Are Two Intrafraction CBCT-Based Corrections Adequate for Online Adaptive Prostate SBRT to Ensure Accurate Target Coverage?
Abstract
Purpose
To evaluate whether two intrafraction CBCT-based motion corrections are sufficient and necessary to ensure accurate target coverage in online adaptive stereotactic body radiotherapy (SBRT) of prostate cancer.
Methods
This retrospective study includes fourteen prostate patients who received online adaptive SBRT on the Ethos platform. After adaptation, patients underwent verification CBCT (CBCT0) with final shifts, and immediately treated with nine IMRT beams. Intrafraction CBCTs acquired after beam 3 (CBCT1) and beam 6 (CBCT2) served as intrafractional surveillance. Prostate displacement was determined by rigidly registering the implanted fiducials on the CBCTs to the adaptive planning CBCT, and corresponding shifts were applied. For a retrospective dose reconstruction, the prostate, bladder, and rectum were automatically segmented using an in-house AI algorithm, and contours were reviewed by expert clinicians. Bladder/rectum contour-guided deformable image registration (C-DIR) was performed between the CBCTs and the adaptive planning CBCT. Subsequently, the delivered dose was reconstructed, accounting for anatomical changes and registration uncertainties. Dose indices for the prostate, bladder, and rectum were extracted and evaluated with and without motion corrections.
Results
Mean 3D displacement of prostate was 1.7±1.4mm (range: 0.2-8.9 mm) between CBCT0 and CBCT1, and 1.1±0.7mm (range: 0-3.4 mm) between CBCT1 and CBCT2. Significant (≥2mm) prostate movement occurred in 19% of the surveillances, with the majority (85%) along the bladder-rectum axis. C-DIR had sub-millimeter accuracy for the prostate, measured by fiducial positions. Correcting twice during delivery, as in the actual treatment, brought the tumor coverage (CTV-D95) within 3% of the planned values for all evaluated cases. In contrast, either correcting once or not correcting at all would cause frequent violations exceeding 3%.
Conclusion
Two intrafraction corrections during online adaptive prostate SBRT treatments provide a balance between treatment accuracy and efficiency. Converting post-treatment to automated online dose reconstruction for real-time, dosimetrics-oriented guidance is underway to further improve treatment efficacy.