Deformable Registration for Geometric Distortion Correction of Diffusion Tensor Imaging

摘要

Geometric distortion of diffusion tensor imaging (DTI) always results in inner brain tissues shift and brain contour deformation and it will certainly lead to the uncertainty of DTI and DTI fiber tracking in the planning of neurosurgeries. In this study, we investigated the accuracy of two deformable registration algorithms for distortion correction of DTI in the application of computer assisted neurosurgery system. The first algorithm utilized cubic B-spline modeled constrained deformation field (BSP) registration of the 3D distorted DTI image to 3D anatomical image, while the second algorithm used multi-resolution B-spline deformable registration. Based on the results, we found that multi-resolution B-spline registration is more reliable than BSP registration for distortion correction of multi-sequence DTI images, the contour deformation and inner brain tissue displacement could be well calibrated in 2D and 3D visualizations. The mesh resolution of B-spline transform plays a great role in distortion correction. This multi-resolution B-spline deformable registration can help to improve the geometric fidelity of DTI and allows correcting fiber tract distortions which is critical for the application of DTI in computer assisted neurosurgery system.