Predictive Variables for Successful Treatment of Trigeminal Neuralgia with Radiosurgery

Gamma Knife Radiosurgery (GKRS) has been reported as a definitive management approach for medically refractory trigeminal neuralgia (TN). Controversy persists regarding factors that predict for treatment success and toxicity. Between September 1999 and December 2008, 777 GKRS procedures for patients with TN were performed at our institution. Follow-up was obtained via patient electronic medical records and telephone interview. Patients reported pain control following GKRS as excellent (100% pain relief no medication), good (100% pain relief with medication), fair (50-99% pain relief) or poor (<50% pain relief). Leksell GammaPlan software was used to measure cisternal nerve length (mm) and maximum dose (Gy) delivered to dorsal root entry zone (DREZ), pons, and the point of trigeminal nerve contact with the petrous dura. Targeting philosophy changed during the course of the series; DREZ was included in the target early in the series while more distal targets were favored later. Median dose for the series was 90 Gy with a single 4mm collimator and collimator output factor of 0.87. Follow-up data were obtainable for 504 patients with initial GKRS procedures for TN. Median follow-up time was 11.8 months. Overall, 83.3% of patients experienced an improvement in quality of life. 26% of patients developed some degree of facial numbness. 26% of patients reported a post-GKRS procedure with median time to next procedure (repeat GKRS, microvascular decompression or other procedure) of 4.4 years among those patients. Post-surgical numbness was a highly predictive variable for >50% pain relief (OR 3.2, p = 0.001). Multivariate analysis revealed that longer cisternal nerve length (OR 0.84, p = 0.002), diabetic status (OR 0.35, p = 0.006), and prior radiofrequency ablation (OR 0.34, p = 0.021) predicted for decreased likelihood of >50% pain relief. Higher DREZ dose predicted for development of facial numbness (mean 55.97 Gy vs. 47.12 Gy, p = 0.001); however, facial numbness was found to be the more significant predictor of treatment success on logistic regression (OR 3.2, p = 0.001). None of DREZ dose, distal nerve dose, or pontine max dose predicted for >50% pain relief. This review is the largest reported GKRS case series for the treatment of TN. The data suggest that the TN targeting location selected for the prescribed dose does not predict for treatment outcome.