Evaluation of Image Quality of Multiplanar Reconstruction Images: Effect of Z-increment of Original Axial Images

Abstract

The aim of our study was to evaluate the image quality of multiplanar reconstruction images (MPRs) focusing on the effect of z-increment of original axial images using signal to noise ratio (SNR) measurement in in- plane and longitudinal directions. SNRs of MPRs were calculated using modulation transfer function (MTF) and noise power spectrum (NPS). We scanned a bead phantom with a diameter of 0.1 mm and a water phantom with a diameter of 250 mm for calculating MTF and NPS using a MDCT with 0.5 s per rotation, 1.0 pitch and 64 x 0.6 mm collimation, and 50 mm field of view. Axial images for generating MPRs were reconstructed with standard kernel (B40), and 1.00 mm slice width. Coronal images were generated from two datasets with 0.1 mm and 0.5 mm z-increments of axial images respectively. For measuring the SNRs, the MTFs and NPSs in in-plane and longitudinal directions of each dataset were calculated from coronal bead images and coronal uniform noise images, respectively. Differences of MTF, NPS, and SNR were compared in in-plane and longitudinal directions. The MTF of longitudinal direction of the dataset with 0.1 mm z-increment was higher than the dataset with 0.5 mm z-increment. 10% MTFs of longitudinal direction with 0.1 mm and 0.5 mm z-increments were 0.75 cycles/mm and 0.68 cycles/mm, respectively. Conversely, the NPS of longitudinal direction of the dataset with a 0.1 mm z-increment was lower than the dataset with a 0.5 mm z-increment. As a consequence, the SNRs of longitudinal direction had relatively no difference between the datasets. In in-plane direction, MTFs, NPSs and SNRs had no differences between the datasets. A tradeoff relationship was indicated between spatial resolution and noise characteristic in the longitudinal direction due to the effects of different z-increment of original axial images used in generating MPRs. MPR using 0.5 mm z-increment of axial images had comparable SNR to MPR using 0.1 mm z-increment of axial images in our experimental condition. CLINICAL RELEVANCE/APPLICATION: Using 0.5 mm z-increment of original axial images for generating MPRs is effective for reducing the data volume, reconstruction time and transfer time without reducing image quality.