Comparisons of image quality and radiation dose between iterative reconstruction and filtered back projection...

The purpose of this study is to compare image quality and radiation dose between iterative reconstruction and filtered back projection (FBP) reconstruction algorithms in 256-MDCT coronary angiography.

One hundred nine consecutive patients were scanned with an ECG-gated helical technique using a 256-MDCT scanner. The patients in group A were scanned using the conventional tube output (120 kVp and 1000 mAs), and images were reconstructed using FBP. The other three groups were scanned with the same peak kilovoltage but with successively reduced tube current-time product, as follows: group B, 600 mAs; group C, 500 mAs; and group D, 400 mAs. Images were reconstructed using different levels of a new iterative reconstruction technique (iDose(4), Philips Healthcare). Noise, contrast-to-noise ratio, effective radiation dose, and scores of sharpness, contrast, and acceptability (where 1 indicates the worst, and 4 indicates the best) were compared between the four groups.

Noise in group A (mean [± SD], 16 ± 2 HU) was significantly higher than that in groups B (12 ± 2 HU), C (14 ± 2 HU), and D (13 ± 2 HU; p < 0.001). The contrast-to-noise ratio in groups B, C, and D was significantly higher than that in group A (p = 0.03). There were no differences in contrast, sharpness, and acceptability of images between group A and groups B and C. Using the criterion of image quality (score, 4), the receiver operating characteristic curve of dose levels and image quality acceptability established a reduction of 50% of tube output (group C) as the optimum cutoff point (area under the curve, 0.70; 95% CI, 0.57-0.82). The effective dose of group C was 55% lower than that of group A (5.5 ± 0.6 vs 12.0 ± 1.5 mSv).

Iterative reconstruction can provide equivalent or improved coronary image quality on coronary CT angiography, compared to routine-dose FBP reconstruction, while enabling radiation dose reductions of 55%.