ESTRO 2023 - Abstract Book

S2092

Digital Posters

ESTRO 2023

bulb), and dose calculations were finally done using the Acuros XB algorithm (dose-to-medium). Comparison of both plans was done using the metrics required by the PACE-B trial: D2%, D98% and maximum dose (Dmax )for the PTV; V40 Gy for the prostate; D5% for the femoral heads; D50%, D20% and D1cm3 for the rectum; D40% and V37Gy for the bladder; and D50% for the penile bulb. In addition, the total delivery times (TDTs) of both plans were compared. A 2 tailed Student t-test ( α = 0.05) was performed to evaluate whether there is a significant difference between both types of plans. Results 1) For each case, both plans met the dose constraints for the target dosage and organs-at-risks required by the PACE-B protocol. 2) Table 1 shows the paired comparison for the metrics. Dual arc VMAT was significantly worse than single arc VMAT for D50% and D20% of the rectum, while dual arc VMAT was significantly better for femoral heads. No significant differences were found for the rest of the metrics.

3) Single arc VMAT provided a significant reduction of the treatment delivery time (1.6 vs. 2.5 min)

Table 1. Statistical comparison between single and dual arc VMAT. Mean and range of values of each metric are displayed. Asterisk mark indicates a significant difference (p < 0.05).

Conclusion Single arc VMAT technique is sufficient for prostate SBRT planning, allowing reducing significantly the delivery time. This is a very important issue in order to minimize the probability of intra-fractional prostate motion (RadiotherOncol. 2020;151:88-94) when an online adaptive strategy is done. The one arc-based VMAT approach is routinely performed in our department for online adaptive prostate SBRT (Pract Radiat Oncol. 2022 Mar-Apr;12(2):e144-e152).

PO-2324 The efficiency of Atlas-Based Segmentation for head and neck patients used in RayStation

L. van Klaveren 1 , J. van der Klein 1 , P. Rietveld 1 , E. Kouwenhoven 1 , J. van Egmond 1 , M. van Ruler 1 , J. Zindler 1 , M. Mast 1

1 Haaglanden Medisch Centrum, Radiotherapy, Leidschendam, The Netherlands

Purpose or Objective Contouring the Organs at Risk (OAR’s) in head and neck cancer patients for the purpose of generating a treatment plan is time consuming. The planningsystem RayStation contains a semiautomatic contouring function, i.e. Atlas-Based Segmentation (ABS), based on contours manually drawn by experts. With this function OAR contours can be generated on a new CT-scan. The question is if these generated structures have the same quality compared to manually drawn structures and how much time can be saved in the delineation process. Materials and Methods The quality of the OAR was determined by the Dice Similarity Coefficient (DSC), this represents the overlap of manually drawn and atlas generated OARs. Also a visual check has been performed on several CT-scans to determine the amount of adaptations that will be needed after generating each OAR. Manually drawn contours of CT-scans of head and neck patients were used for a test phase. The test phase consisted of six atlases containing an increasing number of CT-scans, to determine the influence of atlas on quality of the delineated OAR. On 24 other CT-scans 9 OAR’s were manually contoured by experts only for this research. These scans were used to create the definitive atlas. This atlas was applied on several CT-scans that were also used to create the atlas and compared with the manually delineated OARs (internal cohort). And on CT-scans that already contained clinical used OAR’s (external cohort). The time has been measured on all the steps in the process.