Dosimetric consequences of rotational setup errors with direct simulation in a treatment planning system for fractionated stereotactic radiotherapy

The purpose was to determine dose‐delivery errors resulting from systematic rotational setup errors for fractionated stereotactic radiotherapy using direct simulation in a treatment planning system. Ten patients with brain tumors who received intensity‐modulated radiotherapy had dose distributions re‐evaluated to assess the impact of systematic rotational setup errors. The dosimetric effect of rotational setup errors was simulated by rotating images and contours using a 3 by 3 rotational matrix. Combined rotational errors of [Formula: see text] and [Formula: see text] and residual translation errors of 1 mm along each axis were simulated. Dosimetric effects of the rotated images were evaluated by recomputing dose distributions and compared with the original plan. The mean volume of CTV that received the prescription dose decreased from [Formula: see text] (original) to [Formula: see text] , and [Formula: see text]. Minimal changes in the cold and hot spots were seen in the CTV. In general, the increase in the volumes of the organs at risk (OARs) receiving the tolerance doses was small and did not exceed the tolerance, except for cases where the OARs were in close proximity to the PTV. For intracranial tumors treated with IMRT with a CTV‐to‐PTV margin of 3 mm, rotational setup errors of 3° or less didn't decrease the CTV coverage to less than 95% in most cases. However, for large targets with irregular or elliptical shapes, the target coverage decreased significantly as rotational errors of 5° or more were present. Our results indicate that setup margins are warranted even in the absence of translational setup errors to account for rotational setup errors. Rotational setup errors should be evaluated carefully for clinical cases involving large tumor sizes and for targets with elliptical or irregular shape, as well as when isocenter is away from the center of the PTV or OARs are in close proximity to the target volumes. PACS number: 87.53.Bn