Coronary artery plaques: cardiac CT with model-based and adaptive-statistical iterative reconstruction technique

Objectives: To compare image quality of coronary artery plaque visualization at CT angiography with images reconstructed with filtered back projection (FBP), adaptive statistical iterative reconstruction (ASIR), and model based iterative reconstruction (MBIR) techniques.

Methods: The coronary arteries of three ex vivo human hearts were imaged by CT and reconstructed with FBP, ASIR and MBIR. Coronary cross-sectional images were co-registered between the different reconstruction techniques and assessed for qualitative and quantitative image quality parameters. Readers were blinded to the reconstruction algorithm.

Results: A total of 375 triplets of coronary cross-sectional images were co-registered. Using MBIR, 26% of the images were rated as having excellent overall image quality, which was significantly better as compared to ASIR and FBP (4% and 13%, respectively, all p<0.001). Qualitative assessment of image noise demonstrated a noise reduction by using ASIR as compared to FBP (p<0.01) and further noise reduction by using MBIR (p<0.001). The contrast-to-noise-ratio (CNR) using MBIR was better as compared to ASIR and FBP (44±19, 29±15, 26±9, respectively; all p<0.001).

Conclusions: Using MBIR improved image quality, reduced image noise and increased CNR as compared to the other available reconstruction techniques. This may further improve the visualization of coronary artery plaque and allow radiation reduction.