Quantitative evaluation of respiratory motion‐induced blurring in expiratory‐phase CBCT for pancreatic SBRT

BACKGROUND: Cone-beam computed tomography (CBCT) is widely used in image-guided radiotherapy because it reflects the patient's anatomy immediately before treatment. However, CBCT images of the abdominal organs, including the pancreas, are often degraded by respiratory motion. PURPOSE: To address this limitation, this study quantitatively evaluated expiratory-phase CBCT (EP-CBCT), an acquisition method designed to reduce respiratory motion-induced blurring during stereotactic body radiotherapy (SBRT) for pancreatic cancer. METHODS: This retrospective analysis included 10 patients with pancreatic cancer undergoing SBRT at our institution. Daily CT and EP-CBCT scans were acquired on each treatment day. EP-CBCT images were obtained with repeated end-exhalation breath-holds to collect projection data during the expiratory phase. Because of feasibility limitations in routine clinical practice, a direct intra-patient comparison with conventional free-breathing CBCT was not performed. Image quality was evaluated using edge response width (ERW), structural similarity index (SSIM), and a blinded subjective assessment based on mean opinion scores (MOS). A supplementary retrospective comparison with a historical four-dimensional CBCT (4D-CBCT) dataset was also performed. RESULTS: Differences in ERW between daily CT and EP-CBCT (ΔERW) were evaluated using equivalence testing with a predefined equivalence margin of ± 1 mm. The mean ERW (98% confidence interval) was 2.04 mm (1.95-2.13) for EP-CBCT and 1.73 mm (1.66-1.79) for daily CT. The mean ΔERW was -0.31 mm (-0.40 to -0.23 mm). Equivalence testing demonstrated that EP-CBCT was statistically equivalent to daily CT (p < 0.01). The mean SSIM value was 0.83, indicating high structural similarity between EP-CBCT and daily CT. Subjective image quality assessment showed significantly higher scores for daily CT than for EP-CBCT, whereas EP-CBCT consistently outperformed historical 4D-CBCT, with good interobserver agreement. CONCLUSION: These findings suggest that EP-CBCT can reduce respiratory motion-induced blurring to a degree comparable to that of daily CT, supporting its potential utility for anatomical assessment immediately before irradiation in patients undergoing pancreatic SBRT.