Evaluation of intrafractional motion for patients receiving whole brain radiotherapy in open-faced immobilisation masks

Abstract: Purpose/Background Patients receiving whole brain radiotherapy (WBRT) today are treated with an immobilisation mask which fully covers their face and therefore might be experienced as uncomfortable and claustrophobic. In recent years, optical surface scanning (OSS) systems have been implemented for patient positioning and real-time motion monitoring [1]. OSS systems cannot track the patients surface through the closed masks and are therefore only monitoring the motion of the mask. Using open-faced immobilisation masks would allow the OSS to monitor the patient directly, but may allow for more movement. The purpose of this thesis was to evaluate the intrafractional motion of WBRT patients receiving treatment in open-faced immobilisation masks. Material and Method The accuracy of the x-ray and surface imaging system (EXTD) was investigated using two different rigid head phantoms with internal bony anatomy for matching purposes. The surface imaging is dependant on the phantoms optical characteristics [1] and therefore two different skin colours were used. The phantoms were placed on the treatment couch in nine different rotational angles and the position was verified with CBCT (Cone Beam Computed Tomography). The surface imaging and x-ray systems were compared to the CBCT using a Wilcoxon signed-rank test (α=0.05). The patients included in this study all received WBRT in open-faced immobilisation masks and were monitored using the combined surface and x-ray imaging capabilities of the ExacTrac Dynamic system. In total, 25 treatment fractions from 5 patients were analysed. Results Overall, both x-ray and surface imaging agreed with the CBCT within 0.4 degrees for all investigated angles. The largest deviation was 0.4 degrees and 0.3 degrees in pitch and yaw rotations for the x-ray imaging and surface imaging, respectively. No statistically significant difference was found between the x-ray and CBCT in the roll direction (p>0.05, both phantoms). For the other rotational directions there was a statistically significant difference (p<0.05) between the x-ray and the CBCT for both phantoms and all directions showed a statistically significant difference (p<0.05) between the surface and the CBCT for both phantoms. For all five patients, the median (range) translational vector was 0.3 (0.0-1.3) mm, and 95% of the vector deviations were within 0.7 mm. For the individual translational directions, the absolute median (range) was found to be 0.1 (0.0-1.0) mm in lateral (lat), 0.2 (0.0-1.1) mm in longitudinal (long), and 0.1 (0.0-0.7) mm in vertical (vert), respectively, and for the rotational directions 0.1 (0.0-0.8) degrees in pitch, 0.1 (0.0-0.7) degrees in roll, and 0.1 (0.0-0.7) degrees in yaw, respectively. 95% of the deviations were within 0.4 mm in lat, 0.6 mm in long, 0.3 mm in vert, 0.5 degrees in pitch, 0.3 degrees in roll, and 0.5 degrees in yaw. Conclusion For the two head phantoms, excellent agreement was observed for both x-ray and surface imaging compared to CBCT (within 0.4 degrees), which fulfills the QA guidelines published by ESTRO-ACROP and AAPM TG-302 [1,2]. The EXTD system's accuracy showed no dependance on the colouring of the phantoms. This study shows that the surface guidance real time tracking of the patient's rigid face structures can detect submillimeter patient motion and in combination with its beam hold capabilities, deliver a high-accuracy treatment in the open face masks. The median intrafraction motion observed was 0.3 mm for all patients, and hence, open-faced masks can further be investigated for other patient groups.