ESTRO 2023 - Abstract Book

S1966

Digital Posters

ESTRO 2023

1 Fondazione Policlinico Universitario A. Gemelli, IRCCS, UOSD Fisica Medica e Radioprotezione, Rome, Italy; 2 Fondazione Policlinico Universitario A. Gemelli, IRCCS, UOC Radioterapia Oncologica, Rome, Italy; 3 Università Cattolica del Sacro Cuore, Facoltà di Medicina e Chirurgia, Rome, Italy Purpose or Objective Radiotherapy treatment involves several steps requiring accuracy and precision, performed by different professional figures. Physicians are responsible for the contouring activity, where they identify the tumor as the CTV structure and all the organs at risk (OARs). Despite this practice is regularized by international guidelines shared within the medical community, the inter- and intra-operator variability represents a limit to achieving the optimal results in terms of treatment success. This approach exploits the availability of a certain number of CT series, contoured by an expert physician. For interventional radiation therapy (IRT, brachytherapy) the contouring step is relevant and even more challenging than external beam radiotherapy (EBRT), because of the presence of applicators. The use of atlas-based segmentation in IRT, already tested for EBRT, may improve the treatment delivery in terms of time and repeatability. Materials and Methods CT images acquired on a Discovery CT590 RT CT scanner (GE Healthcare) were collected for 30 patients with cervical cancer and IRT prescription. They were reconstructed using a matrix of 512x512 and thickness of 0.62 mm. Bladder, bowel bag and rectum structures were defined as OARs and manually generated by an expert radiation oncologist on a dedicated treatment planning system (TPS Oncentra Masterplan v 4.6, Elekta). All the enrolled patients received intracavitary high- dose-rate IRT with vaginal applicators of 3 cm diameter. MIM Maestro v.7.1.2 (MIM Software), installed on a workstation with Intel Core i7-7700 CPU and 16 GB RAM, was employed to build a CT atlas for female pelvis. A specific workflow for auto-segmentation of OARs for cervical cancer BT was created. Two algorithms for final contours creation, Majority Vote (MV) and STAPLE (Warfield et al. 2004) were compared, with two kinds of atlases: the first one masking the applicator in order to minimize its effect and the second one without masking the applicator. To evaluate the atlas performances, the Dice Coefficient (DC) and Mean Distance Agreement (MDA) were computed. Results The workflow gave the best results for DC and MDA in the case of the bladder (DC = 0.79 and MDA = 3.06 mm) corresponding to good reproducibility. Lower values were observed (DC = 0.61(0.54) and MDA = 3.54(6.77) mm) for rectum and bowel bag, respectively. The comparison between algorithms showed higher performances with STAPLE than MV (p < 0.05). No relevant differences were stated between atlases. Fig.1Bladder contour obtained with MIM.

Conclusion This study analyzed performances of an atlas-based segmentation method with MIM MAESTRO. The best proposal of bowel bag, rectum and bladder contours in HDR interventional radiation therapy was obtained using the STAPLE algorithm. No difference was found between the two atlases, thus implying that masking the applicator is not necessary. Future steps will involve the increase of the number of atlas’ subjects and the comparison with neural networks performances.

PO-2183 Evaluation of MOSFET response in phantom for in vivo dosimetry in high-rate brachy breast treatments

M. Gutierrez Ruiz 1 , J. Alonso 1 , G. Camacho 1 , A. Reguilon 1 , J. Albendea 2 , F. Pinto 1 , M. Garcia 1 , A. Rivero 1 , C. Uriel 1 , V. Cañon 1 , M. Ferri 1 , J. Anchuelo 1 , R. Astudillo 1 , D. Bruzos 1 , F. Gomez 1 , A. Vazquez 1 , I. Raba 1 , M.T. Pacheco 1 , P.J. Prada 1 , S. Arrebola 1 , R. Fabregat 1 1 Hospital Universitario Marques de Valdecilla, Radiation Oncology, Santander, Spain; 2 Hospital Universitario Marques de Valdecilla, Oncology Radiation, Santander, Spain Purpose or Objective The proposed recommendations by the ICRP and the IAEA and the growing tendency to hypofractioned dose schedules in high-dose rate brachytherapy (HDR-BT) treatments lead us to require in vivo dosimetry (IVD) as an independent and patient-specific quality assurance method to evaluate the delivered dose. A template was manufactured to perform IVD in HDR-BT breast cancer treatments with microMOSFET detectors, in a previous study [1]. Calibration and correction factors were proposed to minimize their angular, temperature, and source- detector distance dependence. The aim of this work is to evaluate the methodology proposed to perform IVD in this type of treatments in a phantom.

Materials and Methods