An Assessment of the Learning Curve Associated with Robotic-Assisted Pedicle Screw Insertion

Presented at SMISS Annual Forum 2014
By Joseph Sclafani MD
With Payam Moazzaz MD, Kevin Liang PhD, Janet Dunlap MD, Neville Alleyne MD,

Disclosures: Joseph Sclafani MD None Payam Moazzaz MD None, Kevin Liang PhD None, Janet Dunlap MD None, Neville Alleyne MD None,

Introduction:
There is increasing interest in new technologies to improve accuracy and consistency during spinal implant placement, especially in cases of complex deformity. Computer-assisted, bone-mounted robotic devices are now available to place pedicle screw instrumentation according to a preoperatively planned trajectory (Renaissance, Mazor Robotics Ltd., Israel). Several studies have reported accurate placement of instrumentation and decreased radiation exposure to the operating surgeon and surgical staff with this technology. However, there is a learning curve associated with new operative techniques that has not been well studied with robotic-assisted pedicle screw placement.

Aims/Objectives:
This study is an assessment of the initial learning curve experienced by three surgeons during their early experience with robotic-assisted pedicle screw placement.

Methods:
Overall screw insertion time, complication rate, and an intraoperative assessment of accuracy were collected during the initial 30 case experiences of robotic-assisted pedicle screw placement by three surgeons. Accuracy was assessed based on intraoperative imaging and electrical resistance across the screw shank after placement. Screw insertion time was pooled for all surgeons based on chronological case number. Statistical analyses were performed via one way analysis of variance with significance defined as p<0.05.

Results:
The mean screw insertion time decreased with chronologic case number (case 1-5= 15.9 minutes, case 6-10= 10.6 minutes, case 11-15= 9.5 minutes, case 16-20= 9.0 minutes, case 21-25= 8.4 minutes, case 26-30= 6.2 minutes, p= 0.0003). Screw insertion time approached an asymptote at approximately chronologic case #20 for all three learning surgeons. The accuracy of pedicle screw placement was maintained above 98% across all chronologic case groups (p>0.05). There were no intraoperative complications or aborted cases recorded.

Conclusions:
This study demonstrates that the learning curve of robotic-assisted pedicle screw insertion is overcome within the initial 20 cases. The main parameter affected during the learning curve is overall screw insertion time. Screw accuracy was consistent with previously published data and did not suffer with decreased pedicle screw insertion times. Future identification of key obstacles encountered during the learning surgeon’s initial case experience will direct development of strategies to shorten the learning curve and improve the adoption rate of robotic-assisted implant placement.