Multidetector-Row CT Angiography Diagnostic Sensitivity in Evaluation of Renal Artery Stenosis

Abstract

The aim of this study was to assess the image quality and interobserver agreement of various multidetector-row computed tomographic angiography postprocessing techniques in the diagnosis of renal artery stenosis (RAS). We studied 36 patients (21 men and 15 women; mean age, 49 years) who underwent computed tomography angiography to assess renal arteries for suspected RAS. Patients were analyzed by using a multidetector-row computed tomography. Computer tomographic scans were obtained after intravenous bolus administration of 110 to 140 mL of nonionic contrast material using a 4- to 6-mL/s flow rate. We assessed every patient by using axial scans, multiplanar reconstruction (MPR), maximum intensity projection (MIP), and volume rendering (VR) techniques. For each patient and for each reconstruction method, the image quality of the main renal artery was scored as 0 for bad-quality, 1 for poor-quality, 2 for good-quality, and 3 for excellent-quality images. Two radiologists reviewed computed tomographic images independently. We calculated interobserver agreement and kappa value. We correlated the stenosis degree observed by the 2 readers with the type of reconstruction used. Overall number of renal arteries studied was 72, and we detected 24 RAS. Quality images obtained an overall (averaged between the 2 observers) value of 133 of 216, 163 of 216, and 145 of 216 for MPR, MIP, and VR, respectively. Our data underlined a statistical difference between MPR images and VR images (P < 0.001). Moreover, we noticed that the images classified as excellent were obtained from a vessel with 350 Hounsfield units or higher. Kappa value was good in MIP and VR methods evaluation but poor with the use of MPR. Reformatting techniques usually provided a high visual impact, and in our study, MIP and VR showed the best diagnostic interobserver agreement in quality and reproducibility of stenosis degree.