ESTRO 2021 Abstract Book

S888

ESTRO 2021

Figure 1: Example of OARs in the automated atlas.

Results Manual adjustments to the 34 OARs were required in ≤ 5% of MRI slices, 6-25% of MRI slices, 26-50% of MRI slices, and > 50% of MRI slices for 44%, 15%, 21%, and 21% of OARs, respectively. In 7/34 (21%) of OARs, adjustments were required in > 50% of cases. These OARs are shown in table 1, as well as the percentage of MRI slices that required adjustments. The observer agreement and the mean volume difference per OAR are also shown in table 1. The outcome of the query from the radiation oncologists was that a significant amount of time was saved as a result of the automated OAR contours with manual fine-tuning compared to manual OAR contouring (approximately 15 minutes per patient versus 90 minutes per patient, respectively).

Conclusion Automated OAR atlas-based segmentation on MRI by the TPS makes OAR contouring according to the EPTN consensus clinically feasible in terms of quality and time. In a majority (79%) of OARs, adjustments were only needed in 4% of MRI slices on average. The process of adjusting the automated OAR contours can be done within 15 minutes. For clinical use, verifying and adjusting of all OARs by a radiation oncologist is still necessary. In the future, more patients will be included to further improve the accuracy of automated OAR delineation and save even more time. References 1. Eekers, D. B., in ‘t Ven, L., Roelofs, E et al.. Radiotherapy and Oncology, 128(1), 37–43. https://doi.org/10.1016/j.radonc.2017.12.013 PO-1067 Exploratory analysis of serial 18F-Fluciclovine PET and mpMRI during chemoradiation for glioblastoma K. Fatania 1 , R. Frood 1 , M. Tyyger 2 , S. Fernandez 3 , D.L. Buckley 4 , L. Murray 3 , A. Scarsbrook 5 , S. Short 6 , S. Currie 1 1 Leeds Teaching Hospitals Trust, Department of Radiology, Leeds, United Kingdom; 2 Leeds Teaching Hospitals Trust, Department of Medical Physics, Leeds, United Kingdom; 3 Leeds Teaching Hospitals Trust, Department of Clinical Oncology, Leeds, United Kingdom; 4 University of Leeds, Leeds Institute of Cardiovascular and Metabolic Medicine, Leeds, United Kingdom; 5 Leeds Teaching Hospitals Trust, Department of Radiology and Nuclear Medicine, Leeds, United Kingdom; 6 University of Leeds, Leeds Institute of Medical Research, Leeds, United Kingdom Purpose or Objective Positron emission tomography (PET) using anti-1-amino-3- 18 fluorine-fluorocyclobutane-1-carboxylic acid ( 18 F- fluciclovine) shows preferential glioma cell uptake with low activity in normal brain. Dynamic contrast- enhanced (DCE) MRI may be abnormal beyond gadolinium-enhancement on conventional post-contrast T1- weighted MR sequences (Gd-T1) of glioma and may reflect tumour infiltration beyond the enhancing margin. There is a paucity of data on how 18 F-fluciclovine uptake changes during radiotherapy and how this correlates with Gd-T1 and DCE-PWI activity. The aim of this pilot study was to compare glioblastoma-related 18 F-fluciclovine uptake, DCE-MRI and Gd-T1 MRI signal intensity over the course of chemoradiotherapy treatment. Materials and Methods