Click here to


Are you sure ?

Yes, do it No, cancel

Beam-Energy and Depth-Of-Interaction Estimation in MV Imaging Using a Dual-Ended Readout

M Myronakis1*, R Berbeco1, (1) Department of Radiation Oncology, Brigham and Women's Hospital, Dana-Farber Cancer Institute & Harvard Medical School, Boston, MA


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: imaging devices have limitations that hamper their application in adaptive radiotherapy (ART). The purpose of this work is to propose a paradigm shift in x-ray imaging detector design to provide beam-related information that can be used to improve image quality and ART.

Methods: pixelated bismuth germanate (BGO) scintillator with 3mm thickness and 0.5mm pixel pitch, was modelled in the GATE Monte Carlo toolkit. Two flat-panel detectors were used to detect the optical signal at the x-ray beam entrance (top) and one at the exit (bottom) of the scintillator. Monoenergetic beams were used to calibrate signal proportions, beam-energy and average depth-of-interaction in the scintillator (Davg). A phantom with three tissues (soft-tissue, bone and lung) was modeled. A 20x20 cm2 square field was simulated with a validated 2.5MV x-ray beam. An image of Davg and an image of the average x-ray beam energy (Eavg) per pixel were generated using the two signal proportions and the calibration curve. Regions-of-interest (ROIs) were defined at each tissue and Eavg and Davg were calculated.

Results: estimated Eavg and Davg for the ROIs in the soft-tissue, lung, and bone ROIs were sufficiently different to enable clear identification. Eavg was 0.148MeV (±5.0×10?4), 0.147MeV (±4.4×10?4), 0.152MeV (±8.0×10?4), respectively. The background (air) had an Eavg of 0.143MeV (±3.5×10?4). The estimated Davg for the same ROIs was 0.88mm (±4.7×10?³), 0.87mm (±4.2×10?³), and 0.91mm (±4.2×10?³) respectively. The background (air) had Davg of 0.84mm (±3.3×10?³).

Conclusion: proof-of-concept imager design with dual-ended detectors and pixelated scintillator demonstrated that x-ray beam energy and depth of interaction can be directly retrieved using signal proportions. Further work is required to optimize the design for improved energy separation and depth estimation in clinical conditions. The readout of average energy per pixel has potential utility in novel image-reconstruction algorithms, in-vivo dosimetry and adaptive radiotherapy, among other clinical applications

Funding Support, Disclosures, and Conflict of Interest: The project described was supported by Varian Medical Systems.


Megavoltage Imaging, Electronic Portal Imaging


IM/TH- RT X-Ray Imaging: Detector development & evaluation

Contact Email