Simulated 50 % radiation dose reduction in coronary CT angiography using adaptive iterative dose reduction in three-dimensions (AIDR3D)

Abstract

To compare the image quality of coronary CT angiography (CTA) studies between standard filtered back projection (FBP) and adaptive iterative dose reduction in three-dimensions (AIDR3D) reconstruction using CT noise additional software to simulate reduced radiation exposure. Images from 93 consecutive clinical coronary CTA studies were processed utilizing standard FBP, FBP with 50% simulated dose reduction (FBP50%), and AIDR3D with simulated 50% dose reduction (AIDR50%). Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were measured within 5 regions-of-interest, and image quality for each reconstruction strategy was assessed by two independent readers using a 4-point scale. Compared to FBP, the SNR measured from the AIDR50% images was similar or higher (airway: 38.3 ± 12.7 vs. 38.5 ± 14.5, p = 0.81, fat: 5.5 ± 1.9 vs. 5.4 ± 2.0, p = 0.20, muscle: 3.2 ± 1.2 vs. 3.1 ± 1.3, p = 0.38, aorta: 22.6 ± 9.4 vs. 20.2 ± 9.7, p < 0.0001, liver: 2.7 ± 1.0 vs. 2.3 ± 1.1, p < 0.0001), while the SNR of the FBP50 % images were all lower (p values < 0.0001). The CNR measured from AIDR50% images was also higher than that from the FBP images for the aorta relative to muscle (20.5 ± 9.0 vs. 18.3 ± 9.2, p < 0.0001). The interobserver agreement in the image quality score was excellent (κ = 0.82). The quality score was significantly higher for the AIDR50% images compared to the FBP images (3.6 ± 0.6 vs. 3.3 ± 0.7, p = 0.004). Simulated radiation dose reduction applied to clinical coronary CTA images suggests that a 50% reduction in radiation dose can be achieved with adaptive iterative dose reduction software with image quality that is at least comparable to images acquired at standard radiation exposure and reconstructed with filtered back projection.

Fig. 1

Hounsfield Unit measurements in 7 regions-of-interest (ROI) placed in the phantom. The white round areas surrounding the numbers indicate each ROI

Fig. 2

Qualitative image quality score from the filtered back projection (FBP) group, filtered back projection with 50 % dose reduction (FBP50 %) and adaptive iterative dose reduction with 50 % dose reduction (AIDR50 %). -Error bars represent the standard deviations. p values from Wilcoxon signed-rank test

Fig. 3

Representative axial (a–c) and corresponding curved multiplanar reformatted images (d–f) of the left anterior descending artery from three reconstruction methods performed on one subject (window width/level = 1,200/200). a and d were reconstructed using conventional filtered back projection (FBP); image noise increases in b and e that were reconstructed with FBP after a simulated 50 % dose reduction (FBP50 %). c and f were reconstructed with AIDR3D applied to the raw data after the application of the 50 % simulated noise reduction (AIDR50 %)

Fig. 4

Representative axial (a–c) and corresponding curved multiplanar reformatted images (d–f) of the right coronary artery with coronary artery disease (arrow) from three reconstruction methods performed on one subject (window width/level = 1,200/200). a and d were reconstructed using conventional filtered back projection (FBP); image noise increases in b and e that were reconstructed with FBP after a simulated 50 % dose reduction (FBP50 %). c and f were reconstructed with AIDR3D applied to the raw data after the application of the 50 % simulated noise reduction (AIDR50 %)