Boosting the Database: Three-Dimensional Visualization of Neurovascular Compression Syndromes Improved by High-Resolution 3 Tesla MRI

Purpose: The aim of this study was an improved three-dimensional (3D) visualization of neurovascular relationships and nerve/vessel conflicts in the posterior fossa based on high-resolution 3D datasets obtained at 3T.

Methods: Between March 2004 and March 2005 12 patients (7 women and 5 men, age range 39–80 ys; average (±SD) 54±13 ys) suspected of having a neurovascular compression syndrome were investigated at 3.0T (GE Signa) equipped with an 8-element brain receiver coil using a high-resolution cycled 3D FIESTA sequence with 1-mm axial sections (interpolated 0.5mm; in-plane resolution 0.4–0.8mm). Out of 12 patients 8 had trigeminal neuralgia, 2 had hemifacial spasm, and 2 had glossopharyngeal neuralgia. The 3D FIESTA datasets obtained served as the source for volume rendering image processing which was performed on a PC (3 GHz processor, 1GByte RAM) using the 3D visualization software Amira 3.1 Developer^®. Image processing included manual segmentation and classification of tiny brainstem structures and an automated computer-based volume rendering process.

Results: The high-resolution 3D data sets obtained in conjunction with 3D visualization software provided for spatially realistic impressions of the brainstem and the posterior fossa. The possibility to turn, to extend, and to look from arbitrary positions at highly resolved structural detail added up to turn 3D visualization into a powerful diagnostic tool. A prerequisite for improved volume viewing was the basic image quality in terms of resolution, contrast, and signal as provided by the T2-over-T1-weighted images. The ability of 3.0T MRI for higher image quality, higher SNR, and higher resolution improved the 3D visualization of the brain nerves compared to 1.5T datasets, enhancing diagnostic confidence. While cycling of the 3D FIESTA had a beneficial effect throughout the study, a higher vulnerability for artifacts noted at 3.0 T rarely was problematic for excellent 3D reconstruction. Automated segmentation was inferior to manual segmentation, because it did not perform well on tiny structures and the separation of vessels and nerves.

Conclusion: The combination of T2-over-T1-contrast imaging at 3.0T, performed with optimized multi-element brain coils, and 3D volume rendering emerges as the superior choice to date in visualizing nerve/vessel conflicts in the posterior fossa.