Coronary CT angiography: image quality, diagnostic accuracy, and potential for radiation dose reduction using a novel iterative image reconstruction technique—comparison with traditional filtered back projection

Abstract

To compare image noise, image quality and diagnostic accuracy of coronary CT angiography (cCTA) using a novel iterative reconstruction algorithm versus traditional filtered back projection (FBP) and to estimate the potential for radiation dose savings. Sixty five consecutive patients (48 men; 59.3 ± 7.7 years) prospectively underwent cCTA and coronary catheter angiography (CCA). Full radiation dose data, using all projections, were reconstructed with FBP. To simulate image acquisition at half the radiation dose, 50% of the projections were discarded from the raw data. The resulting half-dose data were reconstructed with sinogram-affirmed iterative reconstruction (SAFIRE). Full-dose FBP and half-dose iterative reconstructions were compared with regard to image noise and image quality, and their respective accuracy for stenosis detection was compared against CCA. Compared with full-dose FBP, half-dose iterative reconstructions showed significantly (p = 0.001 - p = 0.025) lower image noise and slightly higher image quality. Iterative reconstruction improved the accuracy of stenosis detection compared with FBP (per-patient: accuracy 96.9% vs. 93.8%, sensitivity 100% vs. 100%, specificity 94.6% vs. 89.2%, NPV 100% vs. 100%, PPV 93.3% vs. 87.5%). Iterative reconstruction significantly reduces image noise without loss of diagnostic information and holds the potential for substantial radiation dose reduction from cCTA.