Related Experiment Video
Updated: Apr 23, 2026
05:44
Using Saccadometry with Deep Brain Stimulation to Study Normal and Pathological Brain Function
Published on: July 14, 2016
7.0K
A Learning Based Fiducial-driven Registration Scheme for Evaluating Laser Ablation Changes in Neurological Disorders
Tao Wan1, B Nicolas Bloch2, Shabbar Danish3
1Department of Biomedical Engineering, Case Western Reserve University, OH 44106, USA ; School of Biological Science and Medical Engineering, BUAA, Beijing 100191, China.
Neurocomputing
|September 17, 2014
Summary
A new learning-based fiducial registration (LeFiR) method accurately models localized deformation from laser induced interstitial thermal therapy (LITT) for neurological disorders, outperforming other techniques in clinical validation.
Area of Science:
- Medical Imaging
- Computational Anatomy
- Neurosurgery
Background:
- Magnetic resonance (MR)-guided laser induced interstitial thermal therapy (LITT) is an investigational minimally invasive treatment for neurological disorders.
- Quantitative evaluation of treatment-induced changes using MR imaging markers is needed due to limited long-term outcome data.
- Accurate image registration is crucial for assessing treatment effects by comparing pre- and post-LITT scans.
Purpose of the Study:
- To introduce and validate a novel learning-based fiducial driven registration (LeFiR) scheme.
- To model localized deformations caused by LITT in neurological treatments.
- To quantitatively assess the performance of LeFiR against other registration methods.
Main Methods:
- Developed a LeFiR scheme using point matching for optimal landmark configuration to recover image deformation.
- Applied LeFiR to model localized deformations from MR-guided LITT for glioblastoma multiforme (GBM) and epilepsy.
- Validated LeFiR on a synthetic brain dataset (SBD) and two clinical datasets, comparing it with uniform grid (UniG), SURF, SIFT, and free-form deformation (FFD) methods.
Main Results:
- LeFiR achieved an average 90% improvement in recovering local deformation on the SBD, significantly outperforming UniG (82%), SURF (62%), and FFD (16%).
- On clinical GBM and epilepsy data, LeFiR demonstrated a 28% average improvement over the UniG method.
- LeFiR successfully modeled localized deformations characteristic of LITT procedures.
Conclusions:
- The LeFiR scheme provides a superior method for accurately registering images affected by localized deformations, such as those from LITT.
- LeFiR's performance suggests its potential utility in quantitatively evaluating treatment-related changes in MR imaging for LITT procedures.
- This technique can aid in the clinical validation and broader adoption of MR-guided LITT for neurological disorders.
Keywords:
Fiducial-driven image registrationbrain MRIlaser-induced interstitial thermal therapyminimally invasive therapy
