Voxel-based localization in frame-based and frameless stereotaxy and its accuracy

Area of Science:

• Neurosurgery
• Medical Imaging
• Computational Anatomy

Background:

• Accurate stereotactic localization and registration are critical for neurosurgical interventions using computed tomography (CT) data.
• Conventional stereotactic localizers and frameless registration methods rely on accurate coordinate transformations.
• Patient movement during scanning can degrade the accuracy of stereotactic localization.

Purpose of the Study:

• To address the problem of accurate stereotactic localization and registration in CT data.
• To evaluate the effect of using a single transformation matrix (Volume Transformation Matrix - VTM) on accuracy.
• To present an algebraic approach for calculating stereotactic target coordinates.

Main Methods:

• An algebraic approach was used to calculate stereotactic target coordinates from tomographic data.
• Simulations and analysis of CT datasets from patients with CRW or GTC stereotactic localizers were performed.
• Comparison of Stereotactic Transformation Matrix (STM) and VTM-derived coordinates was conducted.

Main Results:

• The Volume Transformation Matrix (VTM) can lead to accuracy degradation, especially with patient movement.
• Average registration errors were 0.1 mm for anesthetized patients and 0.6-1.4 mm for non-anesthetized patients.
• Using all available fiducials in the region of interest minimizes localization errors for VTM-based methods.

Conclusions:

• The VTM approach requires careful consideration of patient movement and fiducial selection for optimal accuracy.
• Accuracy maps can visualize registration errors, aiding in error assessment.
• Proposed methods aim to minimize registration errors in both frame-based and frameless stereotactic procedures.