Catheter Monitoring in Intelligent Endovascular Navigation Systems: Interactive Simulations and Mixed Reality for Enhanced Navigational Awareness

Technical Summary: Catheter Monitoring in Intelligent Endovascular Navigation Systems

1. Problem Statement

Endovascular procedures (e.g., cardiac interventions) are technically demanding and rely heavily on 2D X-ray fluoroscopy for guidance. This traditional approach presents several critical limitations:

• Radiation Exposure: Both patients and operators are exposed to ionizing radiation.
• Incomplete Visualization: Fluoroscopy provides only 2D projections and suffers from "contrast washout," leading to transient and incomplete views of the vessel lumen.
• Lack of Interaction Awareness: There is a significant lack of direct, real-time visualization of the physical interaction between the catheter and the vessel wall. This makes it difficult to monitor potential vascular damage, especially in tortuous or calcified vessels, increasing the risk of complications.
• Computational Barriers: While physics-based modeling exists, the high computational cost of modeling complex catheter–vessel contact interactions has historically prevented real-time, interactive use in surgical navigation.

2. Methodology

The authors propose a framework that integrates real-time sensing, biomechanical simulation, and Mixed Reality (MR) visualization. The workflow is divided into three stages:

• Initialization:
  • Vessel Modeling: A patient-specific anatomical model is generated from CT data. Instead of modeling the vessel as an isolated tube, the authors embed the lumen within a deformable continuum (Bounding Box) to simulate surrounding soft-tissue constraints. This is discretized using a tetrahedral Finite Element Model (FEM).
  • Catheter Modeling: A geometric "avatar" of the catheter is created using a piecewise-continuous model (spheres and cylinders) to facilitate robust collision handling.
• Registration: A landmark-based Singular Value Decomposition (SVD) method is used to align the Electromagnetic (EM) sensor coordinate frame with the CT anatomical coordinate frame, achieving a Root Mean Square Error (RMSE) of 0.982 mm.
• Runtime (The Interactive Loop):
  • Sensing: A sensorized catheter equipped with Fiber Bragg Grating (FBG) and Electromagnetic (EM) sensors provides real-time 3D shape reconstruction.
  • Simulation: The reconstructed catheter shape is fed into the SOFA framework. A corotational FEM formulation with a Lagrange Multiplier (LM) collision model computes vessel deformation. The LM method is chosen for its parameter-free, physically accurate contact forces.
  • Visualization: The updated geometries are streamed via ROS (Robot Operating System) to a HoloLens 2 (HL2) headset, providing a Mixed Reality interface for the operator.

3. Key Contributions

• Novel Modeling Approach: The use of a deformable continuum (Bounding Box) to represent surrounding tissue and the use of the reconstructed catheter shape as a kinematic boundary condition (rather than a fully simulated object) significantly optimizes computational efficiency.
• Integrated MARS Framework: The study is, to the authors' knowledge, the first to successfully integrate interactive FEM, real-time catheter shape sensing, and Mixed Reality into a single endovascular navigation pipeline.
• Robust Contact Handling: By using the Lagrange Multiplier formulation, the system avoids the need to manually tune empirical parameters like contact stiffness or friction, which are difficult to estimate in vivo.

4. Results

The framework was validated in-vitro using a silicone phantom and high-speed stereo cameras.

• Interactive Performance:
  • The simulation exhibited three phases: startup latency, a low-latency phase (slope 1.12), and a high-latency phase (slope 1.45) as the catheter entered tortuous regions.
  • While the simulation lagged behind physical time (reaching a 45% delay in the most complex sections), the HoloLens 2 rendering remained stable at 35–40 FPS.
• Accuracy:
  • The median relative displacement error remained below 1 mm during early navigation and increased to 2.33 mm in the most tortuous segments.
  • Crucially, these errors remained within the cumulative uncertainty of the experimental ground-truth reconstruction (2.534 mm), validating the FEM's accuracy.
• Visual Fidelity: The FEM successfully captured the "straightening" effect of the catheter as it navigated through the initially tortuous vessel segments.

5. Significance

This research demonstrates the feasibility of using Mixed Reality Augmented Surgical (MARS) interfaces to enhance situational awareness. By providing a 3D, continuous view of catheter–vessel interactions that fluoroscopy cannot offer, the system serves as a powerful decision-support tool. While not yet a standalone guidance system due to computational latencies, it acts as an "interpretable interface" that can bridge the gap between automated robotic systems and human operators, potentially reducing vascular complications and improving the safety of endovascular interventions.