Evaluation of Robotic-Assisted Carotid Artery Stenting in a Virtual Model Using Motion-Based Performance Metrics.

PURPOSE: Robotic-assisted carotid artery stenting (CAS) cases have been demonstrated with promising results. However, no quantitative measurements have been made to compare manual with robotic-assisted CAS. This study aims to quantify surgical performance using tool tip kinematic data and metrics of precision during CAS with manual and robotic control in an ex vivo model. MATERIALS AND METHODS: Transfemoral CAS cases were performed in a high-fidelity endovascular simulator. Participants completed cases with manual and robotic techniques in 2 different carotid anatomies in random order. C-arm angulations, table position, and endovascular devices were standardized. Endovascular tool tip kinematic data were extracted. We calculated the spectral arc length (SPARC), average velocity, and idle time during navigation in the common carotid artery and lesion crossing. Procedural time, fluoroscopy time, movements of the deployed filter wire, precision of stent, and balloon positioning were recorded. Data were analyzed and compared between the 2 modalities. RESULTS: Ten participants performed 40 CAS cases with a procedural success of 100% and 0% residual stenosis. The median procedural time was significantly higher during the robotic-assisted cases (seconds, median [interquartile range, IQR]: 128 [49.5] and 161.5 [62.5], p=0.02). Fluoroscopy time differed significantly between manual and robotic-assisted procedures (seconds, median [IQR]: 81.5 [32] and 98.5 [39.5], p=0.1). Movement of the deployed filter wire did not show significant difference between manual and robotic interventions (mm, median [IQR]: 13 [10.5] and 12.5 [11], p=0.5). The postdilation balloon exceeded the margin of the stent with a median of 2 [1] mm in both groups. Navigation with robotic assistance showed significantly lower SPARC values (-5.78±3.14 and -8.63±3.98, p=0.04) and higher idle time values (8.92±8.71 and 3.47±3.9, p=0.02) than those performed manually. CONCLUSIONS: Robotic-assisted and manual CAS cases are comparable in the precision of stent and balloon positioning. Navigation in the carotid artery is associated with smoother motion and higher idle time values. These findings highlight the accuracy and the motion stabilizing capability of the endovascular robotic system. CLINICAL IMPACT: Robotic assistance in the treatment of peripheral vascular disease is an emerging field and may be a tool for radiation protection and the geographic distribution of endovascular interventions in the future. This preclinical study compares the characteristics of manual and robotic-assisted carotid stenting (CAS). Our results highlight, that robotic-assisted CAS is associated with precise navigation and device positioning, and smoother navigation compared to manual CAS.

In the Fundamentals of Endovascular and Vascular Surgery model motion metrics reliably differentiate competency.

OBJECTIVE: The Fundamentals of Endovascular and Vascular Surgery, a curriculum that includes an endovascular model for skills testing, aims to differentiate between competent and noncompetent performers. The aim of our study was to further validate the model and to test its reliability in assessing the performance of endovascular trainees in an uncontrolled setting. METHODS: The model was tested exclusively in a virtual reality environment. On the basis of their endovascular experience, 52 participants were divided into three groups: novice (<50 endovascular cases), intermediate (50-500 endovascular cases), and expert (>500 endovascular cases). Performance was evaluated in four tasks, measuring the tool tip position and velocity on the virtual model. Average tool tip velocity and movement smoothness in the velocity frequency domain are validated parameters defining proficiency of movement. The data were filtered and interpolated to calculate the metrics. Trials containing critical tool manipulation errors were excluded. RESULTS: In total, 52 tasks completed by novices, 25 completed by intermediates, and 38 completed by experts were analyzed to determine performance. The difference in performance between the novice and expert groups was statistically significant for guidewire smoothness (P < .001). The expert group had a statistically significantly higher average guidewire velocity compared with the novice group (P < .001). CONCLUSIONS: The Fundamentals of Endovascular and Vascular Surgery model continues to differentiate novices from experts on the basis of their handling of guidewire and catheter tools, measured as smoothness and velocity. This model offers a useful instrument to test competency of endovascular surgeons.

Comparing Manual and Robotic-Assisted Carotid Artery Stenting Using Motion-Based Performance Metrics.

Carotid artery stenting (CAS) is a minimally invasive endovascular procedure used to treat carotid artery disease and is an alternative treatment option for carotid artery stenosis. Robotic assistance is becoming increasingly widespread in these procedures and can provide potential benefits over manual intervention, including decreasing peri- and post-operative risks associated with CAS. However, the benefits of robotic assistance in CAS procedures have not been quantitatively verified at the level of surgical tool motions. In this work, we compare manual and robot-assisted navigation in CAS procedures using performance metrics that reliably indicate surgical navigation proficiency. After extracting guidewire tip motion profiles from recorded procedure videos, we computed spectral arc length (SPARC), a frequency-domain metric of movement smoothness, average guidewire velocity, and amount of idle tool motion (idle time) for a set of CAS procedures performed on a commercial endovascular surgical simulator. We analyzed the metrics for two procedural steps that influence post-operative outcomes. Our results indicate that during advancement of the sheath to the distal common carotid artery, there are significant differences in SPARC (F(1, 22.3) = 6.12, p = .021) and idle time (F(1, 22.6) = 6.26, p = .02) between manual and robot-assisted navigation, as well as a general trend of lower SPARC, lower average velocity, and higher idle time values associated with robot-assisted navigation for both procedural steps. Our findings indicate that significant differences exist between manual and robot-assisted CAS procedures. These are quantitatively detectable at the granular-level of physical tool motion, improving the ability to evaluate robotic assistance as it grows in clinical use.