ESTRO 38 Abstract book

S994 ESTRO 38

sparing of OARs and can result in a more homogeneous IMRT plan. EP-1833 Impact of physicist clinical practice on the quality of head and neck IMRT and VMAT plans S. David 1 , M. Hermida-López 1 1 Hospital Universitari Vall d'Hebron, Servei de Física i Protecció Radiològica, Barcelona, Spain Purpose or Objective During treatment plan design many decisions are taken based on the planner's own experience and potential biases. Although the treatment planning system (TPS) is plenty of similar plans that could help in the decision making, it is not easy to access to all this information. These data can be more accessible using Mobius3D (Mobius Medical Systems,LP), an independent dose calculator, own database. We aimed to assess differences in treatment plan quality for head and neck IMRT and VMAT plans between three experienced physicists in our center. Material and Methods After a plan is created, the RTplan, the planning CT and the RTdose DICOM files are exported to Mobius3D (M3D) database. Using a Python script provided by M3D (slightly modified), an Excel file was created for every plan in M3D database. A Matlab script was used to filter the head and neck plans with only two dose levels (54 Gy for the nodes and 70 Gy for the high-dose PTV). Plans from replannings and clinical trials were excluded to ensure there was no bias during plan creation. Sixty-three plans approved by one of our three physicists (16, 15 and 6 years of experience, respectively) were studied. Monitor units (MU) and 40 dosimetric parameters for PTVs and OARs, the most common reported in the literature, were evaluated. Statistically significant differences were analyzed with two-tailed two-samples unpaired t-tests (α=0.05). Results As expected, IMRT plans resulted in a bigger amount of MU (1211±423) than VMAT plans (556±89). A statistically significant difference was found between physicists A and C for IMRT plans. Although VMAT MU values are consistent with those reported in the literature, IMRT MU values are higher with those published by Cozzi et al. [IJROBP, 2004, 58.2]. A possible reason could be different fluence- smoothing procedures among physicists Statistically significant differences were found between physicist A and physicist C for the mean dose and conformity index for the high-dose PTV and for the mean dose, homogeneity, D(2%) and V(107%) for the low-dose PTV. Statistically significant differences were found between physicist A and physicist C for D(50%) and D(66%) for the left parotid, while between physicist A and B for mean dose, D(33%), D(50%) and D(66%) for the right parotid. Physicist A achieved the best result in the 45% of the evaluated parameters, physicist B in the 40%, and physicist C in the 15% left. The standard deviation, taken as a surrogate for the planning consistency, was the smallest in 52.5% of the parameters for physicist A, in 20% for physicist B, and in 22.5% for physicist C. Conclusion Analyzing plan quality parameters have showed differences between physicists, which leads to an opportunity to share practices to improve dosimetric results and decrease self-bias, thus resulting in a better and more homogeneous treatment plan quality. The developed procedure can be easily implemented for other sites and prescriptions. EP-1834 Feasibility study of extreme hypofractionated proton treatment planning for prostate cancer P. Klinker 1 , C. Brouwer 1 , C. Roos 1 , C. Hammer 1 , H. Vanhauten 1 , S. Bijmolt 1 , J. Langendijk 1 , S. Both 1 , A. Van den Bergh 1 , S. Al-Uwini 1

plans were created for each case: one in which the PTV included the laryngeal air cavity and one in which the air cavity was subtracted from the PTV to create a new PTV- air structure. Acuros XB (Varian Medical Systems, Palo Alto, CA) dose calculation algorithm was utilized. Dosimetric variables were collected for PTV-air structure from both IMRT plans, including V100% (volume that received the prescription dose), D98% (dose received by 98% of the volume), D2%, and D0.2%. Dosimetric variables for organs at risk (OARs), including spinal cord and the carotid arteries were also recorded. Homogeneity index (HI) defined as D98/D2 was calculated. Two-sided t-tests were used to compare dosimetric variables. Results The mean V100% was 95.1 ± 2.5% for the PTV plans and 95.4 ± 0.9% for the PTV-air plans, demonstrating adequate target coverage and no statistical difference between the plans (p=0.584). The mean D0.2% (maximum dose) within PTV optimized plan was significantly lower for the PTV-air optimized plans at 108.8 ± 2.1% compared to 111.4 ± 3.3% (p=0.0002). The HI also improved from 0.90 ± 0.03 up to 0.93 ± 0.02 (p<0.0001). Even with these gains, there were no significant differences in any of the OARs. When the PTV-air contour was superimposed onto the PTV plan, the V100% was significantly higher at 97.2 ± 2.3% compared to 95.4 ± 0.9% (p=0.003), showing that the PTV plan was unnecessarily hot. The D2% and D0.2% were also significantly higher (p=0.0002).

Conclusion The removal of the air cavity from the PTV for early-stage glottic cancers does not compromise PTV coverage or