Mo1498 Learning Curve in Diagnosing Dysplasia in Barrett's Esophagus (BE) Using Probe-Based Confocal LASER Endomicroscopy (pCLE): Results From a Randomized Controlled Trial

Abstract

Mo1498 Learning Curve in Diagnosing Dysplasia in Barrett’s Esophagus (BE) Using Probe-Based Confocal LASER Endomicroscopy (pCLE): Results From a Randomized Controlled Trial Prashanth Vennalaganti*, Kavous Pakseresht, Benjamin R. Alsop, Sharad C. Mathur, Sravanthi Parasa, Benjamin Hornung, Neil Gupta, Prateek Sharma Gastroenterology and Hepatology, Veterans Affairs Medical Center, Kansas City, MO; Gastroenterology and Hepatology, University of Kansas Medical Center, Kansas City, KS; Gastroenterology, Loyola University Medical Center, Chicago, IL Background: With the advent of multiple novel techniques, optimal methods to teach advanced imaging techniques like pCLE have not yet been developed. There are limited data regarding the learning curve for trainees for using pCLE in diagnosing BE associated neoplasia. Aim: To determine the learning curve of medical trainees in diagnosing BE associated neoplasia using pCLE. Methods: Videos from a previously conducted multicenter trial evaluating the utility of pCLE in BE were used for this study. Study participants (GI fellows and medical students) with no previous pCLE experience were randomized to either in-class didactic or selfdirected teaching groups stratified by level of training. A standardized power point presentation was developed explaining the technology, the disease state and using previously published validated criteria for diagnosing dysplasia in BE: 1)epithelial surface: saw-toothed 2)goblet cells: sparse 3)glands: not equidistant 4)glands: unequal in size and shape 5)cells: enlarged 6)cells: pleomorphic. For in-class didactic group, an expert in pCLE conducted a classroom teaching session with an opportunity for interaction. Every study participant in the self-directed teaching group was provided with the same teaching power point but with recorded audio explaining the pCLE features to distinguish dysplastic and non-dysplastic BE. Following this initial training, all participants completed 1st set of 20 pCLE videos (pre-feedback) after which they reviewed correct responses with either the expert (in-class didactic group) or on audio power point (self teaching group). Finally, all participants completed next set of 40 videos (post-feedback) and marked their level of confidence in video interpretation for all assessments. Fisher’s exact test was used to calculate statistical differences between groups. Results:60 pCLE videos (nondysplastic BE 40 and dysplastic 20) were assessed by 18 study participants (10 GI fellows and 8 students). The overall accuracy of the entire group significantly improved from 73% to 79% post feedback (pZ0.04) (Table). This improvement was driven by overall improvement post feedback for the in-class didactic group (from 74% to 84%; p!0.01). Students had improvement in the proportion of predictions with high confidence (from 47% to 54.4% p 0.1) while fellows improved their pCLE accuracy in high-confidence predictions (from 82% to 91%, p 0.02) (Table). Finally, accuracy improved with years in endoscopy training (72.08% Students, 77.08% FY1, 82% FY2 and 85% FY3, p 0.003). Conclusions: Trainees taught by an in-class, didactic teaching session (overall 1 hour) had a shorter learning curve and achieved higher accuracies compared to self-directed learning group for the use of pCLE to diagnose BE associated neoplasia. In-class didactic teaching sessions should be the preferred method to teach pCLE for use in BE.