Classifying Inflammation on Intestinal Ultrasound Images and Cineloops－A Learning Curve Study

Abstract

Abstract Background and Aims Intestinal ultrasound has become a crucial tool for assessing inflammation in patients with inflammatory bowel disease, prompting a surge in demand for trained sonographers. Whereas educational programmes exist, the length of training needed to reach proficiency in correctly classifying inflammation remains unclear. Our study addresses this gap partly by exploring the learning curves associated with the deliberate practice of sonographic disease assessment, focusing on the key disease activity parameters of bowel wall thickness, bowel wall stratification, colour Doppler signal, and inflammatory fat. Methods Totals of 21 novices and six certified intestinal ultrasound practitioners engaged in an 80-case deliberate practice online training programme. A panel of three experts independently graded ultrasound images representing various degrees of disease activity and agreed upon a consensus score. We used statistical analyses, including mixed-effects regression models, to evaluate learning trajectories. Pass/fail thresholds distinguishing novices from certified practitioners were determined through contrasting-groups analyses. Results Novices showed significant improvement in interpreting bowel wall thickness, surpassing the pass/fail threshold, and reached mastery level by Case 80. For colour Doppler signal and inflammatory fat, novices surpassed the pass/fail threshold but did not achieve mastery. Novices did not improve in assessing bowel wall stratification. Conclusions We found considerable individual- and group-level differences in learning curves, supporting the concept of competency-based training for assessing bowel wall thickness, colour Doppler signal, and inflammatory fat. However, despite practice over 80 cases, novices did not improve in their interpretation of bowel wall stratification, suggesting that a different approach is needed for this parameter.