Purpose: Prostate SBRT has shown to be a viable treatment for high-risk prostate cancer patients. Online CBCT-guided stereotactic adaptive radiotherapy (CT-STAR) has the potential to account for inter-fraction organ motion and allow for dose escalation to the prostate, while simultaneously treating the nodal volume. We report our workflow for online CT-STAR for prostate cancer using Ethos (Varian Medical Systems).
Methods: A CT simulation was acquired with a full bladder. The prostate (CTV_4000cGy), prostate and proximal seminal vesicles (CTV_3625cGy), and nodal chain (CTV_2500cGy), were delineated by the physician. PTV margins for respective CTVs were built into the plan template to allow for manual online edits of CTVs using the following margins: PTV_4000 (0.2cm S/I,L/R,Ant; 0.0cm Post); PTV_3625 (0.5cm S/I,L/R; 0.3cm Ant/Post); PTV_2500 (0.5cm isotropically). The prescribed dose is to be delivered over 5 fractions given every other day. For treatment, bladder, rectum, and bowel were auto segmented using a convolution neural network as part of the AI driven auto segmentation. The sigmoid was deformably registrered. The CTVs were rigidly propagated and aligned to the respective targets. All OAR and target contours were adjusted at physician discretion within a 3cm isotropic contouring sphere. The plan was optimized so that strict OAR constraints always met in the adapted plan. After adaptation, a repeat CBCT was acquired to ensure patient alignment.This adaptive workflow was used for 7 patients (32 fractions) at our institution. Initial CBCT to post CBCT time was extracted for all patients, capturing the entire adaptive process and treatment time.
Results: All intended adaptive fractions were successfully delivered using the CT-STAR workflow. Median time from acquisition of the initial CBCT to completion of treatment was 42min [29min–69min].
Conclusion: This adaptive workflow will be used as the basis of a phase I/II trial evaluating CT-STAR for whole-pelvis prostate SBRT for high-risk prostate patients.