Abstract
BACKGROUND
Fusion of preoperative and intraoperative magnetic resonance imaging (iMRI) studies during stereotactic navigation may be very useful for procedures such as tumor resections but can be subject to error because of image distortion.
OBJECTIVE
To assess the impact of 3-dimensional (3D) vs 2-dimensional (2D) image distortion correction on the accuracy of auto-merge image fusion for stereotactic neurosurgical images acquired with iMRI using a head phantom in different surgical positions.
METHODS
T1-weighted intraoperative images of the head phantom were obtained using 1.5T iMRI. Images were postprocessed with 2D and 3D image distortion correction. These studies were fused to T1-weighted preoperative MRI studies performed on a 1.5T diagnostic MRI. The reliability of the auto-merge fusion of these images for 2D and 3D correction techniques was assessed both manually using the stereotactic navigation system and via image analysis software.
RESULTS
Eight surgical positions of the head phantom were imaged with iMRI. Greater image distortion occurred with increased distance from isocenter in all 3 axes, reducing accuracy of image fusion to preoperative images. Visually reliable image fusions were accomplished in 2/8 surgical positions using 2D distortion correction and 5/8 using 3D correction. Three-dimensional correction yielded superior image registration quality as defined by higher maximum mutual information values, with improvements ranging between 2.3% and 14.3% over 2D correction.
CONCLUSION
Using 3D distortion correction enhanced the reliability of surgical navigation auto-merge fusion of phantom images acquired with iMRI across a wider range of head positions and may improve the accuracy of stereotactic navigation using iMRI images.
Evaluation of Novel Genetic Algorithm Generated Schemes for Positron Emission Tomography (PET)/Magnetic Resonance Imaging (MRI) Image Fusion