Abstract
OBJECTIVE: This study aims to evaluate the impact of personalized 3D-printed headrest devices on setup errors and dosimetry in nasopharyngeal carcinoma radiotherapy. METHODS: A total of 50 nasopharyngeal carcinoma patients (28 male, 22 female, aged 24-76 years, mean age 45 years) who received radiotherapy at our center between July 1, 2023, and August 10, 2024, were randomly divided into two groups. Group A: 3D-printed headrest and thermoplastic mask (Klarity Medical); Group B: Standard headrest and thermoplastic mask. All patients underwent a simulation CT (Philips IntelliSpace; 120 kV/200 mAs, 3 mm slices) in a supine position. The VMAT plan (70 Gy/33 fx) was designed in Eclipse 15.6: Plan-P: Actual PLA CT values included; Plan-0: PLA CT values set to -1000 (control). Daily CBCT verification (Varian Halcyon) was performed to measure setup errors at the Clivus, C4, and C7 vertebral levels. The Van Herk formula was used to determine the appropriate planning target volume (PTV) margin. Dose measurements used PCI, HI, CR, Dmean, and Dmax to assess PTVs (GTVnx+nd, PTV1, PTV2) and OARs (brainstem, spinal cord, lens, optic nerve). RESULTS: A total of 340 CBCT images were collected from 50 patients, with 164 images from Group A and 176 from Group B. Setup errors in Group B were generally larger than those in Group A. Statistically significant differences were observed in the AP direction at the Clivus, C4, and C7 vertebral ROI registrations. Roll rotational errors showed statistically significant differences in ROI registration at C4 and C7 vertebral levels. The external radiation margins in all directions for Group A were smaller than those for Group B, with the largest difference observed at the C7 vertebral level. The external margins for the C7 ROI registration in the LR, SI, and AP directions were 2.91, 2.97, and 3.01 mm, respectively. Compared with Plan-0, the Dmean and CR of the target volumes PTVnd+nx and PTV1 showed statistically significant differences (p < 0.05). Differences in Dmean, CR, and PCI for PTV2 were also statistically significant (p < 0.05). For the dosimetric evaluation of critical organs adjacent to the target volume, statistically significant differences were observed in the maximum doses to the brainstem, spinal cord, left lens, left optic nerve, and right optic nerve between the two planning approaches (p < 0.05). However, no statistically significant differences were found in the mean doses to these organs at risk (OARs) (p > 0.05). CONCLUSION: The application of 3D-printed immobilization technology significantly improves setup accuracy in nasopharyngeal carcinoma radiotherapy and reduces cervical spine displacement. The incorporation of 3D-printed materials exerts a measurable influence on target volume dose distribution and may notably increase the maximum dose delivered to adjacent critical organs.