In-vivo accuracy of frameless, MLC-based LINAC SRS thalamotomy for essential tremor.

Background: For many years, the standard of care for radiosurgical thalamotomy has been a frame-based treatment, either on Gamma Knife or a linear accelerator with cones. Our institution has recently completed enrollment of phase I/II prospective clinical evaluating the safety and efficacy of a linear-accelerator based treatment with thermoplastic mask utilizing the linear accelerator’s native high-definition collimator to generate a similar rapid-falloff spherical dose distribution. In this study, we assessed the in-vivo accuracy of our treatments. Methods: We treated patients on an IRB-approved clinical trial (ClinicalTrials.gov Identifier: NCT03305588) using a Varian Edge™ (Palo Alto, CA) linear accelerator equipped with high-definition multi-leaf collimator and 10MV flattening-filter free beam using our previously described virtual-cone approach. Optical surface guidance was used to monitor the patient position during treatment. For 15 patients having follow-up imaging, the planned dose distribution was exported in the pre-treatment MPRAGE image space. The pre-treatment MPRAGE was coregistered to the 6-month post-treatment post-contrast T1 SPACE image using a two-stage linear registration (rigid followed by affine). The lesion was segmented using a semi-automated, threshold-based method. The centerof-gravity (COG) of the lesion and of the planned 65 Gy isodose volume were calculated. Results: The mean distances between the center coordinates of the lesion and target were less than 0.5 mm in all axes (mean ± standard deviation between lesion and target: x = -0.1 ± 0.3; y = -0.1 ± 0.5; z = -0.3 ± 0.7 mm). The mean Euclidean distance between the target and lesion COG was 0.86 mm (95% confidence interval 0.58 to 1.13 mm). The distance was significantly correlated with imaging signal-to-noise ratio (SNR), suggesting patients with greater head motion and poorer quality imaging resulted in a greater discrepancy between target and lesion. Conclusion: The distance between the target position and lesion position is consistent with sub-millimeter accuracy for frameless SRS thalamotomy. Imaging quality is a limiting factor when assessing the targeting accuracy. [ABSTRACT FROM AUTHOR]