An Artificial Intelligence-Aided Robotic Platform for Ultrasound-Guided Transcarotid Revascularization

Transcarotid Artery Revascularization (TCAR) is typically performed by manual catheter insertion and implies radiation exposure for both the patient and the surgeon. Taking advantage from robotics and artificial intelligence (AI), this letter presents a robotic ultrasound (RUS) platform for improving the procedure. To this purpose, ultrasound (US) imaging is considered both in the pre-operative stage for procedure planning and in the intra-operative stage to track a catheter. 3D vascular volumes can be precisely reconstructed from sequences of 2D images exploiting robotic probe manipulation and AI-based image analysis. The method proved a median reconstruction error lower than 1 mm. Pre-operative information are mapped to the intra-operative scenario thanks to a US-based registration routine. The automatic probe alignment on the target vessel demonstrated to be as precise as 0.84 & DEG;. The reconstructed 3D model can be exploited to automatically generate a catheter trajectory based on user inputs. Such trajectory enabled automatic insertion of a magnetic catheter steered by an external permanent magnet actuated by the RUS platform. Our results demonstrate a catheter tip target reaching error of 3.3 mm. We believe that these results can open the way for the introduction of robotics and AI in TCAR procedures enabling precise and automatic small-scale intravascular devices control.

Automated summary

This study presents a novel method using a robotic ultrasound (RUS) platform for Transcarotid Artery Revascularization (TCAR) procedures. It utilizes robotic probe manipulation and AI-based image analysis to accurately reconstruct blood vessels and automatically insert a magnetic catheter. The results demonstrate the potential of integrating robotics and AI into TCAR procedures for precise and automatic control of intravascular devices.