HUMAN VS. MACHINE: ARE ROBOTIC NEUROINTERVENTIONS MORE PRECISE IN A CONTROLLED BENCHTOP ENVIRONMENT?

Background

Robotic endovascular treatment are now a reality, and future tools are being developed targeting neurointerventions. Although it is clear that the technology increases operator comfort and reduce their X-ray exposure, it is not certain that it achieves the same precision as manual cases. We compared manual and robotically assisted procedures in a virtual environment with new precision metrics to evaluate robotic technology using the Corpath GRX (Siemens, Erlangen, Germany).

Methods

We defined 3 tasks (easy, medium and hard) of intracranial aneurysm catheterization. We used the G5 Mentice Lab (Mentice, Göteborg, Sweden) to perform the simulated procedures. At the end of every procedure, we could collect the total translation in centimeters and of the microcatheter and the microwire. Two blinded neurointerventionists reviewed the number of re-entry into the aneurysm and catheter or wire “jump”, defined as a quick, uncontrolled movement of the tool to the aneurysm wall.

Results

We performed 68 cases: 34 robotic cases and 34 manual cases, on 3 different aneurysms virtual simulations. We had 12 operators on the easy case and 11 operators on the medium and difficult case. In the difficult case, operators used less microwire translation using the robotic assistance than performing the task manually with a mean of 38,7 cm vs. 108,4 cm, p=0.023 (Figure 1). There was no statistical difference in the easy and medium case. More than one try to catheterize the target was seen 55% of time in the difficult case done manually vs. 9% with the robotic assistance, which almost reach the statistical significance p = 0.063. Dangerous jumps were not statistically different in either 3 cases. Procedural time was not statistically different in the 2 groups.

Conclusion

Robotic procedures seem to be more precise in difficult cases and is associated with less microwire manipulations than in manual cases.