ESTRO 36 Abstract Book

S838 ESTRO 36 _______________________________________________________________________________________________

specialising in H&N RT and their assessments were compared with quantitative analysis. Optimisation objectives and priorities and the use of dummy organs were analysed in conjunction with the dose distributions. Results Each plan met the pre-defined PTV and PORV doses. There was no significant variation in HI (1.05-1.09) regardless of the priorities applied in the optimisation. A larger variation in CI (1.07-1.18) was attributed to use of the Normal Tissue Objective function. There were variations in dose to normal tissues as planners applied varying dose constraints to keep doses as low as reasonably practicable without compromising PTV coverage. Clinician reviews picked up more subtle but potentially clinically relevant variations between plans – high dose spill, dose spread across mid-line and over- zealous sparing of normal tissues. The plans scored most highly by the clinicians were created by the most experienced planners who took less time to reach an “optimal” solution. Conclusion There is a relatively consistent approach to H&N VMAT planning within our institution. This study has highlighted where differences between optimisation objectives and priorities and the use of dummy structures can lead to subtle variations in dose distributions which may not be detected in quantitative analysis but may affect acceptability. More experienced planners familiar with clinician expectations demonstrated improved judgement when determining an optimal plan. Shared learning has enabled a more consistent approach to H&N plan optimisation and an improved understanding of what is achievable and clinically acceptable. This has benefitted both planners and clinicians. In addition, the creation of plan optimisation templates, based on the findings of this study, are aimed at a consistent optimisation resulting in improvements in the patient pathway. EP-1556 Dosimetric commissioning of a TPS for a synchrotron-based proton PBS delivery system G. Kragl 1 , T. Böhlen 1 , A. Carlino 1,2 , L. Grevillot 1 , H. Palmans 3 , A. Elia 1 , B. Knäusl 1 , J. Osorio 1 , R. Dreindl 1 , J. Hopfgartner 1 , S. Vatnitsky 1 , M. Stock 1 1 EBG MedAustron GmbH, Medical Department, Wiener Neustadt, Austria 2 University of Palermo, Department of Physics and Chemistry, Palermo, Italy 3 National Physics Laboratory, Radiation Dosimetry, Teddington, United Kingdom Purpose or Objective To provide an overview regarding dosimetric commissioning of the TPS RayStation for proton PBS delivery installed at a synchrotron-based dual particle facility. 1D/2D commissioning consisted of benchmarking the dose calculation algorithm against measured IDDs, on- axis lateral spot profiles in air and field size factors as well as comparisons of absolute dosimetry. 3D commissioning consisted of absolute dose comparisons in the SOBP of cubic targets in water as well as the characterization of 3D dose distributions with increasing complexity. A robotic patient positioning system was used rather than extendable snouts to reduce the air gap between patient and nozzle. Therefore, special attention was paid to non- isocentric setups. Material and Methods Commissioning was performed for the PB algorithm (version 3.5) integrated in RayStation (version 5.0.2). IDDs were acquired with a Bragg peak chamber (PTW) and corrected for insufficient detector size by means of MC simulations (GATE/GEANT4). Spot profiles in air were acquired with a scintillating screen (Lynx, IBA) at 7 air gaps. Absolute dosimetry was performed with a Roos chamber (PTW) in 12 x 12 cm 2 fields (2 mm lateral spot spacing) for 20 energies. Field size factors were acquired

with a semiflex ionization chamber (PTW) at 3 depths in water for field sizes ranging from 2 x 2 to 20 x 20 cm 2 . 3D dose distributions were characterized using 24-PinPoint chamber arrays (PTW). Results Calculated ranges agreed within 0.2 mm with measured ranges. The integrals of measured and calculated IDDs agreed within 0.5% for clinically relevant ranges. At isocenter, calculated and measured spot sizes (FWHM) differed on average less than 0.4 mm in x- and y- directions. For non-isocentric setups differences were within 0.5 mm. Field size factors always agreed within 4%; deviations were generally low (<1%) and increased only at small field sizes and the highest energies (range >30 cm). For isocentric arrangements, absolute dose agreed within 2.2% in the center of SOBPs of cubic targets with different sizes and at different depths. As expected, the deviations increased for plans with range shifter for non-isocentric arrangements. Variations of up to 3.5% were obtained for modulation widths of 6 cm. Results for more complex geometries are currently under investigation. Conclusion Clinically acceptable results were obtained for open beams. For plans with range shifter, a scaling of dose distributions might be considered until the upcoming MC dose calculation algorithm is available. Minimizing the air gap to reduce modelling inaccuracies with respect to scattered protons in air is beneficial for these cases and realized by non-isocentric treatments. EP-1557 Minimum prescription concept for dose painting increases robustness towards geometrical uncertainty S. Korreman 1 1 Aarhus University Hospital, Department of Oncology, Aarhus C, Denmark Purpose or Objective Dose painting radiotherapy with heterogeneous dose escalation is vulnerable to geometrical errors, which potentially deteriorate the benefits of dose escalation substantially. This study investigates use of a minimum prescription concept to increase plan robustness towards geometrical uncertainties. Material and Methods Dose escalation was prescribed based on PET Cu-ATSM tracer uptake for a head and neck cancer patient, with a high degree of heterogeneity in the uptake. The minimum dose was 60Gy, and dose escalation was prescribed based on a linear correspondence model to the tracer uptake, with a maximum escalation up to ~90Gy. Dose painting plans were optimized using the Eclipse treatment planning system, using modulated arc therapy technique in a contour-based dose escalation scheme (5 levels). Two planning strategies were tested: (1) Minimum and maximum dose constraints imposed on all subvolumes (exact-map), and (2) minimum constraints on all subvolumes with only one overall maximum constraint (minimum-map). Geometrical error was simulated by displacing the isocenter with up to 2 mm. Quality index metrics were compared for the two planning strategies. Results For both strategies, optimizations could be performed with good adherence to dose constraints. For the exact- map technique, the fraction of voxels with quality index within plus/minus 5% of prescription dose was ~79%, and the fraction of voxels above 95% of prescription dose was ~93%. For the minimum-map technique, the fraction of voxels above 95% of prescription dose was ~97%. With displacement of 2 mm, the >95% fraction changed to ~85% for exact-map, and ~95% for minimum-map technique.