Setup uncertainties in linear accelerator based stereotactic radiosurgery and a derivation of the corresponding setup margin for treatment planning

Purpose: In the present study, clinical stereotactic radiosurgery (SRS) setup uncertainties from image-guidance data are analyzed, and the corresponding setup margin is estimated for treatment planning purposes. Methods: Patients undergoing single-fraction SRS at our institution were localized using invasive head ring or non-invasive thermoplastic masks. Setup discrepancies were obtained from an in-room x-ray patient position monitoring system. Post treatment re-planning using the measured setup errors was performed in order to estimate the individual target margins sufficient to compensate for the actual setup errors. The formula of setup margin for a general SRS patient population was derived by proposing a correlation between the three-dimensional setup error and the required minimal margin. Results: Setup errors of 104 brain lesions were analyzed, in which 81 lesions were treated using an invasive head ring, and 23 were treated using non-invasive masks. In the mask cases with image guidance, the translational setup uncertainties achieved the same level as those in the head ring cases. Re-planning results showed that the margins for individual patients could be smaller than the clinical three-dimensional setup errors. The derivation of setup margin adequate to address the patient setup errors was demonstrated by using the arbitrary planning goal of treating 95% of the lesions with sufficient doses. Conclusions: With image guidance, the patient setup accuracy of mask cases can be comparable to that of invasive head rings. The SRS setup margin can be derived for a patient population with the proposed margin formula to compensate for the institution-specific setup errors.