Proctors exploit three-dimensional ghost tools during clinical-like training scenarios: a preliminary study

Abstract

Purpose: In this study, we examine three-dimensional (3D) proctoring tools (i.e., semitransparent ghost tools overlaid on the surgeon's field of view) on realistic surgical tasks. Additionally, we develop novel, quantitative measures of whether proctors exploit the additional capabilities offered by ghost tools.

Methods: Seven proctor-trainee pairs completed realistic surgical tasks such as tissue dissection and suturing in a live porcine model using 3D ghost tools on the da Vinci Xi Surgical System. The usability and effectiveness of 3D ghost tools were evaluated using objective measures of proctor performance based on proctor hand movements and button presses, as well as post-study questionnaires.

Results: Proctors exploited the capabilities of ghost tools, such as 3D hand movement (p < 0.001), wristedness (p < 0.001), finger pinch gestures (p < 0.001), and bimanual hand motions (p < 0.001). The median ghost tool excursion distances across proctors in the x-, y-, and z-directions were 57.6, 31.9, and 50.7, respectively. Proctors and trainees consistently evaluated the ghost tools as effective across multiple categories of mentoring. Trainees found ghost tools more helpful than proctors across all categories (p < 0.05).

Conclusions: Proctors exploit the augmented capabilities of 3D ghost tools during clinical-like training scenarios. Additionally, both proctors and trainees evaluated ghost tools as effective mentoring tools, thereby confirming previous studies on simple, inanimate tasks. Based on this preliminary work, advanced mentoring technologies, such as 3D ghost tools, stand to improve current telementoring and training technologies in robot-assisted minimally invasive surgery.

Keywords: Augmented reality; Ghost tools; Performance metrics; Proctor; Surgeon training; Telementoring.

Fig. 1

a Experimental setup with trainee, proctor, input devices, and three-dimensional display. b Custom wireless input device that resembles da Vinci hand controller at the surgeon console. c Commercial input device from Sixense, Inc. d Image of ghost tools being used for proctoring on a tissue task

Fig. 2

Three clinical-like training tasks on a live porcine model. a Task 1 (suturing): four interrupted sutures were thrown (two half hitches followed by a surgeon’s knot) on the surface of the sigmoid. Trainees used two large needle drivers with 3-0 vicryl (RB-1) cut to a length of 10 cm. Additional suture was supplied, if necessary. b Task 2 (dissection): the ureters were dissected by exposing a window between the medial and lateral leaves of the bladder suspensory ligament. A small grasping retractor was used in the fourth arm to retract the bladder anteriorly to place the suspensory ligaments under tension. Monopolar curved scissors and Maryland bipolar instruments were used for the dissection. c Task 3 (dissection): the left and right uterine horns were mobilized from the broad ligaments. Trainees worked medially starting at the fallopian tubes and ending at the uterine body. They readjusted their fourth arm to provide optimal retraction and used a combination of bipolar and monopolar energy. Finally, once both uterine horns were fully mobilized, trainees amputated the uterine body

Fig. 3

Data of proctor hand movements (right hand/left hand) (a time series, b three-dimensional trajectories) and hand orientation (c roll (red), pitch (green), and yaw (blue) angles)

Fig. 4

a 3D movement interquartile ranges for x-, y-, z-movement directions. b Wrist angle interquartile ranges for roll (R), pitch (P), and yaw (Y) angles. Right and left hand data combined. Horizontal black lines indicate significant differences (p < 0.05) between groups