The accuracy of stereotactic radiosurgery with cone beam computed tomography image guidance for trigeminal neuralgia.

Introduction: Trigeminal neuralgia (TN) causes intense pain along the branches of the fifth cranial nerve. Stereotactic radiosurgery (SRS) may be offered to patients who do not achieve adequate pain response to medical or surgical management. We provide data on the accuracy of SRS with cone beam computed tomography (CBCT) image guidance for TN. Materials/Methods: We identified 9 patients with refractory TN. A stereotactic head frame was placed by a neurosurgeon for patient immobilization. Planning computed tomography (CT) images were acquired (512 x 512 pixels, 120 kVp, 501 mAs, 1 mm slice thickness) and the image set was coregistered with high-resolution 1 mm thin magnetic resonance (MR) images obtained in the T1- and T2-weighted, and fast imaging employing steady-state acquisition (FIESTA) sequences. The radiosurgical target was defined as a short length of the cisternal segment of the involved trigeminal nerve as demonstrated on the FIESTA sequence. Contouring was performed by the treating radiation oncologist. All patients were positioned on a 6 degree of freedom couch. CBCT images were obtained for initial localization. Positional corrections were made for deviations \u3e 0.5 mm or 0.5o followed by further image verification prior to initiating radiosurgery. Midway through treatment, intrafractional motion was assessed with a mid-verification CBCT. Positional corrections were repeated as needed followed by a mid-shift verification CBCT prior to completing the radiosurgical procedure. Results: SRS was delivered in a single fraction using 14 partial arcs with a 4 mm conical cone and 6 MV flattening-filter-free photon energy performed on a linear accelerator-based treatment system. Median number of CBCT images obtained was 4 (range, 3-7). Median translational and rotational positional deviations were 1 mm (range, 0-0.7 mm) and 0o (range, 0-0.4o), respectively. Only 1 patient required a positional correction for a translational deviation of 0.7 mm. Median radiosurgical target volume was \u3c 0.004 cc (range, \u3c 0.004-0.200 cc). Median isocenter dose was 85 Gy (range, 80-90 Gy). Median brainstem dose to 0.035 cc was 10.94 Gy (range, 5.41-14.19 Gy). Conclusion: SRS with CBCT image guidance for TN can achieve submillimeter positional accuracy. Long-term clinical follow-up is needed to assess safety and efficacy of treatment