ESTRO 36 Abstract Book

S334 ESTRO 36 _______________________________________________________________________________________________

distant brain failure, 41 % have been re-irradiated with stereotactic radiotherapy for the additional lesions. Radionecrosis occurred in 35 patients (18.8 %), at a median time of 15 months (3,0 - 38,1) and was associated with the infratentorial location of the metastasis (HR = 2.97 [1.47 - 6.01], p = 0.0025). Conclusion This study demonstrated the safety and efficacy of a 3 x 7.7 Gy HFSRT regimen for irradiation of cerebral metastasis resection cavities. These results compare favorably to historical data undergoing WBRT, SRS or HFSRT. It is an alternative to adjuvant WBRT after surgery, and allow the implementation of salvage therapies. PO-0638 FET PET prior to primary RT of glioblastoma patients - a recurrence pattern analysis D.F. Fleischmann 1,3 , M. Unterrainer 2 , R. Schön 1 , S. Corradini 1 , C. Maihöfer 1 , P. Bartenstein 2 , C. Belka 1,3 , N.L. Albert 2 , M. Niyazi 1,3 1 LMU Munich, Radiation Oncology, Munich, Germany 2 LMU Munich, Nuclear Medicine, Munich, Germany 3 German Cancer Consortium (DKTK) German Cancer Research Center (DKFZ) Heidelberg, Germany, Purpose or Objective 18 F-fluoroethyltyrosine (FET) PET imaging for primary RT treatment planning is of increasing importance. It remains unclear if margin adjustment and reduction is possible through additional FET PET imaging. Margins for FET PET derived biological tumor volumes (BTVs) were compared to MRI-based gross tumor volumes (GTVs) on a recurrence pattern analysis. Material and Methods Treatment plans and clinical outcome of 36 glioblastoma patients receiving FET PET scans before primary radiotherapy were retrospectively analyzed. The minimal margin including the entire tumor recurrence was calculated for the GTV and a union of GTV and BTV with a threshold of 1.6. Recurrence pattern analysis was performed for the original PTV consisting of the clinical target volume (CTV = GTV + 20 mm) and a 3 - 5 mm PTV margin, the original 95-percent isodose, a synthetic PTV derived from the GTV with a fixed 23 mm margin and a union of GTV and BTV with a 18 mm margin. Treatment planning was performed on OTP-Masterplan® and GTV delineation was based on contrast-enhancing regions on MRI or postoperative resection cavity, respectively. Results 36 glioblastoma patients receiving FET PET before initial RT were included in the recurrence pattern analysis. Median age was 65.5 years and median KPS was 90 at the beginning of RT. Median follow-up time from the start of RT was 47 months, median overall survival 23 months and median progression-free survival 7 months. The minimal margin including the recurrent tumor was smaller for the union of GTV and BTV (median 12.5 mm) than for the GTV alone (median 17 mm) with statistical significance on Wilcoxon-Test (p<0.001). Recurrence pattern analysis revealed 33 (91.6%) central, 1 (2.8%) in-field, 1 (2.8%) marginal, 1 (2.8%) ex-field recurrences for the original PTV, 30 (83.3%) central, 3 (8.3%) in-field, 2 (5.6%) marginal, 1 (2.8%) ex-field recurrence for the 95-percent isodose line, 30 (83.3%) central, 4 (11.1%) in-field, 2 (5.6%) ex-field recurrences for GTV with a 23 mm margin and 32 (88.8%) central, 2 (5.6%) in-field and 2 (5.6%) ex-field recurrences for the union of GTV and BTV with a 18 mm margin. The union of GTV and BTV with a 18 mm margin was smaller than the GTV alone with a 23 mm margin (p=0.053 in Wilcoxon-Test). Conclusion FET PET derived BTVs may reduce the minimal margin at the primary RT of glioblastoma patients with an improved delineation of biologically active tumor. The union of GTV and BTV with 18 mm was smaller than GTV with a 23 mm margin and had better results on recurrence pattern

analysis (which corresponds to a 15 mm CTV expansion). These findings have to be prospectively validated. PO-0639 Feasibility of tract based dosimetric analysis in brain tumor patients M. Conson 1 , L. Cella 2 , E. D'Ippolito 3 , F. Piccolo 3 , S. Cocozza 3 , V. D'Avino 2 , R. Liuzzi 2 , M. Quarantelli 1 , R. Pacelli 3 1 Intituto di Biostrutture e Bioimmagini, University of Naples "Federico II"- Department of Diagnostic Imaging and Radiation Oncology- Napoli- Italy, Napoli, Italy 2 Intituto di Biostrutture e Bioimmagini, National Council of Research CNR, Napoli, Italy 3 Department of Advanced Biomedical Sciences, University of Naples "Federico II"- Department of Diagnostic Imaging and Radiation Oncology- Napoli- Italy, Napoli, Italy Purpose or Objective Radiation induced white matter injuries are feared side effects of brain irradiation that may cause cognitive dysfunction. Diffusion tensor imaging (DTI) analysis is a valid tool to evaluate white matter (WM) integrity. A framework for the integrated analysis of tractographic and dosimetric data in brain tumor patients undergoing radiotherapy is proposed. The tool is able to simultaneously measure in specific white matter tracts the diffusion changes due to microstructural alterations and the radiation dose. Material and Methods Ten consecutive patients affected by high grade glioma were treated by conformal radiotherapy with multiple non-coplanar 6 MV photon beams from a linear accelerator with a total dose of 60 Gy in 30 daily fractions of 2 Gy. Two different time-points were planned for MRI execution (before starting RT and one month after RT). Each MRI study included, beside the DTI, three- dimensional T1-weighted gradient-echo sequence (MPRAGE) before and after i.v. injection of contrast medium, as well as FLAIR and TSE-T2W images in the axial plane. The MPRAGE and the TSE-T2W images were co- registered with the b0 image and were used for the delineation of the area of parenchymal distortion (deformed area, DA). The DA included the surgical cavity and/or the area of contrast enhancement and/or the area of abnormal signal on TSE-T2W images. Deterministic fiber tracking across the whole WM was then performed using the fiber assignment continuous tracking algorithm. Fractional anisotropy (FA), radial diffusivity (RD) and axial diffusivity (AD) maps were generated. A ROI-based approach was used to select the fiber tracts. All fiber tracts passing through the DA were excluded though belonging to one of the three structures considered. The fiber tracts were converted in DICOM-RT format. The MPRAGE sequences without contrast medium were used as reference images for the co-registration with the corresponding planning CT-scan. The RT structure-set files were transferred from the MRI-space into the planning CT- space using the resulting co-registration matrix. Superimposing the dose map from each patient, the dosimetric evaluation was performed. The detailed scheme of our framework is reported in Figure 1.