Cone beam CT image guidance for intracranial stereotactic treatments: comparison with a frame guided set-up

An analysis is performed of the setup errors measured by a kV cone beam computed tomography (CBCT) for intracranial stereotactic radiotherapy (SRT) patients immobilized by a thermoplastic mask and a bite-block and positioned using stereotactic coordinates. We evaluated the overall positioning precision and accuracy of the immobilizing and localizing systems. The potential of image-guided radiotherapy to replace stereotactic methods is discussed.

Fifty-seven patients received brain SRT. After a frame-guided setup, before each fraction (131 fractions), a CBCT was acquired and the detected displacements corrected online. Translational and rotational errors were analyzed calculating overall mean and standard deviation. A separate analysis was performed for bite-block (in conjunction with mask) and for simple thermoplastic mask. Interobserver variability for CBCT three-dimensional registration was assessed. The residual error after correction and intrafractional motion were calculated.

The mean module of the three-dimensional displacement vector was 3.0 +/- 1.4 mm. Setup errors for bite block and mask were smaller (2.9 +/- 1.3 mm) than those for thermoplastic mask alone (3.2 +/- 1.5 mm), but statistical significance was not reached (p = 0.15). Interobserver variability was negligible. The maximum margin calculated for residual errors and intra fraction motion was small but not negligible (1.57 mm).

Considering the detected setup errors, daily image guidance is essential for the efficacy of SRT treatments when mask immobilization is used, and even when a bite-block is used in conjunction. The frame setup is still used as a starting point for the opportunity of rotational corrections. Residual margins after on-line corrections must be evaluated.