Spiral CT of renal artery stenosis: Comparison of three-dimensional rendering techniques: Rubin GD, Dake MD, Napel S, Jeffrey RB, McDonnell CH, Sommer FG, Wexler L, Williams DM. Radiology 1994;190:181-9

Abstract

Rubin GD, Dake MD, Napel S, Jeffrey RB, McDonnell CH, Sommer FG, Wexler L, Williams DM. Radiology 1994;190:181-9. Background: Conventional arteriography has long been the standard of reference in the diagnosis of renal artery stenosis (RAS). The search for a less invasive technique has led to the evaluation of numerous radiographic and ultrasonographic techniques. With the recent development of spiral computed tomography (CT), scans can be obtained entirely during the arterial phase of contrast injection, enabling reformation of data to create angiographic projections through any plane. The purpose of this study was to determine the accuracy of CT angiography in depicting hemodynamically limiting RAS as compared retrospectively with conventional arteriography. Methods and Results: CT angiography was performed in 31 patients undergoing conventional renal arteriography. Intravenous contrast volume and rate ranged between 90 ml at 3 ml/sec and 150 ml at 5 ml/sec, with the rate determined to maintain the injection duration constant at 30 seconds. The volume of contrast material was scaled to body weight. CT angiographic data were reconstructed with shaded surface display (SSD) and maximum intensity projection (MIP), two three-dimensional rendering techniques that are currently available on most commercial CT and magnetic resonance systems. Stenosis was graded with a four-point scale: grade 0, 0% to 49% stenosis; grade 1, 50% to 69% stenosis; grade 2, 70% to 99% stenosis; and grade 3, occlusion. The presence of mural calcification, poststenotic dilation, and nephrographic abnormalities was also noted. CT angiography depicted all main (n = 62) and accessory (n = 11) renal arteries that were seen on conventional arteriography. MIP CT angiography was 92% sensitive and 83% specific for the detection of grade 2 to 3 stenoses (=70% stenosis). SSD CT angiography was 59% sensitive and 82% specific for the detection of grade 2 to 3 stenoses. The accuracy of stenosis grading was 80% with MIP and 55% with SSD CT angiography. Poststenotic dilation and the presence of an abnormal nephrography result were 85% and 98% specific, respectively. Summary: The results of this study indicate that spiral CT angiography shows promise in the diagnosis of RAS. The accuracy of CT angiography varied with the three-dimensional rendering technique used and was higher with MIP than with SSD CT angiography. Comments: Spiral CT angiography not only provides detailed and dramatic three-dimensional images but also offers the advantage of being quick and considerably less invasive than conventional arteriography. Current limitations of CT angiography include the inability to obtain satisfactory images in patients weighing more than 125 kg, and the large volume (120 to 150 ml) of contrast material generally required to create diagnostic images. Nevertheless, in this preliminary evaluation, CT angiography was highly sensitive and specific for major RAS, and accuracy can be expected to increase further as spiral CT technology evolves. As this evolution occurs, optimal data collection and image rendering will require continued evaluation. Charles S. O'Mara, MD University Medical Center Jackson, Miss.