Pd27-11 Inter-Dependencies Between Suturing Skills to Improve Automatic Surgical Skill Assessment

You have accessJournal of UrologySurgical Technology & Simulation: Artificial Intelligence II (PD27)1 May 2024PD27-11 INTER-DEPENDENCIES BETWEEN SUTURING SKILLS TO IMPROVE AUTOMATIC SURGICAL SKILL ASSESSMENT Zijun Cui, Runzhuo Ma, Cherine H. Yang, Anand Malpani, Timothy N. Chu, Eman Dadashian, Ahmed Ghazi, John W. Davis, Brian J. Miles, Clayton Lau, Yan Liu, and Andrew J. Hung Zijun CuiZijun Cui , Runzhuo MaRunzhuo Ma , Cherine H. YangCherine H. Yang , Anand MalpaniAnand Malpani , Timothy N. ChuTimothy N. Chu , Eman DadashianEman Dadashian , Ahmed GhaziAhmed Ghazi , John W. DavisJohn W. Davis , Brian J. MilesBrian J. Miles , Clayton LauClayton Lau , Yan LiuYan Liu , and Andrew J. HungAndrew J. Hung View All Author Informationhttps://doi.org/10.1097/01.JU.0001008580.58088.27.11AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Suturing can be broken down into several sequential skills (e.g., needle holding, needle driving, etc., Figure 1). Significant inter-dependencies among these skills have been established by previous studies (e.g., when needle positioning is ideal, needle entry angle has a greater chance of being ideal). In this work, we aim to utilize the inter-dependent nature of skills to enhance automated skill assessment with a machine learning (ML) model. METHODS: A total of 156 virtual reality (VR) videos on the Surgical Science FlexVR™ simulator were collected from 43 surgeons across 5 centers. Kinematic data (XYZ coordinates of the pose of each instrument) was captured. Videos were segmented into sub-phases, encompassing six skills (Figure 1a). Six independent raters provided binary assessment labels for each skill (low vs. high) using the End-To-End Assessment of Suturing Expertise (EASE) system (Figure 1 caption) after standardized training. First, every skill's features were extracted from RGB (colored pixel changes) and optical flow (motion directionality) in video frames using a pre-trained ML algorithm (ConvLSTM). Then features for the six skills were learned simultaneously using the skill inter-dependencies through an ML algorithm (graph attention network (GAT)). Kinematic data (instrument motion tracking) was used as an additional input for the GAT. These learned features were utilized to perform skill assessment classification. RESULTS: Automating the assessment of most skills benefited from using known inter-dependencies, compared to the assessment without inter-dependencies (Figure 1b). One exception was Hold Angle, in which the best performance was achieved without inter-dependencies (AUC 0.53). When available, kinematic data helps further improve the model's assessing performance. One exception was Driving smoothness, in which the best performance was achieved without kinematic data (AUC 0.88). CONCLUSIONS: This study demonstrated the utilization of suturing skill inter-dependencies in improving the performance of a model to automate skill assessment. Leveraging instrument kinematic data can further improve the performance of the skill inter-dependency model. Download PPT Source of Funding: Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number R01CA251579. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health © 2024 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 211Issue 5SMay 2024Page: e555 Advertisement Copyright & Permissions© 2024 by American Urological Association Education and Research, Inc.Metrics Author Information Zijun Cui More articles by this author Runzhuo Ma More articles by this author Cherine H. Yang More articles by this author Anand Malpani More articles by this author Timothy N. Chu More articles by this author Eman Dadashian More articles by this author Ahmed Ghazi More articles by this author John W. Davis More articles by this author Brian J. Miles More articles by this author Clayton Lau More articles by this author Yan Liu More articles by this author Andrew J. Hung More articles by this author Expand All Advertisement PDF downloadLoading ...