MRI sequences for detection of neurovascular conflicts in patients with trigeminal neuralgia and predictive value for characterization of the conflict (particularly degree of vascular compression)

Background and purpose. - A long-term study of the results on trigeminal neuralgia (TN) after microsurgical vascular decompression (Kaplan-Meier Curves at 20 years) showed that cure was achieved in 88.1% of the patients with a neurovascular compression (NVC) producing a large groove on the nerve (Grade 111), 78.3% of the patients with a NVC with nerve distortion or displacement (Grade 11), and 58.3% of the patients with a NVC with simple contact on the nerve (Grade 1). Therefore, preoperative visualization of the NVC by MRI and determination of its grading are important for the therapeutic decision. In this Study, we investigated the predictive value of MRI for detecting and assessing the degree of vascular compression in trigeminal neuralgia. Methods. - The study included 91 consecutive patients with a preoperative MRI (1.5 Testa) using 3D T2-weighted and angio-MR-TOF. NVC prediction and the degree of compression made by an independent observer were correlated with surgical data. Results. - Eighty of the 91 patients had a NVC on MRI, but 83 (91.2%) patients showed a NVC at surgical exploration (eight patients had no NCV). Thus, the sensitivity of imaging in detecting a NVC on the symptomatic nerves was 96% (80/83) and the specificity 100% (8/8). In addition, imaging analysis predicted the responsible vessel in 88.7% (71/80) of the cases and characterized the degree of NVC in 85% (68/80). The Kappa-coefficient (KC) for prediction of the NVC degree was 0.795 for arterial and venous compressions together (p < 0.01; 95% confidence interval, 0.71-0.88). The CK was 0.758 (p < 0.01, good agreement) for grade 1, 0.787 (p < 0.01, good agreement) for grade 11 and 0.824 (p < 0.01, excellent agreement) for grade III. Conclusions. - High-resolution 3D T2-weighted imaging in combination with angio-MR-TOF is a reliable technique for detecting NVC and predicting the degree of the compression in NVC. (C) 2009 Elsevier Masson SAS. All rights reserved.