Premium
Adaptive CaRdiac cOne BEAm computed Tomography (ACROBEAT): Developing the next generation of cardiac cone beam CT imaging
Author(s) -
Reynolds Tess,
Dillon Owen,
Prinable Joseph,
Whelan Brendan,
Keall Paul J.,
O’Brien Ricky T.
Publication year - 2021
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 - 10.1002/mp.14811
Subject(s) - image quality , cone beam computed tomography , contrast to noise ratio , imaging phantom , siemens , cardiac imaging , nuclear medicine , medicine , iterative reconstruction , artificial intelligence , computer science , radiology , computed tomography , physics , image (mathematics) , quantum mechanics
Purpose An important factor when considering the use of interventional cone beam computed tomography (CBCT) imaging during cardiac procedures is the trade‐off between imaging dose and image quality. Accordingly, Adaptive CaRdiac cOne BEAm computed Tomography (ACROBEAT) presents an alternative acquisition method, adapting the gantry velocity and projection rate of CBCT imaging systems in accordance with a patient’s electrocardiogram (ECG) signal in real‐time. The aim of this study was to experimentally investigate that ACROBEAT acquisitions deliver improved image quality compared to conventional cardiac CBCT imaging protocols with fewer projections acquired. Methods The Siemens ARTIS pheno (Siemens Healthcare, GmbH, Germany), a robotic CBCT C‐arm system, was used to compare ACROBEAT with a commercially available conventional cardiac imaging protocol that utilizes multisweep retrospective ECG‐gated acquisition. For ACROBEAT, real‐time control of the gantry position was enabled through the Siemens Test Automation Control system. ACROBEAT and conventional image acquisitions of the CIRS Dynamic Cardiac Phantom were performed, using five patient‐measured ECG traces. The traces had average heart rates of 56, 64, 76, 86, and 100 bpm. The total number of acquired projections was compared between the ACROBEAT and conventional acquisition methods. The image quality was assessed via the contrast‐to‐noise ratio (CNR), structural similarity index (SSIM), and root‐mean square error (RMSE). Results Compared to the conventional protocol, ACROBEAT reduced the total number of projections acquired by 90%. The visual image quality provided by the ACROBEAT acquisitions, across all traces, matched or improved compared to conventional acquisition and was independent of the patient’s heart rate. Across all traces, ACROBEAT averaged 1.4 times increase in the CNR, a 23% increase in the SSIM and a 29% decrease in the RMSE compared to conventional and was independent of the patient’s heart rate. Conclusion Adaptive patient imaging is feasible on a clinical robotic CBCT system, delivering higher quality images while reducing the number of projections acquired by 90% compared to conventional cardiac imaging protocols.