ESTRO 2023 - Abstract Book

S1984

Digital Posters

ESTRO 2023

normal lung patient is compared to illustrate that the model can generate realistic RP patterns(Shown in Figure 2). Second, after randomly splitting the dataset into a training set and a test set at a ratio of 7:3. we trained a DL(RseNet10) model for RP classification on the training set with and without the synthetic images and tested it to show that adding synthetic images can improve DL performance.

Figure 2. Representative results of our work. Every column is a single patient. The “normal” row is the CT without radiation pneumonitis. In “RP added” row, our GAN-based method has added radiation pneumonitis to the image. The different row contains the difference between these two images to show what the algorithm has changed in the image. Results As shown in Figure 2, the synthetic RP images have more complex textures in the lungs compared to normal lungs. As shown in table 1, the performance of the DL model is lower if it is trained only using the original dataset. Synthetic data can improve the model's classification ability, resulting in an improvement in 46% of sensitivity,3% of accuracy and 12% of overall AUC. Table1 Classification model performance trained on the dataset. methods AUC ACC SEN SPE Original data 0.55(0.52-0.57) 0.63(0.60-0.65) 0.27(0.26-0.29) 0.79(0.77-0.81) Original data + synthetic data 0.67(0.66-0.70) 0.66(0.64-0.67) 0.73(0.71-0.77) 0.63(0.60-0.64)

Conclusion Our method can generate realistic RP images and can be utilized as a RP data augmentation method to assist in overcoming the sample imbalance problem and lack of annotated data.

PO-2204 Radiation affects bronchial epithelial progenitor function as assessed by organoid formation assay

M. Kuipers 1 , D. Ninaber 1 , A. Slats 1 , P. Hiemstra 1

1 Leiden University Medical Centre (LUMC), Pulmonology, Leiden, The Netherlands

Purpose or Objective Radiation-induced lung injury (RILI) is a serious side-effect of radiotherapy for lung cancer. We hypothe that radiation- induced effects on the normal lung epithelium play a key role in development of RILI. Therefore, the aim of the present study was to study the effect of ionizing radiation (IR) on cultured primary bronchial epithelial cells (PBEC). Materials and Methods PBEC were cultured submerged/undifferentiated (S-PBEC), at the Air-Liquid-Interface (ALI-PBEC) to allow mucociliairy differentiation, and as 3-dimensional organoids (3D-PBEC). Cells cultured according to these three methods were exposed to ionizing radiation (IR; 0.25, 0.5, 1, 2, 4, 8 Gray (Gy) ), dissociated and reseeded in an organoid model in similar densities 1h, 24h, 7 days and 14 days after IR to observe organoid formation. In addition, in the ALI-PBEC model, a qPCR was performed at 3h, 6h, 24h and 48h after IR, as well as ELISA on conditioned media after 72h after IR, to assess effects of IR. Immunofluorescent staining of γ H2Ax-foci was performed to asses DNA-damage of these cells. Results IR had a marked effect on organoid formation: organoid formation was severely impaired following IR irrespective of culture model in which cells were exposed to IR and at all time points after IR. The effect was strongest with IR-exposed ALI-PBEC, with already a 50% reduction in organoid formation after 0.25 Gy and hardly any organoid formation with an 8 Gy dose (figure 1).