S2344
Physics - Quality assurance and auditing
ESTRO 2026
Digital Poster 1354 Characterisation of self-determined breath holds in dual MR/CT simulation for lung MRgRT Julien Boudet, Aurélie Petitfils, Ilyas Achag, Laurent Delcoudert, Mathieu Gonod, Marie-Hélène Mercier, Léone Aubignac, Axel Martin, Igor Bessieres Department of Medical Physics, Centre Georges- François Leclerc, Dijon, France Purpose/Objective: MRI has historically played a limited role in lung cancer management. The advent of MR-guided radiotherapy (MRgRT) on MR-Linac systems now enables lung treatments1. In MRgRT workflows, electron density mapping requires two consecutive simulations: an MR reference image and a rapid-CT. While external systems can guide patients during Breath Hold (BH) for mobile tumours to improve reproducibility, such dual MR/CT guidance systems remain uncommon. Given the respiratory limitations of lung cancer patients, a self-determined Breath Hold (sBH) approach was implemented instead of the usual exhale BH, allowing a more comfortable BH.This study evaluated the effectiveness of our simulation practices for lung MRgRT. Differences in BH levels between MR and CT acquisitions were quantified, Deformable Image Registration (DIR) outputs were qualitatively assessed. Subsequently, to evaluate the dosimetric impact of non-ideal BH CT scans, alternative CT scans were taken from 4D-CT phases. Material/Methods: This study included 28 reference lung cases treated on an MRIdian system (Viewray, Inc.). Each case included MR and rapid-CT acquisitions under sBH, without external guidance. In addition, a free-breathing 4D-CT was acquired. BH levels were compared via diaphragm position using rigid bone registrations between image sets. DIR quality was visually classified by a reference physicist as ideal or non-ideal. Two 4D-CT phases (corresponding to the closest and furthest positions relative to reference) were used to evaluate dosimetry impact on the CTV, lungs, spinal cord using D98%, D0.035cc and Dmean metrics. Results: For rapid-CTs, 54% of sBH acquisitions corresponded to a free breathing phase, including 18% in exhale. The mean deviation of diaphragm position between MR and rapid-CT BH was 2.4 ± 12.3 mm. For similar BH levels in both MR and CT acquisitions, DIR outputs were rated as ideal in 86% of cases. For BH level difference exceeding 10 mm, it was rated as ideal in only 15% of cases. Dosimetry deviations were minimal for OARs (Dmean < 1% for all, and D0.035cc < 1% for spinal cord) but larger for the CTV, up to 6%, 12% and 4% for D98%, D0.035cc and Dmean, respectively. Conclusion:
The lowest TLCS/ALCS and largest deviations occurred for 2%/2 mm, confirming higher sensitivity for stricter criteria. For 3%/3 mm, only 1.1% of IMRT and 1.0% of VMAT plans showed gamma passing rate %GP < 95%, which was the applied criterion. For 3%/2 mm, these values were 9.3% and 1.0%, respectively (Figure 1). The obtained ALCS and TLCS levels for 3%/3 mm were within the range recommended by AAPM TG-218. Unfortunately, due to automatic overwriting of plans in the Ethos system, it was not possible to recover the original clinical plans with results below ALCS for additional verification.
Conclusion: The determined TLCS/ALCS limits allowed to define the tolerance and action levels for verification plans performed on the Ethos system using the ArcCHECK device in our clinic. The obtained results indicate that the currently applied acceptance criteria can be considered adequate. Since the number of VMAT plans was significantly smaller than the number of IMRT plans, the analysis will be repeated after collecting a larger amount of data in order to verify the stability of the presented results. References: [1] M. Miften et al., „Tolerance limits and methodologies for IMRT measurement - based verification QA : Recommendations of AAPM Task Group No. 218 ”, Med. Phys., t. 45, nr 4, April 2018, doi: 10.1002/mp.12810. Keywords: Adaptive radiotherapy, QA, gamma analysis
Made with FlippingBook - Share PDF online