ESTRO 2021 Abstract Book

S1395

ESTRO 2021

Figure 1: Cross-section of a PET/TC DIR applying a) Eclipse and b) Velocity algorithms with their respective uptake contours. Both images have the same SUV range.

Results Figure 2 shows the percentage of the junction volume in relation to the total volume for each algorithm obtained for each patient, as well as the DSC. The average DSC is 0.8 ± 0.1 and average junction volume to total volume is 82.6% and 80% for Velocity (V) and Eclipse (E) respectively.

Figure 2: Volumetric characteristics of PET uptakes in ten patients of non-microcytic lung carcinoma. Of the cases analyzed, the discrepancy between both methods is 20%, with no appreciable trend. It is worth noting two outliers that deviate sensibly from the average DSC in opposite directions. The differences between the transformation matrices generated by Demons and B-Spline algorithms produce changes in positions and slight variations in the uptake values both in maxima and in thresholds defining the boundaries, causing misalignment between the segmented contours. Conclusion The results of both DIR algorithms with respect to the applied method of relative SUV threshold segmentation, are compatible and no bias is observed between them. PO-1672 Characterization of the geometric distortion and the quality of MR images used for spine SBRT T. Torfeh 1 , R. Hammoud 2 , S. Aouadi 2 , Y. Arunachalam 2 , S. Paloor 2 , N. Al-Hammadi 2 1 Hamad Medical Corporation, Radiation Oncology, Doha, Qatar; 2 Hamad Medical Corporation, Radiation Oncology , Doha, Qatar Purpose or Objective High-quality MR imaging represents a key element for spinal stereotactic body radiation therapy (SBRT) to improve target volume delineation and reduce radiation dose to normal tissues. MR geometric accuracy and image quality are essentials to determine the accuracy of thetreatment planning. The aim of this study is to investigate the geometric distortion and the image quality for the typical sequences used in spinal SBRT and over a clinically relevant Field of View (FOV). Materials and Methods Image quality and geometric distortions on a 1.5T GE MRI-SIM commissioned for RT planning are assessed using the anterior array coil along with the table coil. In-plane and through-plane distortions are characterized using a phantom that covers a large field of view of 200mm and consists of layers of foam embedded with a matrix of ellipsoidal markers.