ESTRO 2020 Abstract book

S812 ESTRO 2020

distributions. For the other 20 patients, VOLO planning was performed without knowledge of what was achieved in the SO plan dose distribution. Prior to dosimetric comparisons, all VOLO and SO plans were renormalized in order to exactly have the clinically required PTV coverage for the prescribed dose: 95% for prostate and 98% for lung. Dosimetric evaluations were based on the clinical protocols. Numbers of nodes, beams, segments, kMU, as well as planning and treatment time were also compared. Two-sided Wilcoxon signed-rank tests were used for statistical analyses with p-values lower than 0.05 indicating statistical significance. Results For prostate cancer, VOLO was highly superior to SO for all clinical requirements apart of D 50% in urethra (39.7 Gy vs 40.0 Gy, respectively). Significant reductions of rectum and bladder mean doses of 34.3% and 23.5%, respectively, were observed, as well as reduction of rectum and bladder D 1cc (30.0 Gy vs 31.5 Gy and 36.7 Gy vs 38.7 Gy, respectively) and urethra D 5% (41.4 Gy vs 43.1 Gy) (all p<=0.001, see also Fig.1). For lung, VOLO dosimetric plan quality was comparable to SO for D 0.03cc in esophagus, trachea and spinal cord. However, significant advantages of VOLO were found for main bronchus D 0.03cc (4.0 Gy vs 4.7 Gy, p=0.02) and thoracic wall D 30cc (21.5 Gy vs 22.6 Gy, p=0.008). For both prostate and lung plans, VOLO was highly favorable compared to SO for the number of applied nodes (48 vs 57 and 40 vs 47, respectively), number of beams (48 vs 71 and 40 vs 47, respectively), kMU (25 vs 47 and 17 vs 24, respectively) and total treatment time (27 min vs 34 min and 27 min vs 33 min, respectively), all p<=0.001. Details are presented in Fig.2. Planning time was also reduced with VOLO: 1 h 10 min vs 6 h and 1 h 40 min vs 4 h for prostate and lung, respectively.

Conclusion The newly implemented VOLO optimizer allowed to substantially improve dosimetric plan parameters. Moreover, VOLO plans showed large reductions in MU, number of nodes and beams, planning time and treatment time. This study show that more investigations on commercially available optimizers are needed in order to avoid sub-optimal treatment with high-end treatment units. PO-1436 Robustness of robustly optimized versus margin based treatment plans for left-sided breast cancer R. Vissers 1,2 , N. Bakx 1 , E. Hagelaar 1 , H. Bluemink 1 , C. Hurkmans 1 1 Catharina Hospital, Radiation Oncology, Eindhoven, The Netherlands ; 2 Eindhoven University of Technology, Applied Physics, Eindhoven, The Netherlands Purpose or Objective Photon radiotherapy conventionally uses a PTV to ensure adequate CTV dose coverage. As the PTV-margin is not adapted to tissue density variations, it becomes flawed near the lungs. Hence, robust planning could improve the treatment quality for left-sided breast cancer patients. Furthermore, since robust plans are optimized and evaluated differently, there are currently no set criteria to compare the quality of robust and non-robust plans to each other. Therefore, this study aims to explore the potential of robust planning for left-sided breast cancer and compare its robustness to non-robust plans. Material and Methods Twelve randomly selected left-sided and node-negative breast cancer patients with a prescribed dose of 15 times 2.67 Gy were re-planned and evaluated in RayStation (RaySearch Laboratories). Non-robust plans used a margin of 5 mm, were clipped 5 mm below the skin, and were optimized for adequate dose coverage ( D 98 ≥ 95 % and V 107 ≤ 2 %) within the PTV. Robust plans were optimized to meet CTV D 98 ≥ 95 % in fourteen error scenarios assuming a universal setup uncertainty of 5 mm and CTV V 107 ≤ 2 % in the nominal scenarios. After fulfilling these goals, the mean heart and lung doses were minimized as much as possible. The robustness of all plans was measured in fourteen error scenarios for a setup uncertainty of 5 mm and evaluated in the voxel-wise (VW) minimum, VW maximum, VW mean,