potentially be enhanced through repeated trials with workshops to build consensus on scoring standardization. Source of Funding: Blue Cross Blue Shield of Michigan PD46-04 VIDEO ANALYSIS OF SKILL AND TECHNIQUE (VAST): MACHINE LEARNING TO ASSESS SURGEONS PERFORMING ROBOTIC PROSTATECTOMY Khurshid R. Ghani*, Yunfan Liu, Hei Law, David C. Miller, James Montie, Jia Deng, for the Michigan Urological Surgery Improvement Collaborative, Ann Arbor, MI INTRODUCTION AND OBJECTIVES: A surgeon’s technical skill may be a major determinant of patient outcomes. Because robotic surgery can be recorded, computer-vision video analysis of skill and technique (VAST) methods may have advantages for assessment that is objective and scalable. To test the hypothesis that specific features in a video can categorize skill, we studied crowdsourced annotated videos of surgeons performing robotic prostatectomy and applied machine learning to determine skill. METHODS: Videos of the anastomosis from 12 surgeons in the Michigan Urological Surgery Improvement Collaborative underwent blinded review by 25 peer surgeons using the Global Evaluative Assessment of Robotic Skills (GEARS) tool (max score 25). Surgeons were categorized into low and high skill based on 0 bimanual dexterity 0 and 0 efficiency 0 . Robotic instruments were annotated by crowdworkers via a custom-designed Mechanical Turk platform. Using the videos we trained a linear support vector machine (SVM), sampling consecutive frames to study VAST metrics for instruments including velocity, tra- jectory, smoothness of movement, and relationship to contralateral in- strument. We applied the SVM to learn and classify videos into high/low skill. To evaluate performance we used 11 videos as training, and tested on the remaining 1 video, repeating it 12 times and averaged the accuracy. RESULTS: GEARS scores ranged from 15.75 to 23.11, with 9 and 3 surgeons categorized into high and low skill, respectively. In total, 146,309 video frames were annotated by 925 crowdworkers. Instrument annotation included individual points as well as wristed joint movement (Figure). SVM accuracy in skill categorization using individual points on an instrument was 83.3%. Accuracy improved to 91.7% when we assessed joint movement. When we combined assessment with the contralateral instrument, accuracy was 100% in categorizing binary skill level. Instrument metrics most closely related to skill prediction were relationship between needle driver forceps and joint, acceleration, and velocity. CONCLUSIONS: Computer video analysis can be used to predict skill in practicing robotic surgeons. In the future, methods uti- lizing deep learning to track instruments and calculate skill, may have significant implications for credentialing and quality improvement. Source of Funding: Blue Cross Blue Shield of Michigan; Intuitive Surgical PD46-05 A RANDOMISED CONTROLLED TRIAL OF COGNITIVE TRAINING FOR TECHNICAL AND NON-TECHNICAL SKILLS IN ROBOTIC SURGERY Nicholas Raison*, Kamran Ahmed, Takashige Abe, Abdullatif Aydin, Oliver Brunckhorst, Haleema Aya, Husnain Iqbal, David Eldred-Evans, London, United Kingdom; Andrea Gavazzi, Florence, Italy; Giacomo Novara, Padua, Italy; Nicolo Buffi, Milan, Italy; Ben Challacombe, London, United Kingdom; Craig McIlhenny, Larbert, United Kingdom; Shamim Khan, London, United Kingdom; Henk Van Der Poel, Amsterdam, Netherlands; Prokar Dasgupta, London, United Kingdom INTRODUCTION AND OBJECTIVES: Cognitive training tech- niques such as mental imagery (MI) have been successfully used as training aids for sport, music and rehabilitation medicine. By stimulating similar neural pathways to motor tasks, MI practice leads to improved motor performance. Studies have shown that MI may be effective in surgery although so far this has been limited to laparoscopic technical skills training. Given the unique training challenges posed by robotic surgery, the potential for MI to supplement training outside of the costly and stressful operating room environment is considerable. This studies aims to establish the feasibility of cognitive training for technical and non-technical skills training in robotic surgery. METHODS: A double blind, randomised controlled trial of 61 robotic novices was performed. ISRCTN registry ID ISRCTN47552076. All participants underwent initial basic robotic skills training using a ro- botic virtual reality simulator. Baseline ability was recorded. Participants were randomised to either MI or standard training in robotic technical and NTS skills. Participants performed 3 dry-lab warm-up exercises before completing a urethrovesical anastomosis (UVA) within a simu- lated operating room environment. Alongside completion of the UVA task, subjects were required to manage 3 NTS scenarios. Perfor- mances were video-recorded and analysed post hoc by blinded, expert robotic surgeons. Technical skills were assessed using GEARS and NOTSS was used for NTS. RESULTS: 28 subjects underwent cognitive training and 33 un- derwent standard training. No significant differences in surgical experi- ence or baseline ability between the 2 study groups. Cognitive training resulted in a significantly better technical performance compared to standard training (total GEARS score 13.37 vs 10.94, p ¼ 0.007). No difference was seen in NTS performance (mean total NOTSS score for MI and standard training respectively 23.5 vs 27.0 p ¼ 0.18). CONCLUSIONS: This RCT provides strong evidence for the role of cognitive training in technical skills training in robotic surgery but not in NTS training. Further assessment of cognitive training in more experienced robotic surgeons is now required to determine the optimal integration of cognitive training into the robotic surgical curriculum. Vol. 197, No. 4S, Supplement, Sunday, May 14, 2017 THE JOURNAL OF UROLOGY â e891