ESTRO 36 Abstract Book

S330 ESTRO 36 _______________________________________________________________________________________________

whereas patients with AC/CC genotypes had a median PFS of 13.7 months ( P =0.0088). No significant differences in PFS and OS for the other genotypes of the investigated MC4R polymorphisms were found. Conclusion The rs489693 AA genotype is significantly associated with a shorter PFS in GBM patients treated with radiotherapy and temozolomide schedule. The present, pilot study may represent the stimulus to prospectively investigate the role of MC4R polymorphisms as a predictive pharmacogenetic marker or as a target for new drug therapies for GBM patients. PO-0631 Molecular subtypes of glioblastoma: immunohistochemical approach and clinical correlations L. Triggiani 1 , M. Cominelli 2 , E. Tratta 2 , P. Ghirardelli 1 , S. Pedretti 1 , N. Pasinetti 1 , M. Buglione 1 , R. Galli 3 , S. Magrini 1 , P.L. Poliani 2 1 Spedali Civili di Brescia, Radiation Oncology, Brescia, Italy 2 Pathology Unit, Department of Molecular and Translational Medicine- University of Brescia-, Brescia, Italy 3 San Raffaele Scientific Institute, Neural Stem Cell Biology Unit- Division of Regenerative Medicine- Stem Cells & Gene Therapy, Milano, Italy Purpose or Objective It is currently accepted that Glioblastomas (GBMs) can be classified according to their transcriptional profile in three major molecular subtypes (classic (CL), proneural/neural (PN/N) and mesenchymal (MES)) associated with different gene signature, specific molecular alterations and different prognosis. However, molecular approach is not routinely feasible in daily clinical practice. Thus, assessment of simple tools for GBMs sub-classification is stringently required, particularly for the development of personalized therapeutic strategies. Goal of our study is to investigate the possibility of classifying GBMs by an immunohistochemical approach and correlate the data with clinical information in a large cohort of patients. Material and Methods We evaluated a wide panel of biomarkers, recognized as subgroup-specific gene classifiers, by IHC on FFPE tissue samples from a large cohort of GBMs (n=151). EGFR amplification and 1p/19q deletion were assayed by FISH. Correlation with histological and prognostic features (including MGMT promoter methylation) were also performed. EGFR status and expression of TP53, IDH1, ASCL1, OLIG2, PDGFRa, PTEN, MET, YKL40 and CD44 were semi-quantitative assayed based on percentage and intensity of expression in immunoreactive cells. Data were also validated at molecular level by mean of RNAseq technology. For statistical analysis, to identify distinct GBM subclasses, we applied consensus clustering to our immunohistochemical data. For clinico-pathological correlations, we retrospectively collected all clinical data, including adjuvant treatment (radiotherapy - chemotherapy). Results By cluster analysis we identified EGFR, MET, YKL40 and ASCL1 as the most sensitive and specific markers. Their combination allowed to recognize GBM subtypes in 78.8% of cases. Moreover, pleomorphism and epithelioid features were found to be strongly associated to MES subtype (46.2% and 25.7% of MES cases respectively) and small cell component to CL subtype (65.1%). Notably, survival analysis showed that MES GBMs were associated with worst prognosis ( p =0.079) and CL GBMs is related to a better radio-chemotherapy response.

Conclusion Although GBMs are morphologically defined as unique entity, they are extremely heterogeneous tumours associated with great variability in term of response to treatment (radiotherapy-chemotherapy) and clinical outcome Our results indicates that using a restricted panel of highly sensitive markers we can identify GBM subgroups with protein expression profiles strongly related to the transcriptional profile. PO-0632 Phase I/II Trial on Intraoperative Radiotherapy (IORT) in Glioblastoma Multiforme (INTRAGO I/II) F. Giordano 1 , S. Brehmer 2 , B. Mürle 3 , G. Welzel 1 , E. Sperk 1 , A. Keller 1 , Y. Abo-Madyan 1 , S. Clausen 1 , F. Schneider 1 , C. Herskind 1 , M. Seiz-Rosenhagen 2 , C. Groden 3 , P. Schmiedek 2 , M. Glas 4 , D. Hänggi 2 , K. Petrecca 5 1 Universitätsmedizin Mannheim- Medical Faculty Mannheim- Heidelberg University, Department of Radiation Oncology, Mannheim, Germany 2 Universitätsmedizin Mannheim- Medical Faculty Mannheim- Heidelberg University, Department of Neurosurgery, Mannheim, Germany 3 Universitätsmedizin Mannheim- Medical Faculty Mannheim- Heidelberg University, Department of Neuroradiology, Mannheim, Germany 4 University of Bonn Medical Center, Divison of Clinical Neurooncology and Clinical Cooperation Unit Neurooncology MediClin Robert Janker Klinik, Bonn, Germany 5 Montreal Neurological Institute and Hospital, Department of Neurology and Neurosurgery, Montreal, Canada Purpose or Objective The median progression-free survival time after standard- of care therapy for glioblastoma multiforme (GBM) is about 7 months. Most GBM recur locally, suggesting that augmenting local treatments may improve outcomes. We here present the results of a phase I/II trial testing local radiation dose escalation to the resection cavity. Material and Methods INTRAGO I/II was a monocentric phase 1/2 trial that included patients with resectable GBM. Patients received intraoperative radiotherapy (IORT) with low-energy photons (50 kV) at three dose levels, whereas dosing started at 20 Gy (prescribed to the cavity margin) and was escalated in 10 Gy increments up to 40 Gy. After surgery