Navigation with Electromagnetic Tracking for Interventional Radiology Procedures: A Feasibility Study

Bradford J. Wood, Hui Zhang, Amir K. Durrani, Neil Glossop, Sohan Ranjan, David Lindisch, Eliott Levy, Filip Banovac, J. Borgert, Sascha Krueger, Jochen Kruecker, Anand Viswanathan, Kevin ClearyJournal of Vascular and Interventional Radiology 2005 · DOI

TL;DR

This paper explores the feasibility of using electromagnetic tracking (EMT) to guide interventional radiology procedures by integrating preprocedural imaging (CT, MR, PET) with real-time device tracking. The study demonstrates that EMT can accurately display device positions within these preprocedural datasets, paving the way for more precise and potentially less invasive interventions.

What Problem Does This Paper Address?

Traditional interventional radiology relies heavily on real-time imaging like fluoroscopy, ultrasound, or CT to guide procedures. While effective, this approach often requires physicians to mentally register anatomical information from other modalities (like MRI or PET) to the live imaging, which can be inexact and prone to error. This mental registration is particularly challenging for complex procedures where precise navigation is critical for patient outcomes. The paper addresses the need for a system that can seamlessly integrate preprocedural, often functional or higher-resolution, imaging with the real-time guidance of interventional tools, reducing reliance on continuous radiation exposure and improving spatial awareness.

Methods

The core of the study involves developing and testing an image-guided intervention software system. Key methodological components include:

  • Electromagnetic Tracking System: A tetrahedral-shaped electromagnetic field generator (Aurora) was used to create a tracking volume. Tiny sensor coils embedded in the tips of catheters, needles, and guide wires detected the magnetic field, allowing for real-time tracking of their position and orientation.
  • Custom Software Development: An open-source software platform (IGBiopsy) was built, integrating components for image handling (DICOM reader), visualization (VTK), tracking, and registration (ITK). This software allowed for loading preprocedural CT, MR, and PET datasets.
  • Registration: The process of aligning the electromagnetic tracking space with the imaging space was crucial. This involved identifying fiducial markers (either external adhesive markers or internal anatomical landmarks) on both the patient/phantom and the preprocedural images.
  • Navigation and Display: Once registered, the software displayed the tracked device's position overlaid on the preprocedural imaging datasets in real-time. This allowed physicians to navigate instruments with guidance from the previously acquired images, supplemented by simultaneous angiography.
  • Phantom and Animal Models: The system was tested on both a vascular phantom and three pigs to evaluate its feasibility and accuracy in simulated and in vivo environments.

Datasets

The study utilized preprocedural imaging datasets from:

  • Computed Tomography (CT): Used for both phantom and animal models. CT scanners from multiple vendors (Philips, GE, Siemens) were employed.
  • Positron Emission Tomography (PET): Used in phantom studies to demonstrate navigation to metabolically active regions.
  • Magnetic Resonance (MR): Used in phantom studies for navigation.

No new benchmarks were introduced; rather, the study focused on integrating existing imaging modalities with the EMT system.

Key Insights and Ideas

  • Bridging the Gap: The most significant insight is the successful demonstration that preprocedural CT, MR, and PET data can be accurately co-displayed with real-time device tracking. This effectively brings high-quality, offline imaging into the procedure room for guidance.
  • Reduced Radiation/Contrast: By using preprocedural imaging for navigation, the system has the potential to significantly reduce the need for continuous fluoroscopy and contrast agents, leading to lower radiation doses and reduced risks for patients.
  • Multimodality Integration: The system's ability to integrate data from different imaging modalities (CT, MR, PET) with real-time tracking opens doors for "multimodality interventions," allowing for more comprehensive guidance.
  • Functional Imaging for Biopsy: The study highlights the potential of using PET data to guide percutaneous biopsies to the most metabolically active regions of a tumor, which could be crucial for targeted therapy.
  • Miniaturization is Key: The paper emphasizes that recent miniaturization of EMT sensor coils is critical, enabling their integration into small interventional devices like guide wires and needles, which was not possible with earlier, larger sensors.

Key Results

  • Accuracy: The system demonstrated manageable error for some applications, with registration error and tracking error reported as less than 5 mm. In phantom studies, average tracked-to-imaged error was 2.5 mm with external fiducials and 1.07 mm with endovascular fiducials.
  • Successful Navigation: Catheters and guide wires were successfully navigated in pigs to various anatomical locations (inferior vena cava, aorta, renal artery, etc.) with angiography confirming accurate selective catheterization.
  • Registration Efficiency: Registration of the patient to preprocedural images could be achieved in a matter of minutes, with actual device positions being located in as little as 35 seconds.
  • Metal Interference: While large metallic objects can cause distortion, the study found that interference was minimal at relevant distances (>10 cm) from the field generator, and turning on the X-ray tube did not affect tracking data.

Limitations and Future Work

  • Respiratory Motion: The study acknowledges that respiratory motion, organ shift, and patient motion remain significant obstacles to routine clinical deployment. The current system did not incorporate respiratory correction or gating.
  • Subjective Accuracy: In pig studies, error estimation was subjective, relying on visual comparison of preprocedural CT with procedural angiography.
  • Fiducial Placement: The selection and placement of fiducials, especially internal ones, can be challenging and may affect accuracy.
  • Operator Time and Cost: Issues like operator time, software setup, automation, integration with existing systems, and cost-effectiveness are noted as unresolved challenges for widespread clinical adoption.
  • Future Work: The authors suggest that image-based registration algorithms could be further developed to automate and speed up the process. Exploring alternate field generator configurations to improve the field of view and reduce inhomogeneities is also a potential area for future research. Further validation in more complex clinical scenarios and with larger patient cohorts is implied.

Summary updated

BibTeX
@article{wood2005navigation,
  author = {Bradford J. Wood and Hui Zhang and Amir K. Durrani and Neil Glossop and Sohan Ranjan and David Lindisch and Eliott Levy and Filip Banovac and J. Borgert and Sascha Krueger and Jochen Kruecker and Anand Viswanathan and Kevin Cleary},
  title = {{Navigation with Electromagnetic Tracking for Interventional Radiology Procedures: A Feasibility Study}},
  journal = {Journal of Vascular and Interventional Radiology},
  year = {2005},
  doi = {10.1097/01.rvi.0000148827.62296.b4},
}
Figure