Abstract book - ESTRO meets Asia

S18 ESTRO meets Asia 2018

OC-047 Initial validation of dose calculation accuracy on Hounsfield Unit-corrected pelvic cone beam CT G. Bengua 1 , S. Lydiard 1 , J. Sandison 1 , S. Andersson 2 , R. Nilsson 2 1 Auckland City Hospital, Radiation Therapy, Auckland, New Zealand 2 Raysearch, RaySearch Laboratories AB, Stockholm, Sweden Purpose or Objective Accurate dose calculations on cone beam CT (CBCT) has a number of challenges; CBCT intensity-electron density relationship is not easily established and is highly dependent on imaging protocol and patient anatomy. An algorithm that converts CBCT intensities to the same scale as CT intensities and corrects for shading artefacts is being developed in the RayStation (RaySearch Laboratories) treatment planning system. This Hounsfield Unit (HU)- correction tool allows the use of CT HU to electron density relationship for CBCT dose calculations, irrespective of the CBCT acquisition system. We report the first validation results of this technique for prostate in an anthropomorphic phantom data and retrospectively Pelvic scans of an anthropomorphic phantom were performed using a CT and two CBCT systems using two routine clinical CBCT imaging protocols. Patient data consisted of previously scanned planning CT (pCT), CBCT and re-CT (rCT). Images were imported into RayStation for all subsequent analysis. Deformable registrations between the CTs and CBCTs were generated. Pelvic structures were contoured on each image set or propagated from the pCT. The HU-correction tool was used to create a corrected CBCT (c_CBCT). Voxel-based analysis of the HU distribution in the CTs, CBCTs, and c_CBCTs was carried out. Doses from a VMAT plan were calculated on the pCT, the bulk density overridden original CBCTs (bd_CBCTs), the c_CBCTs and the rCTs. Dose metrics for selected pelvis structures were compared. Results Using the rCT as the ground truth, HU histogram analysis of pelvic organs yielded absolute difference in the mean HU between the rCT and c_CBCTs that were significantly smaller than those between the rCT and CBCTs. For the dosimetric analysis on the phantom image sets, the relative difference between the mean rCT dose and c_CBCT dose (rCT-c_CBCT) was found to be generally smaller than that of the rCT-bd_CBCT. Both dose differences were consistently smaller for CBCTs acquired using our protocols for the Varian than the Elekta. Similar dosimetric results were obtained from the patient image sets where the doses calculated on the c_CBCT were found to be closer to the rCT dose. Conclusion The CBCT HU correction tool was able to closely estimate the correct HU on c_CBCT for various pelvic structures. This resulted in similar doses in the rCT and c_CBCT in both the phantom and patient image data sets. OC-048 Exclusive dose-escalated chemoradiotherapy in esophageal carcinoma: a randomized phase II study G. Crehange 1 , A. Bertaut 2 , E. Le Prisé 3 , P.L. Etienne 4 , E. Rio 5 , R. Pereira 6 , G. Noel 7 , K. Benezery 8 , D. Peiffert 9 1 Centre Georges-François Leclerc, Radiotherapy, Dijon, France 2 Centre Georges-François Leclerc, Biostatistics, Dijon, France 3 Centre Eugène Marquis, Radiotherapy, Rennes, France 4 Clinique Saint Brieuc, Gastroenterology, Saint Brieuc, France obtained patient data. Material and Methods Proffered papers: Upper GI and thorax

1. Zhao H, Wang B, Sarkar V, Rassiah-Szegedi P, Huang Y, Szegedi M et al. Comparison of surface matching and target matching for image-guided pelvic radiation therapy for both supine and prone patient positions. Journal of Applied Clinical Medical Physics. 2016;17(3):14-24. OC-046 Simple management system of kV imaging doses during image-guided radiotherapy J. Miyata 1,2 , M. Akimoto 3 , M. Kuroda 2 , M. Oita 2 , K. Kondo 1 , S. Itasaka 3 1 Kurashiki Central Hospital, Department of Radiological Technology, Kurashiki, Japan 2 Okayama University, Graduate School of Health Sciences, Okayama, Japan 3 Kurashiki Central Hospital, Department of Radiation Oncology, Kurashiki, Japan Purpose or Objective Image-guided radiotherapy (IGRT) enables instant correction for patient positioning using two-dimensional and three-dimensional kV x-ray images. Meanwhile, IGRT adds the extra imaging doses to the patient in daily treatment fractions. Therefore, the kV imaging doses should be appropriately managed on the consideration of risks and benefits to the patients. Monte Carlo simulations can be used for estimation of the imaging doses; however, it is complicated and current treatment planning systems do not provide the kV imaging dose calculation. In this study, we proposed the simple method for measurement and assessment of the kV imaging doses with half-value layers (HVLs) and source-surface distance (SSD). Material and Methods Entrance Surface Dose (ESD) was measured using a kV imaging dose management system (Accu-Gold; Radcal). Accu-Gold provides convenient and effective measurement of HVLs and ESD using a semiconductor detector (AGMS-D) and the ion chamber (10X-6-6). The detectors were set to the isocenter with the custom-made blocks as shown in Figure. Then, ESD was measured 10 times with On-Board Imager (OBI) equipped on the Clinac iX (Varian) and the ExacTrac (BrainLAB) for protocols of typical treatment sites such as head, thorax, abdomen, and pelvis. Before ESD measurement, the calibration factor (k) of the ion chamber was determined from the measured HVLs. The backscatter factor (BSF) was also determined from the equivalent square field using the published conversion table calculated with Monte Carlo simulation. ESD was calculated with the following equation; M raw × k × BSF × (SDD/SSD) 2 (M raw : measured value, SDD: source-detector distance). The SSD was calculated from the typical body thickness of (19 and 16 cm), (20 and 30 cm), (20 and 30 cm), and (20 and 30 cm) in the frontal and lateral directions for head, thorax, abdomen, and pelvis, respectively. Results Table shows the average ± SDs of ESD with OBI and ExacTrac (x-ray Tube no.1 and 2) for head, thorax, abdomen, and pelvis protocols. The measured imaging dose of OBI was comparable to the calculated skin dose with Monte Carlo simulation which was demonstrated in other reports.

Conclusion Our proposed method can assess the kV imaging doses, which was useful for quality management of kV imaging system because of its instant and convenient procedure.