Experience Matters: Robotic-Assisted Lobectomies in the National Cancer DataBase
*Brian N. Arnold1, *Daniel C. Thomas1, *Raja Narayan2, *Justin D. Blasberg1, *Frank C. Detterbeck1, *Daniel J. Boffa1, Anthony W. Kim3
1Yale School of Medicine, New Haven, CT;2Stanford School of Medicine, Stanford, CA;3Keck School of Medicine, University of Southern California, Los Angeles, CA
Objective: Robotic-assisted lobectomy is being performed with greater frequency in the treatment of lung cancer. Increased use of the robot is assumed to be associated with improved performance. The objective of this study is to measure proficiency with robotic-assisted lobectomy.
Setting: National Cancer DataBase
Patients: Patients undergoing robotic-assisted lobectomy for lung cancer from 2010-2014.
Main Outcome Measures: Conversion to open, 90-day mortality, hospital experience with robotic lobectomy
Results: 7,645 robotic-assisted lobectomies were identified from 465 hospitals. The number of new hospitals performing robotic-assisted lobectomies decreased over time. The new hospitals in each year also trended toward lower mean volume (2.71±2.90 in 2014 vs 3.95±5.59 in 2010, p=0.1602) and were less likely to be academic hospitals (20.4% in 2014 vs. 41.4% in 2010, p=0.0226). The overall conversion rate was 9.2% (702). In a multivariate logistic regression, hospitals in their 3rd through 5th year of experience had a significantly lower conversion rate than hospitals in their first two years of experience. Increasing volume by 5 cases per year also decreased the risk of conversion (OR 0.863, CI 0.798-0.934, p=0.0002). Tumor size >3 cm was associated with an increased risk of conversion. Patients converted to open had worse 90-day mortality (8.0%) compared to totally robotic (3.4%, p<0.0001) and planned open (3.9%, p=0.0002) procedures in a propensity-matched sample.
Conclusions: Robotic-assisted lobectomy is associated with a higher rate of conversion to open lobectomy in the first two years of experience. Patients who undergo conversion have higher mortality than those that do not. Further study is needed to identify the optimal approach to the implementation of robotic technology to minimize the impact of the learning curve.
Back to 2017 Program