S2174
Physics - Inter-fraction motion management and daily adaptive radiotherapy
ESTRO 2026
similarly for both devices (Figure 2b).
device designed for compatibility with a commercial couch top, comparing its positional reproducibility and
user comfort with a conventional open-face thermoplastic mask in healthy volunteers. Material/Methods:
Five volunteers were fitted with two immobilization systems: (1) a conventional open-face thermoplastic mask (3-point BoS, CQ Medical) and MoldCare cushion molded according to clinical procedures, (2) a fully volunteer-specific 3D-printed double-frame immobilization device (Figure 1). The 3D-printed device was created from T1-weighted MR-scans (3T Philips Achieva dStream) using Fusion360 (Autodesk) and fabricated in PLA using a BambuLab X1E 3D- printer.Each volunteer was positioned 10 consecutive times in both setups. Positional reproducibility was evaluated using AlignRT surface imaging (VisionRT), capturing 6-degree-of-freedom deviations relative to a reference surface capture. Deviations were mean- centered per participant, correcting for absence of an absolute positional reference and combined across volunteers. Differences in variances across setups were assessed using Levene’s test. Separately, the volunteer position was captured for 10 minutes in each setup to assess intra-fraction motion.Participants completed a Likert-scale questionnaire evaluating overall comfort and fit.
Conclusion: The 3D-printed immobilization device offers a personalized, digitally designed alternative to conventional thermoplastics, suggesting improved comfort and positioning consistency. These results demonstrate the potential of 3D-printing to enhance radiotherapy workflows and patient experience. Further design optimization and patient evaluation are warranted to confirm clinical applicability. Keywords: 3D-printing, Immobilization, Patient experience Proffered Paper 4173 Anatomically plausible image registration via inverse-consistency constraints in CT/CBCT-guided adaptive radiotherapy Tom Draper, Cornel Zachiu, Bas Raaymakers Department of Radiotherapy, University Medical Center Utrecht, Utrecht, Netherlands Purpose/Objective: Deformable image registration is critical in adaptive radiotherapy (ART), but the accuracy of the resulting deformations for contour propagation and dose accumulation remains an open problem for clinical workflows[1]. Enforcing local constraints in line with physical tissue properties can improve registration accuracy, especially when images are affected by noise or artifacts. However, the definition of such constrained regions typically requires a set of clinically approved contours to already be available on every daily CBCT image. Material/Methods: Here we propose a spatially adaptive registration algorithm requiring delineations only on the offline planning CT. The forward motion field maps the planning CT (moving) to the daily anatomy (reference),
Results: The 3D-printed device demonstrated comparable or improved positional reproducibility relative to the thermoplastic mask. Levene’s test revealed significantly (p<0.05) reduced variance for lateral (SD: 0.46mmTP vs 0.35mm3D), yaw (SD: 0.76°TP vs 0.38°3D), and roll (SD: 0.89°TP vs 0.40°3D) deviations, indicating enhanced positioning consistency, whereas vertical (SD: 0.31mmTP vs 0.48mm3D) deviations showed higher variability (Figure 2a). Intrafraction motion was similar between setups, except for one volunteer where the thermoplastic mask performed better. Outlier analysis confirmed that most deviations originated from this subject, suggesting design-related limitations requiring further refinement.Questionnaire feedback suggested improved comfort with the 3D- printed immobilization device. Participants reported reduced facial pressure and less claustrophobia, while perceiving the headrest as more self-centering. Stability and immobilization confidence were rated
Made with FlippingBook - Share PDF online