PTU-012 Learning curve for optical diagnosis of colorectal polyps using cumulative sum analysis

    loading  Checking for direct PDF access through Ovid



Optical diagnosis (OD) for diminutive and small colorectal polyps is an attractive option to reduce costs and streamline patient care. The American Society of Gastrointestinal Endoscopy Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI) established a 90% diagnostic threshold for real time endoscopic assessment of diminutive colorectal polyps. The learning curve for trainees to achieve the competency has not been fully explored. Our aim is to evaluate the minimum number of polyps to achieve and maintain the OD thresholds per PIVI standards using an upward CUSUM plot.


Four trainees without previous experience in OD at our institution participated in this prospective study. Four weeks before the commencement of the study they were given a training module on OD. OD was based on NICE and WASP classification. During the study period (January 2016-August 2016), each trainee documented the OD of polyps less than 10 mm in size. Patient demographics, polyp details and polyp histology were collected prospectively. OD of each polyp was compared against the polyp histology.


A total of 708 polyp observations were performed by trainees during the study period. Total number of adenomas, hyperplastic polyps and sessile serrated adenomas/polyps (SSA/P) were 3 64 214 and 52 respectively. All 4 trainees achieved sustained accuracy (90% threshold) in OD within 12–58 observations. Image 1 illustrates the upward CUSUM plot of 4 trainees. The number of polyps required to reach the plateau varied between 12 to 58. Every trainee’s confidence level improved over time (from 69% to 89%) and the effect was augmented by in-vivo feedback and revision of training module. Table 1 summarises the optical diagnostic performance of all 4 trainees. Negative predictive value for adenomas were above 90% for all trainees.


The CUSUM scores of all 4 trainees in the study reached the PIVI standards plateau by the 58th polyp observation. In-vivo feedback and continued training appears important to maintain the performance. Our preliminary findings could be used as a guide to plan the certification process for implementation of optical diagnosis.

Disclosure of Interest

None Declared

Related Topics

    loading  Loading Related Articles