The Impact of Different Levels of Adaptive Iterative Dose Reduction 3D on Image Quality of 320-Row Coronary CT Angiography: A Clinical Trial

Abstract

Purpose: The aim of this study was the systematic image quality evaluation of coronary CT angiography (CTA), reconstructed with the 3 different levels of adaptive iterative dose reduction (AIDR 3D) and compared to filtered back projection (FBP) with quantum denoising software (QDS).

Methods: Standard-dose CTA raw data of 30 patients with mean radiation dose of 3.2 ± 2.6 mSv were reconstructed using AIDR 3D mild, standard, strong and compared to FBP/QDS. Objective image quality comparison (signal, noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), contour sharpness) was performed using 21 measurement points per patient, including measurements in each coronary artery from proximal to distal.

Results: Objective image quality parameters improved with increasing levels of AIDR 3D. Noise was lowest in AIDR 3D strong (p ≤ 0.001 at 20/21 measurement points; compared with FBP/QDS). Signal and contour sharpness analysis showed no significant difference between the reconstruction algorithms for most measurement points. Best coronary SNR and CNR were achieved with AIDR 3D strong. No loss of SNR or CNR in distal segments was seen with AIDR 3D as compared to FBP.

Conclusions: On standard-dose coronary CTA images, AIDR 3D strong showed higher objective image quality than FBP/QDS without reducing contour sharpness.

Trial registration: Clinicaltrials.gov NCT00967876.

Conflict of interest statement

Competing Interests: Prof. Dewey has received grant support from the Heisenberg Program of the DFG for a professorship (DE 1361/14-1), the FP7 Program of the European Commission for the randomized multicenter DISCHARGE trial (603266-2, HEALTH-2012.2.4.-2), the European Regional Development Fund (20072013 2/05, 20072013 2/48), the German Heart Foundation/German Foundation of Heart Research (F/23/08, F/27/10), the Joint Program from the German Research Foundation (DFG) and the German Federal Ministry of Education and Research (BMBF) for meta-analyses (01KG1013, 01KG1110, 01KG1110), GE Healthcare, Bracco, Guerbet, and Toshiba Medical Systems. The CARS-320 study has received grants from Bracco. Prof. Dewey has received lecture fees from Toshiba Medical Systems, Guerbet, Cardiac MR Academy Berlin, and Bayer (Schering-Berlex). Prof. Dewey is a consultant to Guerbet and one of the principal investigators of multi-center studies (CORE-64 and 320) on coronary CT angiography sponsored by Toshiba Medical Systems. He is also the editor of Coronary CT Angiography and Cardiac CT, both published by Springer, and offers hands-on workshops on cardiovascular imaging (www.ct-kurs.de). Prof. Dewey is an associate editor of Radiology and European Radiology. Dr. Jörg Blobel and Dr. Joanne Schuijf are full-time employees of Toshiba Medical Systems Europe. Institutional master research agreements exist with Siemens Medical Solutions, Philips Medical Systems, and Toshiba Medical Systems. The terms of these arrangements are managed by the legal department of Charité – Universitätsmedizin Berlin. The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article. The authors have full control control of all primary data. All funders (TMS, Guerbet, CMR Bayer, Siemens) provided support in form of salaries, but did not have any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the “author contributions`section”. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials. Two employees from Toshiba Medical System (JB and JS) gave their input into the manuscript and, due to that, are listed as authors.

Figures

Fig 1. CONSORT Flow Diagram.

Fig 2. Analysis of signal and noise.

Measurement in the vessel and the surrounding tissue at the example of right coronary artery (RCA) being reconstructed by using filtered back projection/ quantum denoising filtering software; axial slices; a/d: proximal RCA measurement point 5 mm distal of the beginning of the vessel; b/e: medial RCA measurement point 5 mm distal of the first branch; c/f: distal RCA measurement point 5 mm distal of the second branch; a-c: field of view of 180 mm; d-f: zooming of a-c; The ROIs in the vessel lumen were placed as large as possible without integrating calcified and non-calcified plaques, stents and also the vessel wall into the analysis (approximately 50% of the vessel lumen).

Fig 3. Analysis of contour sharpness.

The DICOM images were interpolated by Vitrea fx. The measurement is based on a straight line that is placed orthogonal to the vessels course connecting the epicardial tissue on both sides of the vessel at the example of RCA (a-d); e-h: gray values in HU along the line above the vessel for the analysis of contour sharpness; see S5 Table for statistical results; a/e: combined filtered back projection and quantum denoising filtering software; b/f: adaptive iterative dose reduction three-dimensional (AIDR 3D) mild; c/g: AIDR 3D standard; d/h: AIDR 3D strong; i: The analysis of contour sharpness is based on the difference between 25% and 75% of the maximal gray value (max.) and the maximal slope of gray values between two pixels in percent of the difference between minimal (min.) and maximal density.

Fig 4. Comparison of quantitative image quality parameters.

Signal (a) and noise (b) of the reconstructions FBP/QDS, AIDR 3D mild, AIDR 3D standard (STD) and AIDR 3D strong (STR) in the ascending aorta measurement point; c/e/g: signal-noise-ratio (SNR); d/f/h: contrast-noise-ratio (CNR); results in the proximal (c/d), mid (e/f) and distal measurement points (g/h); the same results were found for proximal, medial and distal coronary segments, see S2–S4 Tables for statistical results.

Fig 5. Bland Altman plots of SNR and CNR.

The percentage relative deviation of SNR and CNR of each measurement point from the mean relative deviation for that particular IR level from FBP/QDS is plotted against the corresponding FBP/QDS measurement. Distal segments (RCA3, LAD3 and LCX3) are shown as open marked dots. Mainly, the relative deviation of SNR and CNR in distal segments is not significantly different compared to that of proximal segments; a: AIDR mild; relative deviation is less than 2% of the mean relative deviation of all segments, b: AIDR standard; Relative deviation is less than 1.5% of the mean relative deviation of all segments, c: AIDR strong; Relative deviation is less than 4% of the mean relative deviation of all segments.

Fig 6. Comparison of contour sharpness.

No significant difference was found for the contour sharpness based on the difference between 25% and 75% of the maximal gray value (a) and the maximal slope of gray values between two pixels in percent of the difference between minimal and maximal density (b); comparison between FBP/QDS, AIDR 3D mild, AIDR 3D standard (STD) and AIDR 3D strong (STR).