1. Minimally Invasive Surgery Task Decomposition - Etymology of Endoscopic Suturing Jacob Rosen* Ph.D., Lily Chang** MD, Jeff Brown ***, Andy Isch** MD Blake Hannaford* Ph.D., Mika Sinanan** MD Ph.D., Richard Satava** MD * Department of Electrical Engineering ** Department of Surgery *** Department of Bioengineering University of Washington Seattle, Washington

2. Surgery - The Big Picture Surgeon Virtual Reality Tool Tissue Reality Robot (master-slave) Haptic Device Tissue Tool Human - Machine Interface

3. Evaluation of Surgical Skill

4. Surgery - The Hidden Language Metaphor Surgeon HMI Resident Robot / Tool Target Patient Tool Operator Language

5. Surgery - The Hidden Language Elements

6. Research Aim n Create new quantitative knowledge of the forces and torques (F/T) applied by surgeons to their instruments and the positions and orientations (P/O) of the endoscopic tools during minimally invasive surgery n Develop and evaluate an objective surgical skill scale based multi finite state models (Markov Model) incorporating information regarding the kinematics and the dynamics of the endoscopic tools

7. Instrumented Endoscopic Grasper

8. Blue DRAGON

10. Blue DRAGON - Graphical User Interface

11. Experimental System PC (A/D) VCR (PIP Mode) Video Blue DRAGON Left Blue DRAGON Right Endoscope GUI Mixer Data 13 (Channels) * 2 (Dragons) = 26 Channels

12. Experimental Protocol n Surgical Tasks: n Task 1: Running the bowel, 10x3” (right to left) n Task 2: Running the bowel, 10x3” (left to right) n Task 3: Dissecting Mesenteric Vessels, 4x n Task 4: Passing a Suture, 5x n Task 5: Tying a Knot, 4x n Task 6: Suturing the Bowel and Tying a Knot Task 7: Passing Stomach Fundus behind the Esophagus, 1x n Subjects: 5x (R1, R2, R3, R4, R5, E) - 30 Subjects

14. Normalized Completion Time Learning Curve of Surgical Residents

15. Novice Expert Position (Path) - Raw Data

16. Normalized Trajectory Length Learning Curve of Surgical Residents

17. Novice Expert Forces - Raw Data

18. Novice Expert Torques - Raw Data

19. Tool/Tissue-Object Interactions - Taxonomy

20. F/T Clusters (GR-PS-SW)

21. Generalized Markov Model of MIS Surgery - One Tool

22. Generalized Markov Model of MIS Surgery - Two Tools Right Tool Left Tool

23. Laparoscopic Cholecystectomy Exposure of Cystic Duct Surgeon Non - Surgeon

26. Learning Curve - Markov Model Statistical Distance R1 R2 R3 R4 R5 E

27. Normalized Statistical Distance Learning Curve of Surgical Residents

28. Subjective Score (Video Analysis) - Criteria Suturing the Bowel and Tying a Knot n Overall performance [1-5] n Economy of movement [1-5] n Tissue handling [1-5] n Numbers of errors [No.] –Drop needle –Drop suture –Lose suture loop –Break suture –Needle injury to adjacent tissue –Inability to puncture bowel with needle

29. Normalized Subjective Score (Video Analysis) Learning Curve of Surgical Residents

30. Correlation Between Subjective and Objective Assessment of Surgical Skill

31. Conclusions / Application n Analyzing Minimally Invasive Surgery requires a synthesis between visual and haptic information. n Differences between expert and novice surgeons can be defined in terms of: –Force / Torque / Velocity signatures –State transitions / combinations –Time spent in each state –Trajectory (tool tip) length n Good correlation (R 2 =0.86) between objective (Markov Model) and subjective (video analysis scoring) assessment of surgical skill n The objective methodology for surgical skill evaluation is a modality independent. It can be applied to: –In-vivo surgical conditions –Telerobotic systems –VR haptics simulators

32. 10Q