ESTRO 2025 - Abstract Book

S2986

Physics - Image acquisition and processing

ESTRO 2025

1788

Digital Poster Commissioning of a commercial synthetic CT software for MR-only planning of stereotactic brain treatments Riccardo Dal Bello, Jérôme Krayenbühl, Bertrand Pouymayou, Thomas Rudolf, Nicolaus Andratschke, Matthias Guckenberger, Stephanie Tanadini-Lang Department of Radiation Oncology, University Hospital Zurich and University of Zurich, Zurich, Switzerland Purpose/Objective: MR-only planning can provide several benefits including a simplified imaging workflow, no imaging radiation exposure and a reduction of the registration uncertainties. A synthetic CT (sCT) is generated from MR sequences and may replace the planning CT (CT). This work aims to report the experience gathered during the clinical commissioning of a commercial sCT software (Spectronic Medical AB, Helsingborg, Sweden) for stereotactic brain treatments. Material/Methods: We conducted a retrospective study with 10 patients, generating sCT post-treatment [1], matching these to the CT and re-calculating the dose of the delivered treatment plans. An in-house developed MR-compatible holder for stereotactic masks (Brainlab, München, Germany) has been adopted during the MR simulations [2]. The first part of the analysis assessed the rotations in patient positioning between sCT and CT, which should not exceed 3° if the sCT has to be used for patient positioning based on CBCT. The second analysis aimed to model and reconstruct attenuating materials (mask holder, mask) not present in the sCT (Figure 1). Finally, the third part aimed to assess the dose differences when re-calculating plans on sCT. Results: The first analysis yielded satisfactory results (Figure 2-left). The pitch exceeded 3° only for one patient, which has to be attributed to a limited amount of pillows for patient support. Following MR simulations took this into account and the pitch never exceeded 3°. Jaw and roll were excessive for one case, which could be attributed to large mask spacers used during the MR simulation to increase patient comfort. The second analysis concluded that the mask holder has to be modeled and has an influence up to 0.5% on the dose calculation, but the largest influence is given by the bolus-like effect of the mask itself, which can lead to deviations of up to 3% (Figure 2-center), in agreement with previous reports [3]. Finally, the dose recalculation showed the presence of outliers outside 5% deviation (Figure 3-right).