individual operators. Over time, and with feedback, there is a gradual increase in optical diagnosis accuracy and in the number of colonoscopists achieving the 90 % threshold accuracy target, with a corresponding decrease in the percentage of high confidence calls.

    Citation: Ahmad A, Wilson A, Thomas-Gibson S et al. OP184 LEARNING CURVE OF OPTICAL DIAGNOSIS WITH A RESECT AND DISCARD STRATEGY FOR SCREENING COLONOSCOPY: PRELIMINARY RESULTS FROM THE DISCARD3 STUD. Endoscopy 2021; 53: S76.


    #

    Publication History

    Article published online:
    19 March 2021

    © 2021. European Society of Gastrointestinal Endoscopy. All rights reserved.

    Georg Thieme Verlag KG
    Rüdigerstraße 14, 70469 Stuttgart, Germany