Dose reduction with iterative reconstruction: Optimization of CT protocols in clinical practice

AbstractObjectivesTo create an adaptable and global approach for optimizing MDCT protocols by evaluating the influence of acquisition parameters and Iterative Reconstruction (IR) on dose reduction and image quality.Materials and methodsMDCT acquisitions were performed on quality image phantom by varying kVp, mAs, and pitch for the same collimation. The raw data were reconstructed by FBP and Sinogram Affirmed Iterative Reconstruction (SAFIRE) with different reconstruction kernel and thickness. A total of 4032 combinations of parameters were obtained. Indices of quality image (image noise, NCT, CNR, SNR, NPS and MTF) were analyzed. We developed a software in order to facilitate the optimization between dose reduction and image quality. Its outcomes were verified on an adult anthropomorphic phantom.ResultsDose reduction resulted in the increase of image noise and the decrease of SNR and CNR. The use of IR improved these indices for the same dose without affecting NCT and MTF. The image validation was performed by the anthropomorphic phantom. The software proposed combinations of parameters to reduce doses while keeping indices of the image quality adequate. We observed a CTDIvol reduction between −44% and −83% as compared to the French diagnostic reference levels (DRL) for different anatomical localization.ConclusionThe software developed in this study may help radiologists in selecting adequate combinations of parameters that allows to obtain an appropriate image with dose reduction