Electromagnetic Tracking for Thermal Ablation and Biopsy Guidance: Clinical Evaluation of Spatial Accuracy

Jochen Krücker, Sheng Xu, Neil Glossop, Anand Viswanathan, Jörn Borgert, Heinrich Schulz, Bradford J. WoodJournal of Vascular and Interventional Radiology 2007 · DOI

TL;DR

This paper evaluates the accuracy of an electromagnetic tracking system for guiding minimally invasive procedures like biopsies and thermal ablations. The system demonstrated sufficient spatial accuracy to display preprocedural imaging information in real-time, potentially improving needle placement and procedure outcomes, especially for challenging targets.

What Problem Does This Paper Address?

Minimally invasive image-guided procedures, such as radiofrequency ablation (RFA) and biopsies, are increasingly common. However, accurately placing needles during these procedures can be challenging due to complex anatomy, limited real-time visualization, and the potential for obscuring factors like gas release during ablation. Current imaging modalities like ultrasound (US) and computed tomography (CT) have limitations: US visualization can be poor and is angle-dependent, while CT requires significant radiation exposure and time. This leads to potential inaccuracies in needle placement, repeat procedures, and longer procedure times. The paper addresses the need for improved real-time guidance that can overcome these limitations by evaluating the spatial accuracy of an electromagnetic tracking system for needles.

Methods

The study employed a translational clinical trial approach involving 20 patients undergoing biopsies or RFA. The core technology was a commercial electromagnetic tracking device combined with prototype internally tracked needles and custom software.

Key methodological aspects include:

  • Needle Tracking: Internally tracked needles with sensors embedded in the tip were used. This allows for direct tracking of the needle's motion, correcting for bending, organ motion, and respiration, differentiating it from systems relying on external needle hub position.
  • Spatial Accuracy Evaluation: Tracking error was defined as the difference between the tracked needle position and its actual position determined by verification CT scans.
  • Registration Techniques:
    • Standard Registration: This involved manually identifying skin fiducials on CT scans and correlating them with tracking coordinates.
  • Advanced Registration: This method incorporated CT-confirmed previous needle positions as additional fiducials, allowing for affine (non-rigid) transformations. This was compared to rigid registration.
  • Semiautomatic Registration: This utilized actively tracked fiducials for faster registration.
  • US-CT Fusion: The system also tracked the US transducer, enabling real-time fusion of live US images with spatially corresponding CT multiplanar reconstructions (MPRs). This was achieved through a chain of transformations between US, sensor, tracking system, and CT coordinate systems.

Datasets

The study used data from a clinical trial involving 20 patients undergoing interventional procedures. No new benchmark datasets were introduced; the evaluation was based on clinical data collected during these procedures.

Key Insights and Ideas

  • Internalized Sensors are Key: The use of sensors inside the needle tip is crucial for accurate tracking, allowing the system to compensate for patient and organ motion, which is a significant improvement over external tracking methods.
  • Advanced Registration Significantly Improves Accuracy: Using previous needle positions as additional fiducials for affine registration dramatically reduced tracking error compared to using only skin fiducials. This highlights the importance of incorporating more recent positional information.
  • US-CT Fusion Potential: The successful fusion of real-time US with CT MPRs demonstrates the potential for a more comprehensive real-time visualization tool, combining the strengths of both modalities.
  • Field Distortions are a Factor: The study suggests that distortions from the CT scanner's electromagnetic field contribute to tracking error, a finding consistent with other laboratory evaluations.

Key Results

  • Basic Tracking Error: The mean basic tracking error was 6.4 mm ± 3.7 mm. After excluding one patient with an exceptionally large error, the mean improved to 5.8 mm ± 2.6 mm.
  • Improved Accuracy with Advanced Registration: Using affine registration with at least one prior needle position reduced the tracking error to 3.5 mm ± 1.9 mm, a significant improvement over rigid registration (4.8 mm) and standard registration (7.8 mm).
  • Registration Time: Semiautomatic registration with actively tracked fiducials was significantly faster (2 seconds) compared to manual needle-based registration (30-40 seconds).
  • US-CT Fusion Achieved: The study successfully demonstrated the fusion of tracked real-time US and CT images.
  • Sources of Error Identified: Patient motion, distortion of the tracking system by the CT table and gantry, and issues with fiducial placement were identified as sources of error.

Limitations and Future Work

The authors acknowledge several limitations:

  • Tracking Information Not Used for Guidance: The study evaluated spatial accuracy but did not directly assess the impact of the tracking system on the overall accuracy of needle placement by the interventionalist. This would require a prospective study where the tracking guidance is actively used.
  • Small Sample Size: The study involved a relatively small number of patients (20).
  • Single Operator: Only one physician operator performed the procedures, limiting the generalizability of findings related to operator skill and learning curves.
  • Variety of Procedures: The study included a range of procedures and sites, which could introduce variability.

Future work suggested includes:

  • Prospective Clinical Studies: Conducting studies with multiple operators to demonstrate specific clinical benefits and evaluate the impact of real-time tracking guidance on placement accuracy.
  • Further Optimization: Investigating improvements to fiducial design and placement, as well as exploring more advanced registration algorithms.
  • Real-time Guidance Evaluation: Directly assessing how using the tracking system for real-time guidance affects procedure outcomes, time, and radiation dose.

Summary updated

BibTeX
@article{krucker2007electromagnetic,
  author = {Jochen Krücker and Sheng Xu and Neil Glossop and Anand Viswanathan and Jörn Borgert and Heinrich Schulz and Bradford J. Wood},
  title = {{Electromagnetic Tracking for Thermal Ablation and Biopsy Guidance: Clinical Evaluation of Spatial Accuracy}},
  journal = {Journal of Vascular and Interventional Radiology},
  year = {2007},
  doi = {10.1016/j.jvir.2007.06.014},
}
Figure