Share Email Print
cover

Proceedings Paper

Investigation of calibration-based projection domain dual energy decomposition CBCT technique for brain radiotherapy applications
Author(s): Shailaja Sajja; Masoud Hashemi; Christopher Huynh; James G. Mainprize; Markus Eriksson; Young Lee; Hakan Nordstrom; Anula Nico; Fang-Fang Yin; Arjun Sahgal; Mark Ruschin
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

The purpose of the present study was to develop and evaluate a practical dual-energy imaging approach for enhancing on-board cone-beam CT (CBCT) image quality for brain radiotherapy applications. The proposed primary technique involves a projection domain calibration procedure. In-house fabricated aluminum and acrylic step wedges were stacked and oriented orthogonally to each other to produce 72 unique combinations of two-material path lengths, i.e. 8 acrylic steps × 9 aluminum steps. High (120 kV) and low (70 kV) energy projections were acquired of the step wedges and a 3rd order polynomial fit was used to map the log-normalized projection intensities to the known acrylic and aluminum thicknesses. The subsequent model was tested on two phantoms: 1) in-house DE phantom with PMMA background and calcium inserts of different concentrations (5 mg/mL, 200 mg/mL and 400 mg/mL) and 2) a RANDO head phantom. The decomposed projections were reconstructed separately as aluminum-only and acrylic-only reconstructions. In addition, virtual monochromatic projections (VMPs) obtained by combining the aluminum-only and acrylic-only projections were reconstructed at different keVs. A quantitative improvement was observed in the SDNR (signal difference to noise ratios) of the calcium inserts using Aluminum-reconstructions and synthesized VMPs (40 to 100 keV) compared to the single energy reconstructions. A reduction in beam hardening was observed as well. In addition, a qualitative improvement in soft-tissue visualization was observed with the RANDO phantom reconstructions. The findings indicate the potential of dual energy CBCT images: material specific images as well as VMPs for improved CBCT-based image guidance. The present approach can readily be applied on existing commercial systems and a feasibility study on patients is a worthwhile investigation.

Paper Details

Date Published: 1 March 2019
PDF: 14 pages
Proc. SPIE 10948, Medical Imaging 2019: Physics of Medical Imaging, 109483A (1 March 2019); doi: 10.1117/12.2510430
Show Author Affiliations
Shailaja Sajja, Sunnybrook Research Institute (Canada)
Masoud Hashemi, Sunnybrook Research Institute (Canada)
Christopher Huynh, Sunnybrook Research Institute (Canada)
James G. Mainprize, Sunnybrook Research Institute (Canada)
Markus Eriksson, Elekta Instrument (Sweden)
Young Lee, Univ. of Toronto (Canada)
Hakan Nordstrom, Elekta Instrument (Sweden)
Anula Nico, Univ. of Toronto (Canada)
Fang-Fang Yin, Duke Univ. (United States)
Arjun Sahgal, Univ. of Toronto (Canada)
Mark Ruschin, Univ. of Toronto (Canada)


Published in SPIE Proceedings Vol. 10948:
Medical Imaging 2019: Physics of Medical Imaging
Taly Gilat Schmidt; Guang-Hong Chen; Hilde Bosmans, Editor(s)

© SPIE. Terms of Use
Back to Top