ESTRO 38 Abstract book

S1126 ESTRO 38

ray image and conventional DE image which uses a constant ω throughout the image. Signal-to-noise ratio (SNR) of regions with different soft-tissue and bone thicknesses in step phantom was investigated for both DE techniques. Results The optimal pre-calculated ω values were found to be in a range [0.82, 1.19] depending on the region thickness in step phantom (Figure 1a). The ω values that cancels 25 cm (ROI1) and 10 cm (ROI2) soft-tissue overlapped with 3 cm bone were calculated as 0.93 and 0.83, respectively. Figure 1d demonstrates the ω image of Rando for the given projection view. Figure 2 demonstrates the effectiveness of the PP-DE algorithm compared to the clinical single energy x-ray image and conventional DE images (with ω=0.93 and ω=0.83). The SNR of DE image for ROI1 and ROI2 of the step phantom was 45 and 674 for PP-DE compared to 43 and 203 for conventional DE, respectively.

Fig. 1. Example image. Conclusion

A technique to provide enhanced IQ from a radiotherapy CT scanner has been developed and demonstrated qualitatively to be superior to conventional scans. Major benefits of the technique include no additional time required for the scan, no extra radiation dose and no motion effects introduced with use of the second scan. Further work including a contouring study is planned. EP-2049 Patient specific pixel-based weighting factor bone-only dual-energy x-ray imaging S. Darvish-Molla 1 , M. Reno 2 , M. Sattarivand 1,2,3 1 Nova Scotia Health Authority, Radiation Oncology - Medical Physics, Halifax, Canada ; 2 Dalhousie University, Physics and Atmospheric Science, Halifax, Canada ; 3 Dalhousie University, Radiation Oncology, Halifax, Canada Purpose or Objective Implement the novel pixel-based weighting factor dual energy (DE) algorithm for effective soft-tissue suppression throughout the image to enhance the bone-only image quality and to overcome the limitation of the conventional DE algorithm with constant weighting factor, which is restricted to regions with uniform patient thickness. Material and Methods A step phantom was constructed consisting of slabs of solid water and average bone materials. Thicknesses of bone ranged from 0 to 6 cm in one direction, and solid water from 5 to 30 cm in the other direction, creating 7 by 7 regions of interest (ROIs). The slabs were stacked and placed on a custom-made plastic stand, angled at 42°, such that the central axis of the x-ray beam was perpendicular to the phantom surface. Using Brainlab’s ExacTrac system, projection images of the step phantom were acquired at (60 kVp, 40 mAs) and (140 kVp and 12 mAs). Using an in-house Matlab code, the optimal pre- calculated weighting factors (ω) were found by varying values in the range 0-2 for each ROI such that the soft tissue contrast-to-noise ratio (CNR) reached zero. CT images of Rando phantom were acquired using thorax protocol. Ray tracing technique was used to calculate the bone and soft-tissue DRRs (digitally reconstructed radiographs) (Figure 1b,c) . Weighting factors for different pixel locations on the DRRs were interpolated (or extrapolated) using the pre-calculated weighting factors and the ω image for Rando phantom was generated. By applying the ω image into the DE algorithm and utilizing noise suppression algorithm, the bone-only DE image was generated and compared to the clinical single energy x-

Conclusion Bone-only patient specific pixel-based algorithm was successfully implemented, which could have clinical interests for image guidance for spine SBRT patients. Compared to the conventional DE, this novel technique creates images with improved bone contrast, specifically in soft tissue overlapping regions, improved SNR, thus enhanced image quality for precise tumor localization. EP-2050 Implementation of CT-based attenuation maps of RT positioning devices in PET/MRI - online vs offline