Related Experiment Video
Updated: Jun 17, 2026
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Localization and registration accuracy in image guided neurosurgery: a clinical study
Reuben R Shamir1, Leo Joskowicz, Sergey Spektor
1School of Engineering and Computer Science, The Hebrew University of Jerusalem, Givat Ram Campus, 91904 Jerusalem, Israel. rubke@cs.huji.ac.il
International Journal of Computer Assisted Radiology and Surgery
|December 25, 2009
Summary
Fiducial localization error (FLE) in image-guided neurosurgery varies significantly. Fiducial registration error (FRE) and Fitzpatrick
Area of Science:
- Neurosurgery
- Medical Imaging
- Surgical Navigation
Background:
- Image-guided neurosurgery relies on accurate registration between preoperative images and intraoperative anatomy.
- Assumptions about the consistency of localization and registration errors may influence surgical planning and execution.
Purpose of the Study:
- To quantify and compare clinical localization and registration errors in image-guided neurosurgery.
- To challenge existing assumptions regarding the reliability of error metrics in neurosurgical navigation.
Main Methods:
- Twelve neurosurgery patients with navigation systems were analyzed.
- Landmarks were localized on preoperative MRI and intraoperative anatomy using a tracked pointer.
- Six hundred twelve scenarios were generated to compute fiducial localization error (FLE), target registration error (TRE), fiducial registration error (FRE), and Fitzpatrick's target registration error estimation (F-TRE).
Main Results:
- Image and physical FLE ranged from 0.5-2.0 mm and 1.6-3.0 mm, respectively.
- Measured TRE, FRE, and F-TRE were 4.1 ± 1.6 mm, 3.9 ± 1.2 mm, and 3.7 ± 2.2 mm.
- Low correlations were found between FRE and TRE (0.19) and F-TRE and TRE (0.37).
Conclusions:
- Fiducial localization error (FLE) exhibits significant variability, contrary to common assumptions.
- Fiducial registration error (FRE) and Fitzpatrick's target registration error estimation (F-TRE) are unreliable indicators of target registration error (TRE) in image-to-patient registration.
