ESTRO 2022 - Abstract Book

S174

Abstract book

ESTRO 2022

Conclusion High quality pRads can be obtained with the proposed integrated mode imaging, with quality comparable to single particle tracking. The approach is fast, requires no calibration, and can naturally be extended to proton CT. Future work includes an experimental validation, evaluation of real-time tracking abilities, and potential use for patient positioning.

MO-0217 Deep learning auto-segmentation vs human inter observer variability of normal tissue in the brain

J. Kallehauge 1 , C.S. Byskov 2 , C.R. Hansen 3 , Y. Lassen-Ramshad 1 , A. Trip 1 , S. Lukacova 2 , L. Haldbo-Classen 2 , E.L. Lorenzen 4

1 Aarhus University Hospital, Danish Centre for Particle Therapy, Aarhus N, Denmark; 2 Aarhus University Hospital, Department of Oncology, Aarhus N, Denmark; 3 Odense University Hospital, Laboratory of Radiation Physics, Odense, Denmark; 4 Odense University Hospital, Laboratory of Radiation Physics, Odense, Denmark Purpose or Objective Automatic segmentations of organs at risk (OAR) are often compared to single observer delineations, which especially for small organs may result in higher inter-observer disagreement. A more fair comparison is to validate the auto-segmentation to a dataset where multiple observers have defined the same structures. Hence the objective of this study was to validate a Deep Learning semantic segmentation of selected OAR according to the Danish Neuro Oncology (DNOG) guidelines on multi-observer gold standard segmentations. Materials and Methods A 3D U-Net architecture was trained to segment Brainstem, Hippocampi, Chiasm, Pituitary, and Optical Tracts on 58 glioma patients' 3D T1w MRI of the brain (46 patients were scanned on a 3 T MRI and 12 were scanned at 1.5 T). This network was subsequently evaluated on a holdout test set consisting of 13 patients (nine patients were scanned at 3T MRI and four were scanned at 1.5T MRI). The performance of the automatic segmentation was finally evaluated using Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95) on three randomly chosen patients from the gold standard data having 13 patients in total. Each structure for each of the three patients were defined on average by 8 radiation oncologists with experience in neuro-oncology. The automatic segmentation was compared to each of the 8 expert segmentations and the mean metric HD95 and DSC was determined and compared to the mean expert metrics. Results The auto-segmented structures performed similar to expert delineated structures in respect to DSC and HD95. (figure 1 and figure 2). Although, the optical tracts appear to perform slightly worse when comparing DSC. The auto-segmentation showed similar or reduced variability especially for the pituitary gland but generally across all the investigated structures. Few expert segmentations were identified with large deviations from the median metrics indicative of being outliers. Especially some segmentations with HD95 well beyond the interquartile range are suspicious.