ESTRO 2020 Abstract Book

S480 ESTRO 2020

Material and Methods Six patients with GBM underwent <72-hour and 1-week delayed post-surgical MRI, with gross tumour volumes (GTV) delineated independently by a neuroradiologist (O nr ), and two oncologists (O 01 , O 02 ). Clinical target volumes (CTV) were generated using 25 mm isotropic margins, adjusted for anatomical boundaries, and planning target volumes (PTV) were grown using 5 mm isotropic margins. Sørensen–Dice coefficient (DSC) and volumetric analysis was performed with in-house software, in which GTVs were discretised into individual voxels and O nr contours were utilised as reference contours. Following local planning criteria 60 Gy in 30# treatment plans were generate for all individual PTVs, with dosimetric differences assessed through dose-volume histogram statistics of the delayed target volumes and organs at risk (OAR). Results One patient underwent rapid progression between MR acquisitions and was excluded. At a cohort level, GTV and PTV volumes decreased by 20 cm -3 and 71 cm -3 respectively for 1-week delayed MR delineation, whilst DSC values were approximately invariant (table 1) at 0.82.

MRI. Additionally, delineation on post-resection MRI was disfavoured due to potential for under-treatment; due to progression and/or non-rigid deformation. PO-0898 Primary tumor-status relevance in NSCLC with brain metastases undergoing radiosurgery A. Gonzalez Lopez 1 , F. Suarez 1 , D. Büchser 1 , A. Frias 1 , A. Lasso 1 , E. Mayrata 1 , P. Gonzalez-Fernandez 2 , E. Boveda 1 , P. Bilbao 1 1 Hospital de Cruces, Radiation Oncology, Barakaldo, Spain ; 2 Hospital de Cruces, Endocrinology, Barakaldo, Spain Purpose or Objective Brain metastases (BMs) have poor prognosis, therefore identifying subgroups of patients’ candidates to aggressive therapy might improve survival. Prognostic indices such as disease-specific graded prognostic assessment (DS-GPA) are used to select patients´ candidates for radiosurgery (SRS). Non-Small cell lung cancer (NSCLC) DS-GPA doesn´t consider primary tumor (PT) status. The aim of this study was to analyze if there are differences in overall survival (OS) after SRS between patients with controlled PT that develop BMs (Group 1) and newly diagnosed NSCLC with synchronous BMs (Group 2). The secondary objective was validating DS-GPA in this population. Material and Methods Retrospective analysis of 82 NSCLC patients with BMs´ undergoing SRS at our center between 2011 and 2018 (47 in group 1 and 35 in group 2). Patients were stratified according to NSCLC DS-GPA to evaluate the accuracy of survival estimates. Results Median OS was 17 months for group 1 (95% CI 5.7-28.3) and 12 for group 2 (95% CI 1-22.9), with no statistically significant differences between groups (p=0.654).In multivariate analysis, Karnofsky performance-status (p=0.005) and BMs´ surgery (p=0.005) were significant predictors of OS. Median OS stratified according to DS-GPA for scores 1.5, 2.5-3, and 3.5-4 were 12, 21, 51 months, respectively (p=0.028). No differences were observed between both groups´ regarding OS. According to our results median OS stratified according to DS-GPA was consistent with OS reported in previous studies.

A broad range of PTV and CTV inter-operator dose differences (fig. 1) were determined, with the largest variations for primary tumours proximal to multiple OARs. Despite OAR dose dependency on location of target volumes, systematic reductions in all OAR doses were observed for delayed delineation, including a crucial median 3.3 Gy reduction in mean brain dose.

Conclusion Tumour delineation at 1 week post-resection, once transient post-surgical effects dissipated produced smaller target volumes, alongside target dose increases and OAR dose reduction. Inter-operator variability remained similar whether employing post-surgical or delayed