Aims
Accurate pre-procedural lesion size estimation is crucial in endoscopic submucosal dissection (ESD) to guide treatment decisions. This study aimed to compare the accuracy of lesion size estimation by community endoscopists and hospital-based expert endoscopists against the actual lesion size measured from resected ESD specimens.
Methods
A total of 250 paired pre-procedural lesion size estimates were analyzed from community endoscopists and from hospital based expert endoscopists. These estimates were compared with the actual lesion size using Bland-Altman analysis, Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Lin’s Concordance Correlation Coefficient (CCC). Regression analysis was performed to assess proportional bias and trends over time, while ANOVA tests were used to determine whether lesion classification (Paris classification) or anatomical location influenced estimation errors.
Results
Hospital-based endoscopists demonstrated superior accuracy compared to community endoscopists, with a lower bias of 5.22 mm, narrower limits of agreement, and a higher concordance correlation coefficient (CCC) of 0.785. Lesions larger than 11.5 mm were associated with increased estimation errors, particularly in community endoscopist estimates, which exhibited significant proportional bias (p < 0.001). Among hospital-based endoscopists, neither Paris Classification nor anatomical location showed a significant association with estimation errors. However, for community endoscopists, lesions located in the greater curvature were significantly overestimated (p = 0.019), with an average overestimation of 8.50 mm. No significant improvement in estimation accuracy over time was observed, suggesting that error rates have remained stable despite potential advancements in training or experience.
Conclusions
These findings highlight the need for standardized measurement techniques and underscore the potential benefits of enhanced training programs and adjunctive imaging technologies to improve pre-procedural lesion size assessment in endoscopic practice