ESTRO 2022 - Abstract Book

S281

Abstract book

ESTRO 2022

Results The proposed methods achieved a performance similar or better than the one achieved by the multi-organ model (Table 1). In particular, the proposed methods achieved a better Dice Score on three OARs out of six and a better HD95 in five OARs out of six when compared to the multi-organ model. In particular, the results of HD95 might suggest that combining single-organ models allows to compute a more reliable segmentation.

Conclusion The proposed methods achieved a performance similar or better than the one achieved by the multi-organ model (Table 1). In particular, the proposed methods achieved a better Dice Score on three OARs out of six and a better HD95 in five OARs out of six when compared to the multi-organ model. In particular, the results of HD95 might suggest that combining single-organ models allows to compute a more reliable segmentation.

PD-0319 Evaluation of a synthetic-CT generation method for MRI-only head and neck cancer radiotherapy

S. Riga 1 , C. Carsana 2 , M. Felisi 1 , D. Sibio 2 , A.F. Monti 1 , D. Lizio 1 , R.G. Pellegrini 3 , D. Curto 1 , G. Muti 1 , O.E. Panchi Maigualca 1 , B. Bortolato 2 , F. Bracco 1 , A. Vanzulli 4 , M. Palazzi 2 , A. Torresin 1 1 ASST Grande Ospedale Metropolitano Niguarda, Medical Physics Department, Milan, Italy; 2 ASST Grande Ospedale Metropolitano Niguarda, Radiotherapy Department, Milan, Italy; 3 Elekta AB, Global Clinical Science, Stockholm, Sweden; 4 ASST Grande Ospedale Metropolitano Niguarda, Radiology Department, Milan, Italy Purpose or Objective This work aims to evaluate the feasibility of a radiation therapy (RT) workflow based only on magnetic resonance (MR) images for head and neck (H&N) cancer, using the resources available in a general hospital, i.e., a diagnostic MR system and a conventional linac to realize full MR guided Radiotherapy (MRgRT). The RT planning performed on MR images (MRI- only radiotherapy) allows overcoming some limitations of the combined use of CT and MR images, like defects in co- registration. Unlike CT, MRI signals do not depend on electron densities (EDs), required by the treatment planning systems (TPS) to calculate the dose distribution. Therefore, it is necessary to associate MR data with ED maps, thus obtaining a synthetic-CT (sCT). Materials and Methods Thirteen patients with H&N cancer were scanned using a T1w 3D VIBE Dixon gradient echo sequence on a 1.5 T Magnetom Aera scanner (Siemens Healthcare), with a large field of view to cover the entire shoulder region. The patients were acquired in the same setup of CT simulation, using a customized thermoplastic mask fixed on a rigid table and MR compatible markers. An 18-channel body coil placed above the patient's head and shoulders was used. At the same time, the CT acquisitions were performed with a Brilliance CT scanner (Philips Medical System), in order to validate the sCT generation method, comparing the dose distributions. The sCT images are obtained from MR images, using a combination of a multi atlas-based approach and the bulk ED assignment to the contoured organs-at-risk (OARs). Furthermore, air cavities and bone structures should be properly contoured and assigned average EDs. First, MR images were auto-contoured using ADMIRE® software (research version 3.28, Elekta AB) and verified by a radiation oncologist. VMAT plans, optimized with a Monaco TPS (Elekta AB), on the sCT were recalculated on the CT deformably registered on the MR. Deviations between dose calculations on CT and sCT were evaluated using dose-volume histograms (DVH) and gamma analysis.