Image quality improvement in submillisievert computed tomographic colonography using a fast 3-dimensional noise reduction method.

Abstract

The objective of this study was to evaluate the image quality in submillisievert computed tomographic colonography (CTC) images using a structure preserving diffusion denoising method. Image quality was compared before and after denoising in 31 patients. One hundred twenty-kilovolt, 30-mAs prone CTC scans were used as reference and compared with submillisievert 140-kV, 10-mAs supine scans. Two readers assessed 2-dimensional and endoluminal image quality. The image noise and the signal-to-noise ratio were measured. After denoising, image quality scores improved in both supine series and prone series (P < 0.0001), with the submillisievert denoised images being equal to or better than the native prone reference images. In both the supine images and the prone images, the noise was reduced by a factor of 2 and the signal-to-noise ratio was significantly higher (P < 0.001). The signal-to-noise ratio in the denoised submillisievert images was higher than those in the native prone images (P < 0.001). The structure preserving diffusion denoising method preserves the image quality in submillisievert CTC images compared with the native 30-mAs reference images.