Bifurcation Identification for Ultrasound-driven Robotic Cannulation

TL;DR

BIFURC addresses the critical need for rapid, safe intravascular access by autonomously identifying vessel bifurcations and optimal needle insertion sites from ultrasound data in a robotic cannulation system. The method blends RESUS-based vessel segmentation with an erosion-detection-tracking-merging pipeline to reconstruct vessel centerlines and locate bifurcations using temporal information from multiple robot poses, followed by a rule-based selection of the insertion point. Validated on a medical phantom and six live-pig experiments, the approach achieves expert-aligned bifurcation localization and insertion-site placement with low latency, while outperforming alternative segmentation strategies in this setting. The work demonstrates that integrating expert heuristics with deep learning can enable near-expert autonomous performance in ultrasound-guided cannulation, potentially reducing procedure time and reliance on specialized personnel, with future work aimed at generalization to multiclass vessel typing, broader anatomy, and movement/deformation handling.

Abstract

In trauma and critical care settings, rapid and precise intravascular access is key to patients' survival. Our research aims at ensuring this access, even when skilled medical personnel are not readily available. Vessel bifurcations are anatomical landmarks that can guide the safe placement of catheters or needles during medical procedures. Although ultrasound is advantageous in navigating anatomical landmarks in emergency scenarios due to its portability and safety, to our knowledge no existing algorithm can autonomously extract vessel bifurcations using ultrasound images. This is primarily due to the limited availability of ground truth data, in particular, data from live subjects, needed for training and validating reliable models. Researchers often resort to using data from anatomical phantoms or simulations. We introduce BIFURC, Bifurcation Identification for Ultrasound-driven Robot Cannulation, a novel algorithm that identifies vessel bifurcations and provides optimal needle insertion sites for an autonomous robotic cannulation system. BIFURC integrates expert knowledge with deep learning techniques to efficiently detect vessel bifurcations within the femoral region and can be trained on a limited amount of in-vivo data. We evaluated our algorithm using a medical phantom as well as real-world experiments involving live pigs. In all cases, BIFURC consistently identified bifurcation points and needle insertion locations in alignment with those identified by expert clinicians.

Figures (4)

• Figure 1: Left: Center lines of vessels in a real-world experiment. Without the BIFURC algorithm, the presence of noise greatly hampers the identification of individual vessels and subsequent detection of bifurcations. In contrast, data processed with BIFURC (right) yields a clearer depiction of distinct vessels and the relevant bifurcation, denoted by the red circle. Vessel IDs and colors have been chosen arbitrarily.
• Figure 2: BIFURC is a deep learning technique augmented with expert-derived heuristics designed to identify bifurcations and optimal needle insertion sites. First, the robot scans the leg and collects 2D ultrasound images alongside their poses. We then utilize a model to segment the vessels from these images. Next, we apply an erosion algorithm to distinguish vessels with overlapping segmentation masks. Using robot poses and vessel centerlines, we apply heuristics to track and merge distinct segments to identify vessel bifurcations. Finally, we locate a safe needle insertion spot, which is at least 2cm away from the identified bifurcation.
• Figure 3: Ultrasound data from pigs and phantoms. The phantom images are cleaner, with vessels appearing well-separated and elliptical. In contrast, the pig images are noisy with asymmetric vessels situated closer together. In the segmentation, these vessels may appear merged, which could be mistaken for a bifurcation if not examined thoroughly.
• Figure 4: Robotic ultrasound scanning system is shown with a needle insertion mechanism attached to the end-effector of a 6-DOF Universal Robots UR3e Serial Manipulator.