Towards a validation of atlas warping techniques

Area of Science:

• Neurosurgery
• Medical Imaging
• Computational Anatomy

Background:

• Functional neurosurgery relies on pre-operative MRI and CT scans for planning.
• Subcortical targets like the thalamus and basal ganglia are difficult to precisely localize due to imaging limitations.
• Digital atlases are crucial for accurate surgical targeting by warping to patient anatomy.

Purpose of the Study:

• To compare three distinct atlas-to-patient warping techniques for functional neurosurgery.
• To evaluate the accuracy of different methods in localizing subcortical nuclei.
• To identify the optimal technique for customizing anatomical atlases to patient-specific data.

Main Methods:

• Comparison of two template-based nonlinear warping methods using MRI as an intermediary.
• Evaluation of a novel method using a pseudo-MRI derived from a basal ganglia and thalamus atlas.
• Validation using manual segmentations of subcortical nuclei and intra-operative thalamic stimulation data.

Main Results:

• The template-based atlas-to-patient warping technique demonstrated superior performance.
• This method effectively customized the anatomical atlas to individual patient imaging data.
• The novel pseudo-MRI method did not outperform the established template-based approaches.

Conclusions:

• Template-based atlas-to-patient warping is the most effective method for precise subcortical target localization in neurosurgery.
• Accurate atlas warping is critical for improving targeting accuracy in procedures like deep brain stimulation.
• Further research may explore refinements to existing template-based methods or novel approaches for enhanced accuracy.