To shift or not to shift: identifying and correcting patient motion after couch rotations in non‐coplanar intracranial radiosurgery with stereoscopic X‐ray imaging

Abstract Background Frameless linear accelerator (linac)‐based image‐guided stereotactic radiosurgery (SRS) or fractionated stereotactic radiotherapy (FSRT) are a widely used treatment option for intracranial lesions. Given the high radiation doses involved, it is crucial to maintain precise patient positioning throughout treatment. This requires that geometric inaccuracies arising from patient motion or setup errors are identified and corrected. With frameless immobilization, image‐guidance has a greater impact, especially in non‐coplanar settings that can lead to patient motion and discrepancies between couch and radiation isocenters. Purpose Both patient and phantom studies were conducted to assess and quantify the magnitude of geometric uncertainties after couch rotations, aiming at evaluating the clinical need for their correction to warrant a precise treatment delivery. Methods Intrafraction X‐ray data, performed by ExacTrac Dynamic (ETD) to monitor and correct patients’ position throughout treatment delivery, were collected from 50 patients treated for brain metastases in stereotactic non‐coplanar schemes and immobilized by stereotactic double‐layered thermoplastic mask systems: 26 patients treated in 40 single‐fraction SRS (168 stereoscopic X‐ray images); 24 treated with FSRT in 128 fractions (278 stereoscopic X‐ray images). Additionally, a head phantom was utilized and 350 measurements under two different couch loads were carried out to distinguish true patient motion from deviations caused by couch rotations or system‐related effects. For both studies, ETD stereoscopic X‐rays were acquired after each couch rotation and the first measured positioning deviation was calculated by comparing X‐ray images to the treatment plan's digitally reconstructed radiographs. Results Clinically relevant deviations were observed, exceeding clinical tolerance (≥ 0.5 mm/0.5°) mostly in the lateral and yaw directions and requiring repositioning in nearly half of the couch rotations. These deviations measuring up to 2 mm, revealed to be emerging mainly from patient motion rather than linac setup, as the phantom study showed maximum deviations of up to 0.6 mm and 0.4° when simulating a patient treatment and an interquartile range that did not exceed 0.2 mm and 0.2°. Conclusions These findings demonstrate the importance of a continuous intrafraction motion monitoring and repositioning in cranial stereotactic treatments, especially in non‐coplanar settings.