Reliability of renal length measurements made with ultrasound compared with measurements from helical CT multiplanar reformat images.

Abstract

The purpose of this article is to determine the reliability of sonographic renal length measurements compared with measurements obtained from helical CT multiplanar reformat images and compared with standard renal growth curves. A retrospective review was performed of 76 subjects who underwent both renal ultrasound and abdominal CT within 2 weeks of one another. Renal lengths were measured using oblique coronal reformat images of helically acquired CT data by two observers on two occasions. Intraobserver and interobserver error for these measurements were calculated. Ultrasound renal length measurements were compared with CT measurements. Measurement variation was compared with standard renal growth curves. The mean (± SD) of the absolute value of interobserver error of CT measurements was 0.9 ± 0.8 mm. Compared with CT, individual ultrasound measurements underestimated renal length by 1.5 ± 5.6 mm on average, with a 95% CI of -12.5 to 9.5 mm. When the maximum of three ultrasound renal length measurements was used, the SD was 4.7 mm, with a 95% CI of -8.2 to 10.1 mm of the reported renal length. This corresponds to greater or less than 3.3 years of normal renal growth. Lack of renal growth can be asserted only when renal length falls below the growth curve, taking into account the corresponding measurement error limits, which we found to be greater or less than 9.3 mm. If the follow-up measurement falls within these limits, one should not infer lack of appropriate renal growth, even if the renal length measurement decreases or remains unchanged for up to 3 years.