Background: Stereotactic radiosurgery (SRS) is an effective treatment for trigeminal neuralgia (TN). Nevertheless, a proportion of patients will experience recurrence and treatment-related sensory disturbances. In order to evaluate the predictors of efficacy and safety of image-guided non-isocentric radiosurgery, we analyzed the impact of trigeminal nerve volume and the nerve dose/volume relationship, together with relevant clinical characteristics. Methods: Two-hundred and ninety-six procedures were performed on 262 patients at three centers. In 17 patients the TN was secondary to multiple sclerosis (MS). Trigeminal pain and sensory disturbances were classified according to the Barrow Neurological Institute (BNI) scale. Pain-free-intervals were investigated using Kaplan Meier analyses. Univariate and multivariate Cox regression analyses were performed to identify predictors. Results: The median follow-up period was 38 months, median maximal dose 72.4 Gy, median target nerve volume 25 mm 3 , and median prescription dose 60 Gy. Pain control rate (BNI I-III) at 6, 12, 24, 36, 48, and 60 months were 96.8, 90.9, 84.2, 81.4, 74.2, and 71.2%, respectively. Overall, 18% of patients developed sensory disturbances. Patients with volume ≥ 30 mm 3 were more likely to maintain pain relief (p = 0.031), and low integral dose (< 1.4 mJ) tended to be associated with more pain recurrence than intermediate (1.4-2.7 mJ) or high integral dose (> 2.7 mJ; low vs. intermediate: log-rank test, χ 2 = 5.02, p = 0.019; low vs. high: log-rank test, χ 2 = 6.026, p = 0.014). MS, integral dose, and mean dose were the factors associated with pain recurrence, while re-irradiation and MS were predictors for sensory disturbance in the multivariate analysis.