ESTRO 2023 - Abstract Book

S1322

Digital Posters

ESTRO 2023

Purpose or Objective Meningioma is usually irradiated with a dose of 12 Gy in a single fraction. In cases where the tumor overlaps critical organs. or the tumor volume is significant, fractional irradiation is used. The goal of this study is to evaluate the automated treatment planning and irradiation of patients with meningioma using a combination of Knowledge-Based Planning (RapidPlan, RP) with HyperArc automated brain irradiation. Materials and Methods Data from 57 manually prepared clinical HyperArc plans with a prescribed dose of 54 Gy delivered in 30 fractions were used to create a model in RapidPlan module. The model was validated on 15 cases not used in the model by comparing dose distributions in the RP-generated and clinical plans. For PTV were compared D98%, D95% and D2%. For all organs at risk (brain stem, optic nerves, chiasm, eyes, lens and inner ears) Dmax and mean dose statistics were analyzed. Eclipse TPS (Varian Medical Systems, v15.6) and AXB algorithm were used, X6MV FFF. One optimization was performed without the planner’s intervention. The Wilcoxon signed pairs rank test was used to test the statistical difference between the analyzed parameters (p<0.05). Results Knowledge-based plans can achieve comparable PTV coverage as clinical plans for D98%, D95% and D2% (96.7±1.2 Gy vs 96.4±1.7 Gy, 97.2±1.0 Gy vs 96.8±1.7 Gy and 102.1±0.9 Gy vs 102.1±0.5 Gy, respectively). The doses to all OARs are also comparable to the clinical plans. Dose differences were not statistically significant, except for Dmax for the optic nerves left and right, left eye and lens (11.3±11.1 Gy vs 13.4±11.5 Gy, 18.3±13.0 Gy vs 21.3±14.8 Gy, 5.5±4.2 Gy vs 7.3±5.4 Gy and 2.6±0.8 Gy vs 3.2±1.3 Gy, respectively ) and mean dose for left eye and lens ( 2.1±1.1 Gy vs 3.0±1.8 Gy and 1.6±0.7 Gy vs 2.0±1.0 Gy, respectively). For these organs, the automatically generated plans obtained lower doses than in the clinical plans. Conclusion We have demonstrated that it is possible to achieve in a single optimisation a dose distribution equivalent to that in clinical plans, which, combined with the HyperArc technique, makes it possible to reduce the preparation and treatment time for meningioma patients. 1 Delft University of Technology, Technical Medicine, Delft, The Netherlands; 2 Leiden University Medical Center, Radiation Oncology, Leiden, The Netherlands Purpose or Objective Adjuvant radiotherapy (RT) after breast-conserving surgery for breast cancer reduces the risk of local recurrence and increases overall survival. Accurate segmentation of the target volumes and organs at risk (OARs) is a crucial step in the RT workflow to deliver an adequate therapeutic radiation dose to the target volumes while sparing the OARs. Target volumes comprise the tumor bed, whole breast, and the axillary (L1-L4) and interpectoral (IP) lymph nodes. Delineating these target volumes is a labor intensive task and is known to have a high interobserver variability (IOV). This study aims to compare the IOV of manual segmentations of the regional lymph node levels to the IOV of deep learning (DL)-based segmentations manually corrected by radiation oncologists (ROs). Implementation of the DL-based segmentation in the RT workflow as initial segmentation might improve consistency and reproducibility, and reduce the workload of ROs. Materials and Methods Five experienced ROs delineated the axillary (L1-L4) and IP lymph node levels on the planning CT in the RayStation TPS (RayStation 10B, RaySearch Laboratories) of two breast cancer patients requiring locoregional adjuvant RT. Additionally, DL-based segmentations were generated using a 3D U-net convolutional neural network (CNN) available in the TPS, which was pretrained on delineations following the ESTRO delineation guidelines. The ROs corrected the DL-based segmentations to clinically acceptable segmentations after a two-week interval. The time to create the manual and cDL segmentations was recorded in minutes for each RO. IOV within both the manual segmentations and corrected DL (cDL) segmentations was compared using the Dice Similarity Coefficient (DSC), surface DSC (sDSC) and 95% Hausdorff Distance (HD). A Wilcoxon signed-rank test was conducted to assess the statistical differences in IOV between the manual and cDL segmentations. Results An example of the delineations is shown in Figure 1. The median DSC of all lymph node levels combined for the manual and cDL segmentations was 0.67 (± 0.071 IQR) and 0.90 (± 0.040 IQR), respectively. The cDL segmentations had a significantly higher DSC for all individual lymph node levels (Figure 2). Additionally, all median 95% HD and sDSC were improved (Table 1). The median time investment for the manual and cDL segmentations was 13 minutes (10-17) and 9 minutes (5-16), respectively. PO-1627 Interobserver study of deep learning-based segmentation for nodal target volumes in breast cancer C. Adriaans 1,2 , F. Dankers 2 , C. Papalazarou 2 , N. Hoekstra 2