Back to 2016 Annual Meeting
Development and Validation of an Ex Vivo Trainer for Robotic Vesicourethral Anastomosis
Kevin Koo, MD, MPH, MPhil1, Xiaotian Wu, BEng2, Ryan J. Halter, PhD2, Fady M. Ghali, BS3, Elias S. Hyams, MD1.
1Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA, 2Thayer School of Engineering, Hanover, NH, USA, 3Geisel School of Medicine, Hanover, NH, USA.
BACKGROUND: The vesicourethral anastomosis (VUA) in robotic prostatectomy is a challenging task for novices due to delicate tissues and difficult suturing angles. Though digital simulation is useful for developing certain robotic skills, there are no viable digital models for anastomotic suturing. We developed and validated a 3D-printed model of the VUA for ex vivo training.
METHODS: VUA Model: A dome-shaped shell (12 × 12 × 5.5 cm) with a tapered opening for the bladder neck (2 cm) and urethral cylinder (0.5 cm thick, 1 cm protrusion, on a 5 cm base) were developed. The models were molded with Smooth-On Ecoflex Supersoft two-part silicone using 2-part inverse ABS plastic molds designed on SolidWorks and printed with the Stratasys Mojo Desktop 3D printer. The bladder shell and urethra were attached to a rigid but adjustable acrylic/polycarbonate frame (Figure). Subjects: Ten junior surgical residents (PGY 1-3) naive to robotic surgery were enrolled. Five completed a curriculum on the da Vinci simulator and five did not, matched by PGY. Four faculty fellowship-trained in robotic uro-oncology were enrolled. All subjects attempted a VUA on the model. Non-simulator trained subjects were given a 10-minute practice period to familiarize with the robotic controls. Percentage completion of the anastomosis within 15 minutes was recorded. Integrity of the anastomosis was graded (excellent, moderate, or poor). Face (realism) and content (training utility) validity were assessed via 1-10 scoring by subjects. Construct validity (differentiation in performance) was assessed by comparison of scores between groups.
RESULTS: Mean (range) percentage completion of the anastomosis was 20% (10-30%), 54% (40-70%), and 96% (85-100%) by non-simulator-trained, simulator-trained, and expert surgeons, respectively (p < 0.05). Integrity was similar between groups with one "poor" grade in the non-simulator-trained group. Face validity was rated 8 by all expert surgeons. Content validity was rated 10 by all subjects.
CONCLUSIONS: We demonstrated face, content, and construct validity of a 3D-printed model for the VUA. Digital simulation significantly improved trainees' performance on the ex vivo model. While digital simulation is likely to become more sophisticated, ex vivo models can be realistic and useful training tools for novice robotic surgeons prior to in vivo performance.
Back to 2016 Annual Meeting