Direct validation of atlas-based red nucleus identification for functional radiosurgery

Area of Science:

• Neurosurgery
• Medical Imaging
• Neuroanatomy

Background:

• Functional neurosurgery targets deep brain structures like the thalamus and basal ganglia for movement disorders.
• Radiosurgery offers a noninvasive lesioning alternative but requires precise, noninvasive target localization.
• Accurate noninvasive target identification is crucial for effective functional neurosurgery, motivating atlas-based approaches.

Purpose of the Study:

• To validate an atlas-based method for noninvasively identifying the red nuclei (RN) near the subthalamic nucleus (STN).
• To assess the accuracy of automated RN localization by comparing it with manual delineation on patient MRI scans.

Main Methods:

• RN coordinates from the Talairach and Tournoux atlas were transformed to the Montreal Neurological Institute (MNI) atlas space, creating a mask.
• The MNI atlas volume was nonrigidly registered to patient MRI, and the deformation field applied to the RN mask.
• Manual delineation of RN on 1.5 T T2-MRI was performed, and coordinates of the center of mass were compared with automated identification.

Main Results:

• High agreement was found between manually and automatically identified red nuclei structures across ten patients.
• Volumetric overlapping indices confirmed the accuracy of the atlas-based automated identification method.
• The study confirmed the reliability of the atlas-based approach for functional neurosurgery target localization.

Conclusions:

• The validated atlas-based method provides accurate, noninvasive localization of deep brain targets like the red nuclei.
• This technique enhances the precision of radiosurgery and other noninvasive functional neurosurgery procedures.
• Further fine-tuning of the automatic identification method is possible based on these findings.