Improved Setup Accuracy for Proton Dibh Thoracic Treatments Using CT-on-Rails–Based 3D-3D Image Registration
Abstract
Purpose
Accurate patient setup in deep inspiration breath hold (DIBH) proton therapy is critical due to steep dose gradients, limited range robustness, and the frequent use of tight clinical target volume (CTV) margins. This study introduces a CT-on-Rails (CTOR)–based 3D–3D image registration workflow for proton DIBH treatment of thoracic tumors to address limitations of conventional 3D–2D image guidance in resolving complex anatomical changes, particularly in multi-site disease.
Methods
Proton therapy patients with thoracic tumors treated under DIBH conditions are being retrospectively analyzed and categorized by tumor location as midsternal or diaphragm adjacent. At the time of submission, four patients were included (aim of having at least ten patients by the time of conference presentation). For each fraction, setup was evaluated using 3D–2D orthogonal radiographs, followed by CTOR-based 3D–3D registration. CTOR-based setup corrections were quantitatively analyzed to assess CTV alignment improvements relative to clinical tolerance thresholds, and preliminary clinical guidance for applying CTOR-based image guidance in proton DIBH treatments for moving thoracic targets was developed.
Results
CTOR-based 3D–3D image registration improved target alignment compared with conventional 3D–2D radiographic guidance across all patients. The average CTOR couch shift following initial X-ray–based bony alignment was 0.51 ± 0.32 cm. The largest alignment improvements were observed in patients treated at multiple anatomical sites. Based on these findings, a setup discrepancy threshold of approximately 0.3 cm was identified as a practical trigger for CTOR-based intervention. For fractions exceeding this threshold, CTOR-guided correction resulted in visibly improved target alignment relative to surrounding anatomy.
Conclusion
CTOR-based 3D–3D image guidance enhances patient positioning and confidence in dose delivery for proton DIBH treatments. The proposed workflow and correction thresholds provide a practical, clinically actionable framework to reduce geometric uncertainty and improve treatment robustness, particularly for multi-site proton therapy.