Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print June 13, 2005, 10.1148/radiol.2361040158
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2361040158v1
236/1/276    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Riccabona, M.
Right arrow Articles by Mache, C. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Riccabona, M.
Right arrow Articles by Mache, C. J.
(Radiology 2005;236:276-283.)
© RSNA, 2005

Pediatric Imaging

Hydronephrotic Kidney: Pediatric Three-dimensional US for Relative Renal Size Assessment—Initial Experience1

Michael Riccabona, MD, Gerald A. Fritz, MD, Helmut Schöllnast, MD, Thomas Schwarz, MD, Michael J. Deutschmann, MD and Christoph J. Mache, MD

1 From the Department of Radiology, Divisions of Pediatric Radiology (M.R., M.J.D.), General Radiology (G.A.F., H.S.), and Nuclear Medicine (T.S.), and Department of Pediatrics (C.J.M.), University Hospital Graz, Auenbruggerplatz 9, A-8036 Graz, Austria. Received January 28, 2004; revision requested April 6; final revision received September 6; accepted September 29. Address correspondence to M.R. (e-mail: michael.riccabona{at}meduni-graz.at).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To prospectively evaluate accuracy of three-dimensional (3D) ultrasonography (US) for assessment of relative renal size in infants and children with hydronephrosis.

MATERIALS AND METHODS: Informed consent was obtained from parents and also from children who were older than 8 years. Study was approved by ethics committee. Two-dimensional (2D) US, 3D US, and scintigraphy were performed in 40 patients with hydronephrosis (age range, neonate to 16 years; seven girls, 33 boys) without acute renal disease. Twenty patients also underwent magnetic resonance (MR) urography. US and MR urography were performed by one experienced pediatric radiologist; 3D US and MR urographic volume calculations were performed by specifically trained radiologists. Three-dimensional US was performed with integrated 3D volume probes or external system based on electromagnetic positioning devices. At 2D US, kidney volume was calculated with application of ellipsoid equation. At MR urography and 3D US, real renal parenchymal volume was calculated with subtraction of dilated collecting system. Split renal function was assessed with static renal scintigraphy. Three-dimensional US results were graded with respect to image quality and compared with results of 2D US, scintigraphy, and MR urography by using mean difference percentage and standard deviation of the difference. All investigations were performed with blinding. Inter- and intraobserver variability were calculated with coefficient of variation.

RESULTS: In 76 of 80 kidneys, 3D US image of diagnostic quality was obtained. Three-dimensional US volume measurements compared well with MR urographic measurements (mean difference, –2.5% ± 7.8 [standard deviation] vs 25.8% ± 32.2 for 2D US) and with scintigraphically assessed split renal function (mean difference, 1.2% ± 9.2 vs 15.9% ± 43.8 for 2D US). Intra- and interobserver variability were ±6.4% and ±9.9%, respectively.

CONCLUSION: Initial experience with renal 3D US indicates that it is an accurate method for assessment of renal parenchymal volume and relative renal size, provided there is no acute renal disease.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Ultrasonography (US) has become the basic imaging modality in pediatric patients who are suspected of having urinary tract disease. Conventional two-dimensional (2D) US is used as a screening modality for evaluation in patients with acute disease such as urinary tract infection or acute obstruction, for evaluation and follow-up in patients with congenital urinary tract malformations, and for assessment in patients with chronic urinary tract disease (1,2). Two-dimensional US is routinely used for diagnosis and follow-up in infants and children with hydronephrosis, such as that which results from ureteropelvic junction obstruction, dilating (ie, high-grade) vesicoureteral reflux, primary obstructive megaureter, or simple nonobstructive dilatation such as hydrocalycosis. Assessment of renal parenchymal volume in particular, which is performed with repeated static renal scintigraphy, is one essential requisite for therapeutic decision making. Relative renal size and split renal function are critical factors not only initially but also during follow-up. Renal (parenchymal) growth is an essential aspect of patient treatment (2–6). It has been shown that US measurement of relative renal volume correlates with scintigraphically assessed split renal function in normally shaped kidneys without dilatation of the collecting system and with the absence of any acute urinary tract disease, such as an infection (7–9). Volume calculations that are based on geometric estimates at 2D US, however, may be inaccurate in organs with a nongeometric shape, particularly in patients with irregularly shaped kidneys or with hydronephrosis (10). These limitations may create inaccuracies in comparisons of 2D US results with those of other modalities or with those at follow-up and may obscure subtle changes, such as a slow decrease in renal parenchymal volume, in an evolving disease.

Three-dimensional (3D) US has become an established imaging tool in many specialties. In clinical and research settings, investigators have shown that 3D US is beneficial in imaging in the fields of obstetrics and gynecology, cardiovascular disease, oncology, neurology, and neurosurgery, as well as in interventional radiology (11–16). In addition, 3D US reportedly offers a marked improvement in organ volume assessment, with superior standardization and lower inter- and intraobserver variabilities (17–27). Yet, few reports have been published about the use of pediatric 3D US, and those that have been published mainly deal with the neonatal brain (21,27–31). To our knowledge, only a few reports have been published to date about the use of 3D US for imaging the adult or pediatric urinary tract to demonstrate the overall feasibility and accuracy of 3D US volume assessment in pediatric and adult kidneys (32–35). Thus, the aim of this study was to prospectively evaluate the accuracy of 3D US in the assessment of relative renal size in infants and children with hydronephrosis.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
GE Medical Systems Kretztechnik, Zipf, Austria, provided the Voluson 3D US device. The authors had complete control of the data and information for publication; no authors were employees of the company or had any financial interest with the funding source.

Study Population
The study population consisted of 40 patients (33 boys and 7 girls; age range, newborn to 16 years; mean age, 1.75 years; median age, 0.26 year) who were referred to the Division of Pediatric Radiology at the Department of Radiology, University Hospital, Graz, Austria, for evaluation of hydronephrosis. They all underwent 2D US of the urinary tract; immediately afterward, they underwent 3D US of the kidneys after informed consent had been obtained from the parents and also from any children who were older than 8 years. The study had been approved by our ethics committee. In all patients, renal scintigraphy was performed within 14 days (before or after US). Twenty patients additionally underwent MR urography within 14 days after 2D US for standard clinical indications (eg, a preoperative imaging finding, a complex abnormality, or a suspected crossing vessel). Sedation was used in infants and young children who were 1–6 years old for MR urography and in part for scintigraphy; no sedation was administered for 2D or 3D US. Final diagnoses were established by using all available images (voiding cystourethrogram or MR urographic, US, static and dynamic renal scintigraphic images) read in consensus, with surgical confirmation in those patients who eventually needed surgery.

US and Evaluation
Two-dimensional US was performed by an experienced pediatric radiologist (M.R., with 20 years of experience with pediatric US) by using a variety of age-adapted multifrequency transducers (1–15 MHz) of a US system (Sequoia 512; Acuson/Siemens, Mountain View, Calif). Each kidney was measured in two planes and documented in the picture archiving and communication system. At 2D US, volume calculations were performed after each study on the basis of the ellipsoid equation (V = {pi}/6 x l x dx h, where V is volume, d is diameter, and h is height), as reported by Dinkel at al (36). Grading of hydronephrosis was performed by using a scale adapted to pediatric uses, based on the classification of the Society for Fetal Urology, as follows: grade 0, normal kidney without dilatation; grade 1, renal pelvis that is just visible; grade 2, some visible normal-shaped calyces; grade 3, marked dilatation of renal calyces and pelvis, with loss of normal calyceal shape (ie, rounded papillary impression, with flattening of the fornices but without parenchymal narrowing); and grade 4, additional parenchymal narrowing caused by gross dilatation of the entire collecting system (37).

Three-dimensional US was performed by the same investigator immediately after initial 2D US with one of two systems as available.

The first system (hereafter referred to as system 1) included freehand scanning with an electromagnetic positioning device attached to various multifrequency transducers (1–15 MHz) of the US imaging device, with an external personal computer–based 3D US system (EchoTech, GE Medical Systems, Munich, Germany).

The second system (hereafter referred to as system 2) included mechanically driven multifrequency transducers (3–8 MHz), such as the 3D volume probe (Voluson 730; GE Medical Systems Kretz-technik). The transducers had a small integrated motor that was attached to the scanner head and automatically performed a sweep over a predefined area.

At least two different acquisitions in arbitrary or orthogonal acquisition planes (axial and longitudinal section in the kidney) were performed. Time for the 3D US acquisition for both kidneys was 5 minutes on average, with slightly longer duration of 3D US for system 1 (mean, 6.3 minutes; range, 4–15 minutes) than for system 2 (mean, 4.1 minutes; range, 2–9 minutes). Three-dimensional US images were classified on the basis of quality into four groups by two investigators in consensus (M.R., 20 years of experience with pediatric US, and G.A.F., 2 years of experience with US). The classification was as follows: grade 1, excellent (the entire kidney was included in the 3D US data set, with well-defined border delineation and no artifacts); grade 2, good (the entire kidney was included in the 3D US data set, with well defined border delineation and only minor artifacts); grade 3, diagnostic (the entire kidney was included in the 3D US data set, with partially degraded border delineation and/or more minor artifacts, such as minor motion artifacts that did not impair border delineation or distort any structures); and grade 4, nondiagnostic (substantial parts of the kidney were not included in the 3D US data set or severe artifacts or unclear border delineation over a substantial part of the kidney was present). Kidneys with nondiagnostic 3D US quality (group 4) were not included in the statistical analysis. At 3D US, volume calculations were performed after the patient had left the imaging suite by using the integrated software from the two systems. For both systems, the reconstructed kidney was positioned in a standardized upright and enlarged view. Review was conducted by scrolling up and down the view of the kidney, and measurements and outlining were conducted by rotating the view of the kidney around its longitudinal axis.

The renal parenchymal volume assessment consisted of two steps: calculation of the entire kidney volume and subtraction of the dilated collecting system in patients with hydronephrosis. The volume assessment was based on a semiautomated, interactive, and adaptable delineation of the outer kidney contour. A rotation-symmetric algorithm was used for calculation of the entire kidney volume with both systems, with a slightly different user interface but the same underlying calculation.

The subtraction of the dilated system was slightly different in the two systems: In system 1, manual outlining throughout the 3D data set for planimetric calculation of the volume of the dilated collecting system was mandatory to then deduct this subvolume from the entire kidney volume (Fig 1). In system 2, not only the definition of the kidney border and surface geometry but also the outlining of the collecting system could be performed with the semiautomated approach. Calculation with this approach was based on rotation of the image plane about a fixed axis and a manual definition of the 2D contours in each plane. In this computerized calculation performed with the software of the system, the surface geometry was defined by 3D triangularization of the 2D contours, meaning that each point of the 2D contour in one plane is connected through a triangle mesh to corresponding points in both neighboring planes. The procedure started with outlining first the outer kidney contour and then manually tracing the dilated collecting system throughout the 3D data set. The subvolume of the collecting system thus was already delineated and subtracted by means of the manual tracing. In addition, system 2 also offered the possibility for manual outlining of the dilated collecting system for calculation of this subvolume separately and then for subtracting it from the entire kidney volume (Fig 2). If the kidney had an irregular outer border or the collecting system delineation was inaccurate, the planimetric approach with manual outlining of the outer contour and the dilated system was applied also for calculation of the entire kidney volume with both systems, and the results of the semiautomated and the manual approach were compared to assess potential variations. The outlining and the delineation of inner and outer contours were performed by one of the radiologists who performed the volume calculations (G.A.F., H.S., M.J.D.). All 2D and 3D US evaluations were performed with blinding to the results of other investigations. In addition, all 3D US volume measurements were performed twice by two different blinded radiologists (G.A.F., H.S.) for assessment of interobserver variation.



View larger version (73K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. Images obtained at 3D US depict renal parenchymal volume calculation in a hydronephrotic kidney by using system 1. (a) Images show volume assessment of entire kidney. Left: Longitudinal sections through the kidney show delineation of the outer contour in serial stages by using a semiautomated rotation-symmetric approach. Outlined areas signify the consecutively included area. Right: Two sectional views of the defined kidney that resulted from this procedure. (b) Longitudinal section obtained through the kidney shows delineation of dilated collecting system. Left: View shows manual outlining of the outer kidney contour performed with planimetric approach. Middle: Inverted maximum signal-intensity-weighted rendered 3D view that resulted from interactive definition of the subvolumes. Right: Rendered 3D box view in red shows subtraction from the entire volume, which resulted in the real renal parenchymal volume.

 


View larger version (31K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. Images obtained at 3D US depict renal parenchymal volume calculation in a hydronephrotic kidney by using system 1. (a) Images show volume assessment of entire kidney. Left: Longitudinal sections through the kidney show delineation of the outer contour in serial stages by using a semiautomated rotation-symmetric approach. Outlined areas signify the consecutively included area. Right: Two sectional views of the defined kidney that resulted from this procedure. (b) Longitudinal section obtained through the kidney shows delineation of dilated collecting system. Left: View shows manual outlining of the outer kidney contour performed with planimetric approach. Middle: Inverted maximum signal-intensity-weighted rendered 3D view that resulted from interactive definition of the subvolumes. Right: Rendered 3D box view in red shows subtraction from the entire volume, which resulted in the real renal parenchymal volume.

 


View larger version (85K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a. Images obtained at 3D US depict renal parenchymal volume calculation in a hydronephrotic kidney by using system 2. (a) Multiplanar view depicts definition of outer contour for calculation of the entire kidney volume. Left: Upper and lower images are longitudinal orthogonal sections. Right: Upper image is a purely reconstructed transverse section created by using the semiautomated rotation-symmetric approach. Lower image is rendered view (in red) that shows cropped volume. (b) Multiplanar display shows delineation of the dilated collecting system for subtraction from the entire kidney volume to obtain renal parenchymal volume. Left: Upper and lower images are longitudinal orthogonal sections. Right: Upper image is transverse section. Lower image is rendered view (in red) that depicts manually delineated subvolume of the dilated calyces and renal pelvis.

 


View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b. Images obtained at 3D US depict renal parenchymal volume calculation in a hydronephrotic kidney by using system 2. (a) Multiplanar view depicts definition of outer contour for calculation of the entire kidney volume. Left: Upper and lower images are longitudinal orthogonal sections. Right: Upper image is a purely reconstructed transverse section created by using the semiautomated rotation-symmetric approach. Lower image is rendered view (in red) that shows cropped volume. (b) Multiplanar display shows delineation of the dilated collecting system for subtraction from the entire kidney volume to obtain renal parenchymal volume. Left: Upper and lower images are longitudinal orthogonal sections. Right: Upper image is transverse section. Lower image is rendered view (in red) that depicts manually delineated subvolume of the dilated calyces and renal pelvis.

 
MR Urography, Scintigraphy, and Evaluation
MR urography was performed on the basis of clinical indications, such as a preoperative imaging finding, a complex abnormality, or a suspected crossing vessel. Therefore, only 20 patients underwent MR urography, which was performed with a 1.5-T system (Magnetom; Siemens Medical Systems, Erlangen, Germany) and age-adapted coils for transverse and coronal acquisitions. The other parameters were as follows: section thickness, 3–5 mm; matrix, 512 x 256. The field of view was varied according to the age of the patient and the size of the patient, the latter of which was defined primarily by height and weight (in regard to selection of the MR imaging coil and the contrast medium dose, respectively). T2-weighted sequences (half-Fourier rapid acquisition with relaxation enhancement [repetition time msec/echo time msec, 1000/396]), including respiratory-gated (diaphragmatic tracking) true fast imaging with steady-state precession (5/2.5), T1-weighted sequences (619/15), and 3D gradient-echo sequences (4.8/1.4) for dynamic assessment were performed before and after intravenous administration of a gadolinium-based contrast agent (0.1 mL/kg) and intravenous administration of furosemide (0.5 mg/kg; maximum dose, 20 mg). The latter sequence was also used for MR angiography (eg, for a suspected crossing vessel). Volume calculations were performed with integrated region-growing software (MagicView110; Siemens Medical Systems) at the workstation by using the nephrographic phase of the contrast material–enhanced 3D gradient-echo T1-weighted sequence by one radiologist (G.A.F.). Reading of MR urographic images was performed by an experienced pediatric radiologist (M.R., who had 5 years of experience working with pediatric MR urography), who was blinded to the results of all other imaging tests.

Static renal scintigraphy was performed in all patients within 14 days of 3D US. Acquisition was started 2–3 hours after intravenous administration of 0.1 mCu/kg (3.7 MBq/kg) technetium 99m (99mTc) dimercaptosuccinic acid to the physiologically hydrated patients. All images were acquired by using a dual-head large-field-of-view camera (e.cam; Siemens Medical Systems) with low-energy general-purpose collimators (zoom factor, 1.7 in children older than 1 year and 2.0 in children younger than 1 year of age), with 500 000 counts each in posterior, left posterior oblique, and right posterior oblique projections and loaded in a 256 x 256 matrix. Computerized split renal function assessment was based on the automated calculation of relative renal uptake (geometric mean of anterior and posterior acquisition) by using a semiautomatic definition of the region of interest for the kidneys and for background subtraction. Scintigraphy with 99mTc–dimercaptosuccinic acid and relative renal size and function were assessed and reported by a senior nuclear medicine specialist (T.S.), who was blinded to the results of the other investigations.

Statistical Analysis
Age and sex differences were evaluated by using the Wilcoxon-Mann-Whitney test. Relative renal volume expressed as a percentage and assessed with 2D and 3D US were compared with scintigraphically assessed split renal function by using the mean difference percentage and the standard deviation of the difference. In those patients in whom assessment was performed with MR urography, the absolute renal volume (in milliliters) determined at MR urography was compared with that determined at 2D US and at 3D US; the mean difference percentage and the standard deviation of the difference were calculated. Whenever possible, absolute numbers were used for comparison (MR urography vs 2D US and 3D US) and absolute (difference) percentages were used to account for dependency and clustering in the comparison of 2D US, 3D US, MR urography, and scintigraphy.

In addition, the differences in the results determined with systems 1 and 2 versus those determined with scintigraphy and MR urography were compared by using the numbers derived from previously mentioned calculations, and the results of the two 3D US calculation approaches (manual outlining or semiautomated volume calculation) were compared too. The null hypothesis that the difference between the measurements was zero was tested with two-tailed one-sample t tests, with a type I error probability of 5%. Furthermore, after 4 weeks, intra- and interobserver variabilities were evaluated with reassessment of the 3D US data sets, from which patient identifying information was removed. The standard deviations of the differences between volume measurements were divided by the mean of all volume measurements, yielding a coefficient of variation of differences. The coefficient of variation of single volume measurements was derived from this value by dividing it through the square root of two.

Software (SPSS 12.0 for Windows; SPSS, Chicago, Ill) was used for statistical calculations. A P value of less than .05 was considered to indicate a significant difference.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Eighty kidneys in 40 patients were evaluated. The age and sex differences were not statistically significant; furthermore, they constitute the typical sex distribution for this condition (median age for both sexes, 0.26 year; median age for male patients, 0.23 year; median age for female patients, 0.34 year [P = .67]). Forty-six renal units had a higher grade of hydronephrosis (grade 3 or 4, according to the Society for Fetal Urology classification). Of those, 22 kidneys with hydronephrosis of grade 3 or 4 were observed in the 20 patients who underwent MR urography in addition to 2D and 3D US and scintigraphy. The final diagnoses were: 32 cases of ureteropelvic junction obstruction in 28 patients, six cases of unilateral dilating high-grade vesicoureteral reflux in six patients, and six cases of unilateral primary megaureter in six patients. In addition, assessment of contralateral kidneys yielded these results: one patient with hydrocalycosis and one patient with nonobstructive dilatation; contralateral kidneys in the other patients were normal. Seventeen patients were examined with system 1, and 23 were examined with system 2.

Comparisons
Three-dimensional US images of diagnostic quality or better (grade 1–3) could be obtained in 76 kidneys. A nondiagnostic 3D US (grade 4) image was obtained in four hydronephrotic kidneys in four infants; a 2D US image also was of suboptimal quality because of motion artifacts, and these images could not be included for further evaluation. At 2D US, calculations of relative kidney volume showed a poor correlation with scintigraphically assessed split renal function in all kidneys with hydronephrosis (mean difference, 19.3% ± 61.8 [standard deviation]; P = .04). For all kidneys, the correlation was shown with a mean difference of 15.9% ± 43.8. Relative kidney volume also compared badly with the absolute kidney volumes determined at MR urography, with a mean difference of 25.8% ± 32.2 (P < .001), except for the 34 contralateral kidneys without dilated collecting systems, with a mean difference of 11.8% ± 21.6 for that vs scintigraphy (n = 34, P = .004) and of 2.4% ± 16.6 for that vs MR urography (n = 18, P = .56).

Relative renal parenchymal volume calculations determined with planimetry-based 3D US in patients with hydronephrosis showed a mean difference of 2.8% ± 10.7 (P = .09) and good correlation (r = 0.92, P < .001) when they were compared with scintigraphic calculations and a mean error of difference of 1.4% ± 7.7 (P = .41) and good correlation (r = 0.98, P < .001) for the comparison with MR urographic calculations (Fig 3). The mean difference for values for normal kidneys versus scintigraphic calculations (percentage difference) was –0.9% ± 7.2 and for normal kidneys versus MR urographic calculations (absolute difference) was –7.4% ± 5.9 (P < .001). In all kidneys, these values yielded an overall difference of 1.2% ± 9.2 (P = .25) for comparison with scintigraphic calculations and –2.5% ± 7.8 (P = .05) for comparison with MR urographic calculations. The 26 higher-quality 3D US data (ie, grade 1 and 2 quality for scans of hydronephrotic kidneys) offered even more accurate results (one kidney with grade 1 and 25 kidneys with grade 2 quality), with a mean difference of 0.2% ± 5.6 (P = .86) compared with those at scintigraphy and a mean difference of 1.2% ± 7.4 (P = .53) compared with those at MR urography. There was a higher accuracy for the assessment of hydronephrotic kidneys with system 2 (mean difference, 2.4% ± 6.8 [P = .11], compared with that at scintigraphy, and a mean difference of 0.9% ± 6.6 [P = .64], compared with that at MR urography) than there was for assessment with system 1 (mean difference, 3.5% ± 15.8 [P = .38], compared with that at scintigraphy, and a mean difference of 4.5% ± 9.1 [P = .18], compared with that at MR urography). Volume measurements in hydronephrosis had a lower error with the planimetric approach for volume calculation (numbers as given previously) versus the semiautomated volume calculation (mean difference, 5.6% ± 15.5 [P = .02], compared with the calculation at scintigraphy, and 3.8% ± 9.4 [P = .08], compared with that at MR urography). A little more time was necessary for the volume calculation with the planimetric analysis (6.5 minutes per kidney; range, 5–12 minutes) than was necessary for that with the semiautomated approach (4.6 minutes per kidney; range, 3–8 minutes). When we excluded the infant who had gross hydronephrosis in whom 3D US image quality was grade 3 and in whom the largest error occurred, the mean 3D US percentage error compared with that at scintigraphy was 1.4% ± 5.9 (P = .27) with the planimetric approach for volume calculation and 3.4% ± 7.5 (P = .07) with the semiautomated volume calculation.



View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. Plots show correlation between assessment at 3D US and at other modalities. (a) Plot of relative renal parenchymal size (expressed as a percentage) at 3D US compared with relative split renal function (expressed as a percentage) determined with scintigraphy in all kidneys of the entire study population, with good correlation (r = 0.92). (b) Plot of measured renal parenchymal volume (expressed in milliliters) at 3D US compared with MR urographic assessment of renal parenchymal volume (expressed in milliliters) in 22 hydronephrotic kidneys, with good correlation (r = 0.98). Note that volumes determined with MR imaging are generally larger, partly caused by the effect of the contrast medium during the nephrographic phase, and may increase renal parenchymal volume.

 


View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. Plots show correlation between assessment at 3D US and at other modalities. (a) Plot of relative renal parenchymal size (expressed as a percentage) at 3D US compared with relative split renal function (expressed as a percentage) determined with scintigraphy in all kidneys of the entire study population, with good correlation (r = 0.92). (b) Plot of measured renal parenchymal volume (expressed in milliliters) at 3D US compared with MR urographic assessment of renal parenchymal volume (expressed in milliliters) in 22 hydronephrotic kidneys, with good correlation (r = 0.98). Note that volumes determined with MR imaging are generally larger, partly caused by the effect of the contrast medium during the nephrographic phase, and may increase renal parenchymal volume.

 
In summary, we found significant differences in volume calculations with 2D US versus MR urography and with 2D US versus scintigraphy in patients with hydronephrosis, in calculations with 2D US versus scintigraphy in subjects with normal kidneys, and in calculations with semiautomated 3D US versus scintigraphy in patients with hydronephrosis. All other calculations yielded no statistically significant differences. Intra- and interobserver variabilities for the difference of two 3D US volume measurements were ±6.4% and ±9.9%, yielding variabilities for single volume measurements with 3D US of ±4.5% and ±7.0%, respectively.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Relative renal parenchymal size and split renal function assessments are an essential aspect of initial diagnosis and follow-up in patients with marked hydronephrosis. At present, only static renal scintigraphy provides this information, but repeated investigations during follow-up pose a considerable radiation burden to the patient (38). Radiation protection is considered most important, particularly in children; therefore, a nonionizing imaging modality that can be used to reliably assess relative renal parenchymal size would be appreciated.

US is the basic imaging tool for diagnosis and follow-up in patients with hydronephrosis. It is noninvasive and nonionizing and can even be used at the patient's bedside. Our results confirm that 2D US volume estimates of relative renal parenchymal size, however, are unreliable in patients with hydronephrosis because of the dilated collecting system. The renal volumes that are based on 2D US ellipsoid calculations, particularly in unilateral and gross hydronephrosis, do not correspond with relative renal parenchymal size or function. In addition, one must consider that US is used to calculate renal volume, whereas the calculation with scintigraphy is based on the tracer uptake by functioning tubular cells (3,39). Thus, both methods are used to assess different structures: with US, milliliters of tissue is measured, and with scintigraphy, number (ie, tracer uptake) of tubular cells with unimpaired function is measured. Scintigraphy specifically is used to evaluate the mass of functioning renal tissue; it does not include scarred tissue that is included in a US size assessment; furthermore, it adequately represents poorer functioning areas by means of the reduced local uptake. Despite these differences and US restrictions, a good correlation of determination of relative renal parenchymal size with US and with scintigraphically assessed split renal function has been shown for nonhydronephrotic kidneys without acute renal disease (7,8). This correlation indicates that there might be a clinically useful application of 3D US in the assessment of renal parenchymal volume in infants and children with hydronephrosis, provided an accurate subtraction of the dilated collecting system is achievable.

Three-dimensional US reportedly offers an increased accuracy at improved standardization (15,17–28,32). Our results demonstrate that 3D US in patients with hydronephrosis is accurate and reliably helps in the assessment of renal parenchymal volume because it allows subtraction of the dilated collecting system and enables a reliable calculation of relative renal parenchymal size. This calculation, at present, is achieved by means of measurement of the entire kidney volume and subtraction of the manually outlined collecting system. The volume calculation procedure is time consuming; in the future, semiautomated and interactive visually controlled segmentation for computerized calculations that are similar to MR imaging and CT techniques will be available and will help to speed up measurement time. An approach in which threshold segmentation is used for delineation of the collecting system cannot at present provide constant, reproducible, and reliable results and, therefore, was not included for this study.

Besides the improved volume assessment 3D US offers, other benefits that support its use in pediatric US of the urinary tract (35) are as follows: Three-dimensional US requires less time for examination of the patient, although more time is necessary afterward for evaluation of the 3D US volume data. Three-dimensional US enables a multiplanar view that may show abnormalities in critical sections, a capability that is not achievable by using 2D US (19–31,35). Three-dimensional US offers the potential for creation of rendered views (eg, for comprehensive demonstration of the dilated collecting system in patients with hydronephrosis [35]). Three-dimensional US offers an improved standardization with better inter- and intraobserver variabilities. Three-dimensional US creates optimal data for comparison of data obtained during US follow-up and/or for comparison of results with those of other imaging modalities. Finally, 3D US may serve as an ideal teaching and counseling tool, as one can virtually "rescan" the area and examine any anatomic relationship or structure, without the patient's presence being required, just by scrolling through the 3D US data set.

Some potential errors, limitations, and intrinsic restrictions of this study and of 3D US in general, however, have to be acknowledged: The patient group was relatively small and preselected, and the differences among the patients consecutively did not reach a high significance level. Larger cohorts will have to be investigated in a multi-institutional approach not only to confirm 3D US accuracy and its reduced observer dependency but also to evaluate therapeutic effects, economic aspects, and long-term effectiveness. Three-dimensional US relies on the quality of the 2D US source images. Since a good 2D US acquisition is compulsory, some investigator dependency will remain (15,29). Structures and regions not seen with 2D US (eg, because of interfering structures, restricted acquisition size, or insufficient penetration) or areas not included for reconstruction cannot be visualized, and repeated 3D US acquisitions may be necessary (eg, for large abnormalities that are difficult to image in their entirety, particularly when the motor-driven 3D US transducers with a restricted acquisition range are used) (15,29,32,35).

A variety of artifacts that are present at 2D US may then be imported into the 3D US data set and distort the 3D US image by being amplified by the reconstruction algorithm throughout all planes. Furthermore, phenomena and artifacts specific to 3D US (derived from data acquisition, reconstruction, or rendering) have to be considered (40). Patient motion (eg, that caused by breathing or crying and, particularly, the use of the external positioning device with a longer acquisition time) distorts 3D US data, even at real-time 3D US (15,29,32,35,40). Repeated 3D US acquisitions may be necessary. In our experience, 3D US could not be performed at all in some uncooperative infants. In these infants, however, 2D US examinations also were difficult to perform and were of suboptimal quality; thus, these patients were not included in the study population. Electromagnetic positioning devices may produce misregistration when materials are present that distort the magnetic field; acquisitions, therefore, must always be checked for distortion or artifacts (15,29,35,40).

The field of view is restricted by acquisition size, sweep, and 2D US accessibility, as defined by the site of the targeted area, with some differences between the various pieces of equipment; their effect on clinical utility yet remains to be clarified. Furthermore, resolution at 3D US is worse than that at 2D US, particularly in purely reconstructed planes (15,29,35). This difference in resolution may cause inaccuracies, such as those that may occur in the measurement of a very small residual parenchymal rim, as observed in a patient with a marked hydronephrosis. In that patient, consecutively, the largest errors in measurement in the entire study population were exhibited, and, thus, the accuracy of the data was decreased. Without this patient, at 3D US, error values in the patients with hydronephrosis would have been even better for both approaches and both systems: 1.4% ± 5.9 with the planimetric approach and 3.4% ± 7.5 with the semiautomated volume calculation, compared with scintigraphy. The differences between the systems may be explained by the higher quality of the acquisitions with system 2, with reduced motion artifacts, because of the faster sweep at 3D US. During initial stages of performance of 3D US, some training is necessary to become accustomed to the scanning technique and to the viewing modalities; more time is necessary for reconstruction and volume calculation particularly during this phase. At 3D US, the size of the data set and the lack of Digital Imaging and Communications in Medicine standards for 3D data sets can create problems in data storage, in networking, and in integrating 3D US data into an existing picture archiving and communication system.

Finally, by using error and difference of relative renal volume or split renal function expressed as a percentage, calculation of the difference from this percentage was suboptimal; however, as scintigraphy provided only these values, the comparison with 2D and 3D US could be achieved only when we used the same scale.

We considered both advantages and limitations of 3D US and conclude from our experience that 3D US is feasible in most infants and children without sedation. On the basis of our results, 3D US appears to be accurate for assessment of real renal parenchymal volume and relative renal size in patients with hydronephrosis and is comparable to split renal function determined with scintigraphy, provided there is absence of acute and active renal disease. Bearing these factors in mind, on the basis of our initial experience, we suggest that 3D US of the kidney may become a helpful adjunct to 2D US of the urinary tract in patients with long-standing hydronephrosis to assess renal parenchymal volume.


   ACKNOWLEDGMENTS
 
The authors thank Franz Quehenberger, PhD, Institute for Medical Informatics, Statistics and Documentation, University of Medicine, Graz, Austria, for his review of statistical data.


   FOOTNOTES
 

Abbreviations: 3D = three-dimensional • 2D = two-dimensional

See Materials and Methods for pertinent disclosures.

Author contributions: Guarantor of integrity of entire study, M.R.; study concepts and design, M.R.; literature research, M.R., G.A.F., M.J.D.; clinical studies, M.R., C.J.M., G.A.F.; data acquisition, M.R., G.A.F., T.S.; data analysis/interpretation, M.R., G.A.F., T.S., H.S., M.J.D.; statistical analysis, G.A.F., H.S.; manuscript preparation and revision/review, G.A.F., M.R., H.S.; manuscript definition of intellectual content, M.R.; manuscript editing, M.R., H.S.; manuscript final version approval, all authors


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Dacher JN. Diagnostic procedures excluding MRI, nuclear medicine and video-urodynamics. In: Fotter R, ed. Pediatric uroradiology. Berlin, Germany: Springer, 2001; 1–15.
  2. Gordon I, Riccabona M. Investigating the newborn kidney: update on imaging techniques. Semin Neonatol 2003; 8:269–278.[CrossRef][Medline]
  3. Gordon I. Diagnostic procedures: nuclear medicine. In: Fotter R, ed. Pediatric uroradiology. Berlin, Germany: Springer, 2001; 27–40.
  4. Lavocat MP, Granjon D, Guimpied Y, et al. The importance of 99mTc-DMSA renal scintigraphy in the follow-up of acute pyelonephritis in children: comparison with urographic data. Nucl Med Commun 1998; 19:703–710.[Medline]
  5. O'Sullivan DC, Dewan PA, Guiney EJ. Compensatory hypertrophy effectively assesses the degree of impaired renal function in unilateral renal disease. Br J Urol 1992; 69:346–350.[Medline]
  6. Rossleigh MA, Farnsworth RH, Leighton DM, Yong JL, Christian CL. Technetium-99m dimercapto-succinic acid scintigraphy studies of renal cortical scarring and renal length. J Nucl Med 1998; 39:1280–1285.[Abstract/Free Full Text]
  7. Riccabona M, Ring E, Mileder P. Sonographie beim Harnwegsinfekt im Säuglingsalter. Klin Padiatr 1991; 203:372–376.[Medline]
  8. Sargent MA, Gupta SC. Sonographic measurement of relative renal volume in children: comparison with scintigraphic determination of relative renal function. AJR Am J Roentgenol 1993; 161:157–160.[Abstract/Free Full Text]
  9. Troell S, Berg U, Johansson B, Wikstad I. Comparison between renal parenchymal sonographic volume, renal parenchymal urographic area, glomerular filtration rate and renal plasma flow in children. Scand J Urol Nephrol 1988; 22:207–214.[Medline]
  10. Sargent MA, Long G, Karmali M, Cheng SM. Interobserver variation in the sonographic estimation of renal volume in children. Pediatr Radiol 1997; 27:663–666.[CrossRef][Medline]
  11. Downey DB, Fenster A. Vascular imaging with a three-dimensional power Doppler system. AJR Am J Roentgenol 1995; 165:665–668.[Free Full Text]
  12. Marx GR, Sherwood MC. Three-dimensional echocardiography in congenital heart disease: a continuum of unfilled promises? No. A presently clinically applicable technology with an important future? Yes. Pediatr Cardiol 2002; 23:266–285.[CrossRef][Medline]
  13. Merton DA, Bega G, Goldberg BB. Multiplanar 3-dimensional neonatal neurosonography: initial experiences and potential benefits. J Diagn Med Sonogr 2001; 17:3–15.[CrossRef]
  14. Merz E, ed. 3D-ultrasound in obstetrics and gynecology. Philadelphia, Pa: Lippincott, Williams & Wilkins, 1993.
  15. Nelson TR, Pretorius DH, Fenster A, Downey D, eds. Three-dimensional ultrasound. Philadelphia, Pa: Lippincott, Williams & Wilkins, 1999.
  16. Nelson TR, Pretorius DH, Sklansky M, Hagen-Ansert S. Three-dimensional echocardiographic evaluation of fetal heart anatomy and function: acquisition, analysis and display. J Ultrasound Med 1996; 15:1–9.[Abstract]
  17. De Odorico I, Spaulding KA, Pretorius DH, Lev Toaff AS, Bailey TB, Nelson TR. Normal splenic volumes estimated using three-dimensional ultrasonography. J Ultrasound Med 1999; 18:231–236.[Abstract]
  18. Elliot TL, Downey DB, Tong S, McLean CA, Fenster A. Accuracy of prostate volume measurements in vitro using three-dimensional ultrasound. Acad Radiol 1996; 3:401–406.[CrossRef][Medline]
  19. Gilja OH, Thune N, Matre K, Hausken T, Odegaard S, Berstad A. In vitro evaluation of three dimensional ultrasonography in volume estimation of abdominal organs. Ultrasound Med Biol 1994; 20:157–165.[CrossRef][Medline]
  20. Lang H, Wolf GK, Prokop M, et al. Three-dimensional ultrasound for volume measurement of liver tumors. Chirurg 1999; 70:246–250.[CrossRef]
  21. Nagdyman N, Walka MM, Kampmann W, Stover B, Obladen M. 3D ultrasound quantification of neonatal cerebral ventricles in different head positions. Ultrasound Med Biol 1999; 25:895–900.[CrossRef][Medline]
  22. Riccabona M, Nelson TR, Pretorius DH, Davidson TE. Distance and volume measurements using three dimensional ultrasound. J Ultrasound Med 1995; 14:881–886.[Abstract]
  23. Riccabona M, Nelson TR, Pretorius DH, Davidson TE. In vivo three-dimensional sonographic measurement of organ volume: validation in the urinary bladder. J Ultrasound Med 1996; 15:627–632.[Abstract]
  24. Schlogl S, Werner E, Lassmann M, et al. The use of three-dimensional ultrasound for thyroid volumetry. Thyroid 2001; 11:569–574.[CrossRef][Medline]
  25. Tong S, Cardinal HN, McLoughlin RF, Downey DB, Fenster A. Intra- and inter-observer variability and reliability of prostate volume measurement via two-dimensional and three-dimensional ultrasound imaging. Ultrasound Med Biol 1998; 24:673–681.[CrossRef][Medline]
  26. Treece G, Prager R, Gee A, Berman L. 3D ultrasound measurement of large organ volume. Med Image Anal 2001; 5:41–54.[CrossRef][Medline]
  27. Gilmore JH, Gerig G, Specter B, et al. Infant cerebral ventricle volume: a comparison of 3-D ultrasound and magnetic resonance imaging. Ultrasound Med Biol 2001; 27:1143–1146.[CrossRef][Medline]
  28. Kampmann W, Walka MM, Vogel M, Obladen M. 3-D sonographic volume measurement of the cerebral ventricular system: in vitro validation. Ultrasound Med Biol 1998; 24:1169–1174.[CrossRef][Medline]
  29. Riccabona M, Nelson TR, Resch B, Pretorius DP. Potential of three-dimensional ultrasound in neonatal and pediatric neurosonography: a pictorial essay. Eur Radiol 2003; 13:2082–2093.[CrossRef][Medline]
  30. Salerno CC, Pretorius DH, Hilton SW, et al. Three-dimensional ultrasound imaging of the neonatal brain in high risk neonates: a preliminary study. J Ultrasound Med 2000; 19:549–554.[Abstract]
  31. Stanojevic M, Hafner T, Kurjak A. Three-dimensional (3D) ultrasound: a useful imaging technique in the assessment of neonatal brain. J Perinat Med 2002; 30:74–83.[CrossRef][Medline]
  32. Fritz GA, Riccabona M, Bohdal G, Quehenberger F. Accuracy of renal volume assessment in children using three-dimensional sonography [in German]. Rofo 2003; 175:540–546.[Medline]
  33. Gilja OH, Smievoll AI, Thune N, et al. In vivo comparison of 3D ultrasonography and magnetic resonance imaging in volume estimation of human kidneys. Ultrasound Med Biol 1995; 21:25–32.[CrossRef][Medline]
  34. Matre K, Stokke EM, Martens D, Gilja OH. In vitro estimation of kidneys using three-dimensional ultrasonography and a position sensor. Eur J Ultrasound 1999; 10:65–73.[CrossRef][Medline]
  35. Riccabona M, Fritz G, Ring E. Potential applications of three-dimensional ultrasound in the pediatric urinary tract: pictorial demonstration based on preliminary results. Eur Radiol 2003; 13:2680–2687.[CrossRef][Medline]
  36. Dinkel E, Ertl M, Dittrich M, Peters H, Berres M, Schulte-Wissermann H. Kidney size in childhood: sonographical growth chart for kidney length and volume. Pediatr Radiol 1985; 15:38–43.[CrossRef][Medline]
  37. Fernbach SK, Maizels M, Conway JJ. Ultrasound grading of hydronephrosis: introduction to the system used by the Society for Fetal Urology. Pediatr Radiol 1993; 23:478–480.[CrossRef][Medline]
  38. Tubiana M. Carcinogenic effects of low radiation dose. Cancer Radiother 1999; 3:203–214.[Medline]
  39. Piepsz A. Radionuclide studies in paediatric nephro-urology. Eur J Radiol 2002; 43:146–153.[CrossRef][Medline]
  40. Nelson TR, Pretorius DH, Hull AD, Riccabona M, Sklansky MS, James G. Sources and impact of artifacts on clinical 3DUS imaging. Ultrasound Obstet Gynecol 2000; 15:1–11.



This article has been cited by other articles:


Home page
 Am. J. Roentgenol.Home page
P.-H. Vivier, M. Dolores, I. Gardin, P. Zhang, C. Petitjean, and J.-N. Dacher
In Vitro Assessment of a 3D Segmentation Algorithm Based on the Belief Functions Theory in Calculating Renal Volumes by MRI
Am. J. Roentgenol., September 1, 2008; 191(3): W127 - W134.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2361040158v1
236/1/276    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Riccabona, M.
Right arrow Articles by Mache, C. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Riccabona, M.
Right arrow Articles by Mache, C. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE