Abstract Book

S108

ESTRO 37

Radiosurgery (GKRS) alone or in conjunction with other techniques has been increasingly used for salvage treatment of recurrent grade III-IV gliomas. The steep dose fall-off of the radiosurgery treatment and the definition of the target as the lesion detectable in the diagnostic images (without margins) make the presence of the microscopic tumor spread a matter of concern when dealing with high grade gliomas. Aim of this work is to investigate the impact of the presence of tumor cells beyond the target boundaries on the treatment outcome of radiosurgery for invasive high grade gliomas. Material and Methods A novel mathematical model able to predict the spatial density distribution of cancerous cells outside the tumor boundary was implemented on a virtual brain tumor. The algorithm calculates the concentration of tumor cells beyond the delineated tumor border by accounting for the motility of cancerous cells through the host tissues relative to white matter. Calculations are iteratively performed until agreement with histopathological observations is reached. Based on the density distribution resulting from the mathematical model, three targets are defined: the Gross Tumor Volume (GTV), the volume having 100%-60% of the GTV cell density (zone 1), and the volume having 60%-0% of the GTV cell density (zone 2). Eight different treatment plan strategies were considered with different dose prescription levels administered to GTV and peripheral zones. In plan 1, 20 Gy at 50% isodose level were prescribed to the GTV implicitly accounting for the presence of the microscopic cell spread following the judgement of an expert planner (spatial dose distribution resulted as not strictly conformal to the delineated GTV). In plan 2, 20 Gy at 50% isodose were kept strictly conformal to the GTV. In plans 3 to 6 the target was zone 1 and the dose prescribed was 20 Gy at the 50%, 40%, 30%, and 20% isodose level, respectively. Plans 7 and 8 had as target zone 2 and dose prescriptions equal to 20 Gy at 50% and 20% isodose level, respectively. Results Treatment plans were evaluated based both on conformity index and tumor control probability (TCP). Results are shown in Table 1. As seen, treatment plans with similar values of conformity indexes resulted in rather different TCP values. As an example, results for one of the plans considered in this study are presented in Figure 1.

Conclusion Accounting for the potential infiltration of tumor cells outside the delineated GTV is of key importance in GKRS. Radiobiological parameters (e.g. tumor cell density distribution after treatment and TCP) have the potential to differentiate between otherwise similar plans and could be used for treatment plan evaluation together with well-established conformity indexes. PV-0200 Benchmark Case results from the EORTC Lungtech trial of SBRT for patients with centrally NSCLC M. Lambrecht 1 , E. Clement 2 , J.J. Sonke 3 , U. Nestle 4,5 , S. Adebahr 5 , M. Guckenberger 6 , D.C. Weber 7 , C.W. Hurkmans 1 1 Catharina Hospital, Department of Radiation Oncology, Eindhoven, The Netherlands 2 EORTC, Headquarters, Brussels, Belgium 3 The Netherland Cancer Institute, Department of Radiation Oncology, Amsterdam, The Netherlands 4 Kliniken Maria Hilf, Department of Radiation Oncology, Moenchengladbach, Germany 5 University Medical Center Freiburg, Department of Radiation Oncology, Freiburg, Germany 6 University Hospital Zurich, Department of Radiation Oncology, Zurich, Switzerland 7 Paul Scherrer Institute- ETH DomainSwitzerland, Center for Proton Therapy, Villigen, Switzerland Purpose or Objective The European Organisation for Research and Treatment of Cancer (EORTC) prospective multicentre Lungtech trial assesses the safety and effectiveness of SBRT in patients with inoperable centrally located non-small cell lung cancer. As part of the quality assurance, a Benchmark Case (BC) was distributed for delineation and planning. Material and Methods BCs were reviewed stepwise; only when delineations were deemed acceptable the treatment plan was generated and subsequently reviewed. A retrospective study was performed of the 15 first accepted submissions looking at the dice similarity coefficients of the OARS and the target delineations. In addition, a set of gold standard CTV contours combined with each institutions specific CTV to PTV margin were linked to the dose submissions to evaluate the influence of delineation on dosimetry. Results Four (27%) BCs were accepted at their first attempt. Ten (67%) institutions had to revise their delineation at least once and three had to revise their planning once. The mean dice coefficient varied significantly between OARs: The lowest values were found for the brachial plexus 0.23 (0.11-0.58;average volume 9.0 cm 3 ) and the highest for the spinal cord 0.88 (0.78-0.94;average volume 57.8 cm 3 ). The mean PTV dice coefficient was 0.71 (0.63- 0.95). Applying the gold standard contours, only one