Premium
SU‐D‐141‐03: Evaluation of Cone‐Beam CT Imaging Quality Using a Clinical Detective Quantum Efficiency Measure (cDQE)
Author(s) -
Samant S,
Lee S,
Gopal A
Publication year - 2013
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4814034
Subject(s) - detective quantum efficiency , imaging phantom , cone beam computed tomography , image quality , optics , collimated light , physics , medical imaging , optical transfer function , dosimetry , nuclear medicine , materials science , computer science , medicine , artificial intelligence , image (mathematics) , laser , computed tomography , radiology
Purpose: To quantitatively analyze the image quality of cone‐beam CT(CBCT) imaging using a robust image evaluation metric, a clinical analogue of detective quantum efficiency (DQE), to allow for image assessment under variable acquisition parameters. Methods: CBCT image quality assessment, especially between different systems, remains semi‐quantitative and lacks the fundamental metrics found in 2D x‐ray imaging. Clinical DQE(cDQE) is based on measuring modulation transfer function (MTF) and noise power spectrum(NPS) using a cylindrical acrylic phantom containing modules with line‐pairs of various spatial frequencies, and uniform section. Imaging was acquired using a linac based CBCT (Elekta XVI) and a commercial cylindrical imaging QA phantom (Phantom Laboratory Catphan 500). CBCT imaging was obtained with various acquisition settings for collimation size (small S20, medium M20 field‐of‐views), beam filter (with and without bow‐tie filter), and adjustable scatter‐to‐primary ratios(SPRs). Incident beam fluence was determined based on dose measurements, and fluence‐dose conversion ratio computed by Monte Carlo simulations. A standard vendor supplied reconstruction algorithm was used with various SPRs (0, 0.1, 0.2, 0.33). Results: Use of bow‐tie filter improved CBCT image quality, as measured by cDQE, by 42% compared to imaging with neutral filter. As expected, cDQE confirmed that small collimation produced superior imaging compared to medium collimation due to reduced scattering. Interestingly, CBCT images reconstructed under various SPRs (0, 0.1, 0.2, 0.33) showed no significant difference in cDQE values. The results of cDQE calculations were consistent with evaluations of the phantom imaging based on traditional contrast‐to‐noise ratio, spatial resolution visibility, and qualitative observer assessment. Conclusion: cDQE was shown to be an effective metric to assess CBCT imaging under variable acquisition parameters. The proposed approach can be extended to other imaging phantoms and CBCTs, and has significant potential as a QA tool in monitoring and optimizing the imaging performance of CBCT systems for clinical use.