Comparison of patient setup accuracy for optical surface-guided and X-ray-guided imaging with respect to the impact on intracranial stereotactic radiotherapy.

Purpose: The objective of this work is to estimate the patient positioning accuracy of a surface-guided radiation therapy (SGRT) system using an optical surface scanner compared to an X‑ray-based imaging system (IGRT) with respect to their impact on intracranial stereotactic radiotherapy (SRT) and intracranial stereotactic radiosurgery (SRS).

Methods: Patient positioning data, both acquired with SGRT and IGRT systems at the same linacs, serve as a basis for determination of positioning accuracy. A total of 35 patients with two different open face masks (578 datasets) were positioned using X‑ray stereoscopic imaging and the patient position inside the open face mask was recorded using SGRT. The measurement accuracy of the SGRT system (in a "standard" and an SRS mode with higher resolution) was evaluated using both IGRT and SGRT patient positioning datasets taking into account the measurement errors of the X‑ray system. Based on these clinically measured datasets, the positioning accuracy was estimated using Monte Carlo (MC) simulations. The relevant evaluation criterion, as standard of practice in cranial SRT, was the 95th percentile.

Results: The interfractional measurement displacement vector of the SGRT system, σ, in high resolution mode was estimated at 2.5 mm (68th percentile) and 5 mm (95th percentile). If the standard resolution was used, σ increased by about 20%. The standard deviation of the axis-related σ of the SGRT system ranged between 1.5 and 1.8 mm interfractionally and 0.5 and 1.0 mm intrafractionally. The magnitude of σ is mainly due to the principle of patient surface scanning and not due to technical limitations or vendor-specific issues in software or hardware. Based on the resulting σ, MC simulations served as a measure for the positioning accuracy for non-coplanar couch rotations. If an SGRT system is used as the only patient positioning device in non-coplanar fields, interfractional positioning errors of up to 6 mm and intrafractional errors of up to 5 mm cannot be ruled out. In contrast, MC simulations resulted in a positioning error of 1.6 mm (95th percentile) using the IGRT system. The cause of positioning errors in the SGRT system is mainly a change in the facial surface relative to a defined point in the brain.

Conclusion: In order to achieve the necessary geometric accuracy in cranial stereotactic radiotherapy, use of an X‑ray-based IGRT system, especially when treating with non-coplanar couch angles, is highly recommended.