ESTRO 37 Abstract book

S886

ESTRO 37

Material and Methods Brain MR images of 20 children (12 M, 6 F, age ranging from 8 months to 8 years) were retrospectively evaluated. All studies were performed on a 3 Tesla magnet. Subjects underwent a standardized protocol which included T1 weighted sequences. Inclusion criteria were the absence of pathological findings and more specifically the presence of a normal myelination pattern related to subject age. As image quality was paramount for our purpose, sedation was required in a number of subjects. Hippocampal region contouring was performed on axial slices by a panel of radiation oncologists and radiologists according to the atlas provided by the Radiation Therapy Oncology Group (RTOG) trial 0933. Qualitative analysis of contours was performed by comparing the pediatric brains with RTOG images obtained from adult subjects in terms of morphologic appearance and anatomic landmarks. Results In our cohort adult brain appearance was reached at about 12-15 months of age on T1-weighted images analysis. These results are consistent with findings in literature. In children older than 12-15 months hippocampus appearance was similar to adults for contornation purposes. In children younger than 12-15 months the poor contrast between white and grey matter did not allow to differentiate hippocampal structures on T1 weighted images. Specifically, one of the main contouring rules proposed in RTOG atlas, i.e. to contour only the hypointense grey matter and not the white matter could not be used. The cerebrospinal fluid containing structures indicated as landmarks in RTOG atlas were instead visible in this subgroup. Specifically, the temporal horns of the lateral ventricles and the quadrigeminal cistern could be used to delineate the temporo-mesial structures from the lateral portions of the temporal lobe. Conclusion Hippocampal contouring in childhood is feasible as soon as the myelination process allows to discriminate the hippocampal structures on T1-weighted imaging (12-15 months). In this group RTOG atlas recommendations can be used. In younger subjects accurate hippocampal contouring is not feasible because the myelination process is still incomplete. One possible solution could be sparing all temporo-mesial structures. Further studies are required to develop and validate ad hoc atlases. EP-1644 Potential gain of MRI-guided IMRT planning versus current clinical CBCT-guided VMAT for Wilms’ tumor F. Guerreiro 1 , E. Seravalli 1 , G.O. Janssens 1 , J.J.W. Lagendijk 1 , B.W. Raaymakers 1 1 UMC Utrecht, Department of Radiotherapy and Imaging Division, Utrecht, The Netherlands Purpose or Objective At University Medical Center Utrecht, radiotherapy treatment (RT) for Wilms' tumor is currently done using a cone beam computed tomography-guided volumetric modulated arc therapy (VMAT CBCT ) workflow. By adding real-time imaging guidance to the RT with the introduction of the magnetic resonance imaging (MRI)- linac, the use of small safety margins due to the better visualization of the target and the organs at risk (OAR) can be achieved. The purpose of this study is to quantify the potential reduction of planning target volume (PTV) margins and its dosimetric impact when using MRI-guided

) instead of

intensity modulated radiation therapy (IMRT MRI

workflow.

the clinical VMAT CBCT

Material and Methods Imaging data, including 4D-CT, CBCT and MRI-scans acquired during planning and treatment of 15 patients (average 3, range 1-8 years), were used to estimate the PTV expansion for each workflow. Both intra- and inter- fraction tumor bed motion (using four clips as surrogates) and patient positioning systematic (∑) and random (σ) errors were determined individually for each direction and patient. The average PTV expansion (PTV m ) was calculated using the van Herk recipe (2.5∑+1.7σ). For the dosimetric comparison, RT plans were generated for both delivery techniques using 5 margin scenarios (PTV m ± 0, 1 and 2 mm). All plans, with prescription doses (PD) ranging from 10.8-25.2 Gy, were optimized to achieve the same PTV coverage (D98% > 95%PD). Furthermore, the IMRT MRI plans were optimized with a 1.5T transverse magnetic field turned on to realistically model a MRI-linac treatment. Plan quality was evaluated by assessing mean dose (D mean ) and dose-volume statistics of the PTV and OAR (p<0.05, Wilcoxon). Results The calculated PTV m was 7 mm in cranio-caudal (CC) and 6 mm in anterior-posterior (AP) and left-right (LR) directions for the VMAT CBCT and 4 mm in CC and 3 mm in AP and LR directions for the IMRT MRI (Table 1). For both delivery techniques in all scenarios, PTV coverage was fulfilled. A reduction of the D mean to the OAR up to 18% was achieved with the IMRT MRI when using the estimated PTV m : D mean to the contralateral kidney was reduced from 4.3 ± 2.5 to 3.2 ± 2.2 Gy (p=6e-5), to the liver from 6.0 ± 3.1 to 5.0 ± 2.9 Gy (p=6e-5), to the spleen from 7.1 ± 8.1 to 6.1 ± 7.4 Gy (p=1e-4) and to the pancreas from 12.2 ± 5.6 to 10.9 ± 5.8 Gy (p=1e-4). The maximum reduction of the D mean to the OAR between the largest and smallest margin scenarios was up to 19% for VMAT CBCT (range 0.1- 2.7 Gy) and 21% for IMRT MRI (range 0.2-3.0 Gy).