HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jones, R. A. Articles by Grattan-Smith, J. D. Search for Related Content PubMed PubMed Citation Articles by Jones, R. A. Articles by Grattan-Smith, J. D. Social Bookmarking                      What's this?

Dynamic Contrast-Enhanced MR Urography in the Evaluation of Pediatric Hydronephrosis: Part 1, Functional Assessment

¹ Department of Radiology, Emory University School of Medicine, Atlanta, GA.
² Department of Radiology, Children's Healthcare of Atlanta, 1001 Johnson Ferry Rd., Atlanta, GA 30342.
³ Department of Biostatistics, Emory University, Atlanta, GA.
⁴ Department of Pediatric Urology, Emory University School of Medicine, Atlanta, GA.
⁵ Department of Pediatric Urology, Children's Healthcare of Atlanta, Atlanta, GA.

Received September 30, 2004; accepted after revision March 8, 2005.

 
Address correspondence to R. A. Jones (richard.jones{at}choa.org).

Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
OBJECTIVE. The purpose of our study was to derive time-intensity curves for the renal cortex and medulla from 3D dynamic MR urography and to assess whether these curves are predictive of obstruction.

MATERIALS AND METHODS. Fifty-nine examinations were performed in 53 pediatric patients and the degree of obstruction assessed using the renal transit time. The cortex and medulla were segmented using a semiautomatic method, and mean time-intensity curves were derived for the segmented volumes. The basic parameters of the curves (amplitude, washout) were assessed, as was the presence of certain characteristic features of the curves.

RESULTS. The images allowed clear visualization of three phases of the uptake of contrast material in the cortex, the medulla, and the collecting system. Both the amplitude of the curves and the washout of the contrast material were predictive of obstruction. The distal tubular peak was reliably detected in the cortex of nonobstructed kidneys.

CONCLUSION. Combining signal-intensity-versus-time-curve analysis with the other parameters that can be derived from the same MR urography data set provides a powerful tool for the diagnosis of obstruction.

Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Obstructive uropathy refers to the obstruction of urine flow from the kidney to the bladder [1]. In children, this is usually a result of chronic partial obstruction, which is typically related to ureteropelvic junction (UPJ) obstruction or obstructive megaureter. The consequences of the obstruction not only depend on the degree of obstruction but also occur as a result of a complex syndrome resulting in alterations of both glomerular hemodynamics and tubular function caused by the interaction of a variety of vasoactive factors and cytokines [2]. Any attempt to separate obstructed from nonobstructed kidneys as distinct entities is artificial and unrealistic because a continuum of degrees of obstruction exists [3, 4]. All hydronephrotic systems are obstructed to some degree, and we need to develop sensitive measures to detect early renal functional deterioration. The ultimate goal of the management of obstruction is the preservation of renal function. Safe and effective interventions to correct urinary tract obstruction due to UPJ obstruction exist, but the necessity for, and timing of, these interventions remain controversial [5, 6]. Currently, no imaging technique can accurately assess the degree of obstruction or identify which kidneys are at risk for a progressive loss of renal function.

The Whitaker test is often considered the reference test for detecting and grading obstruction, but this invasive test is infrequently performed, provides no information on renal function, lacks objective criteria for pediatric studies, and uses nonphysiologic flow rates [7]. Thus, the default standard is diuresis renal scintigraphy, which uses either technetium-99m DTPA (dimethyltriamine pentaacetic acid) as a glomerular tracer or ^99mTc mercaptoacetyltriglycine (MAG3) as a tubular tracer, and an injection of furosemide to provide a diuretic challenge to the kidney. A number of protocols have been developed for diuresis renal scintigraphy using different timing for the administration of furosemide relative to the injection of the tracer [8]. Regardless of which protocol is used, time activity curves are derived for each kidney, and these are characterized by indexes such as the rate of washout of the tracer. The principal limitations of diuresis renal scintigraphy are that it relies on projection images, that it provides limited anatomic assessment of the urinary tract, and that it needs depth correction of counts and background subtraction.

MR urography provides anatomic images of the kidneys and ureters with excellent spatial and contrast resolution, and this can be combined with dynamic contrast-enhanced MR urography, which uses high temporal resolution to follow the passage of a contrast agent through the kidney. Dynamic MR urography can accurately assess the split renal function [9, 10] and provide useful information on renal function; however, the interpretation of the images is not straightforward and often involves extensive postprocessing [9, 10]. Recently, we have shown a simpler approach using hydration before scanning and the administration of furosemide 15 min before administering contrast material that allows the signal intensity to be used directly [11, 12]. This approach has been used to derive values for the differential renal function that were well correlated with those measured using diuresis renal scintigraphy [11-13]. The renal transit time is defined as the time between the arrival of contrast material in the cortex and its arrival in the ureter. The renal transit time has been shown to be well correlated with the half-life (T_1/2) washout time, which is widely used to characterize diuresis renal scintigraphy studies [4].

View larger version (83K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1A —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

View larger version (94K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1B —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

View larger version (68K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1C —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

View larger version (76K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1D —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

View larger version (98K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1E —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

View larger version (61K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1F —Dynamic images of 3-month-old boy referred for mild bilateral hydronephrosis. This patient had renal transit times of 1 min 37 sec on the left and 3 min 30 sec on the right, with slight delay on right still being within normal limits. Split renal function was calculated to be 49% on left and 51% on right. Images A-C show maximum intensity projections derived from volumes corresponding to three separate time points, whereas D-F show same slice from each of the three data sets. A and D show vascular or cortical phase. B and E were acquired 1 min 45 sec after vascular phase and show homogeneous enhancement of cortex and medulla. C and F were acquired 3 min 30 sec after vascular phase and show enhancement of calyces and ureters of both kidneys.

Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
MRI Methods
This study was approved by our hospital's institutional review board. All children referred by the urologists for suspected hydronephrosis were included in the study. A total of 107 examinations were performed on sedated patients between November 2002 and December 2003; five of these were unsuccessful because of failed sedation, motion, or problems with the injection of contrast material. Only patients with two functioning kidneys and with each kidney having a single collecting system were included in this study, resulting in a total of 59 examinations in 53 patients (39 boys, 14 girls), with six of the patients having a repeat examination. None of the patients had bilateral obstruction. The average age of these patients was 1.5 ± 1.9 (SD) years; the range of ages was 0.16-8.5 years. Thirty-four (58%) of the examinations were performed on patients who were younger than 1 year old.

View larger version (32K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2A —Relative signal-versus-time curves for 3-month-old boy shown in Figures 1A, 1B, 1C, 1D, 1E, and 1F. Graphs show signal from calyces (A) and salient features of curves (B). Specifically, note vascular peak in cortex (VASC), cortical distal convoluted tubular (DT) peak, and peak caused by contrast material in loop of Henle (LH) in medulla. Calyces show slight early signal enhancement, which is probably due to either contrast material in vessels running over surface of calyces or partial volume effects. Relative signals from cortex and medulla can be seen to fall in linear range of phantom data shown in Figure 3, whereas those from calyces are in nonlinear range. Hence, relative signal levels in calyces are somewhat underestimated and form of calyceal curves will be distorted.

View larger version (16K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3 —Graph shows relative signal intensity from saline phantom. Illustrated straight fit was calculated for gadopentetate dimeglumine concentrations in range of 0-10 mmol/L. At higher concentrations, response becomes nonlinear.

View larger version (28K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2B —Relative signal-versus-time curves for 3-month-old boy shown in Figures 1A, 1B, 1C, 1D, 1E, and 1F. Graphs show signal from calyces (A) and salient features of curves (B). Specifically, note vascular peak in cortex (VASC), cortical distal convoluted tubular (DT) peak, and peak caused by contrast material in loop of Henle (LH) in medulla. Calyces show slight early signal enhancement, which is probably due to either contrast material in vessels running over surface of calyces or partial volume effects. Relative signals from cortex and medulla can be seen to fall in linear range of phantom data shown in Figure 3, whereas those from calyces are in nonlinear range. Hence, relative signal levels in calyces are somewhat underestimated and form of calyceal curves will be distorted.

Once the patient was positioned in the scanner, scout images were acquired to determine both the positioning of the kidneys and bladder and the combination of spine coil elements required to optimize the signal-to-noise ratio for these anatomic structures. After the scout images were completed, furosemide (1 mg/kg of body weight; maximum, 20 mg) was administered IV. Coronal 2D T1-weighted (TR/TE, 40/17) and T2-weighted (5,500/210; echo-train length, 29) series and a respiratory-triggered, heavily T2-weighted 3D sequence (TE, 600; echo-train length, 109) were then acquired [4]. The acquisition of these three series required approximately 15 min. The 2D series provided detailed anatomic references, and the heavily T2-weighted 3D series provided the basis for an unenhanced maximum intensity projection (MIP) of the collecting system, ureters, and bladder [10]. To create the MIP, other T2 structures with long T2 relaxation times, such as CSF and the gallbladder, were manually edited out of the images.

Once these sequences were complete, the acquisition of a 3D dynamic gradient-echo sequence (3.4/1.5; flip angle, 30°) oriented along the axis of the kidneys was initiated. The start of the dynamic series was approximately 15 min after the injection of furosemide, which coincides with the maximum effect of the furosemide [14]. A bolus injection of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist, Berlex Laboratories) was hand-injected during the acquisition of the fourth dynamic image. Each time point of the dynamic sequence consisted of 36 slices, with the outer three slices on each side being discarded to limit variations in the flip angle related to the slice profile and also to limit wrap-around artifacts. The scans were acquired contiguously for the first 5 min, resulting in a temporal resolution of 15 sec. Subsequently, intervals of 45 sec were inserted between the scans, resulting in a temporal resolution of 1 min. For each volume, an MIP of the whole volume was automatically generated. If both ureters were clearly visualized 10 min after the injection of contrast media, no further dynamic series were acquired. Otherwise, the acquisition of dynamic images was continued for an additional 5 min at 1-min intervals.

After completion of the dynamic series, sagittal, axial, and coronal 3D images with high spatial resolution were acquired for calculating a contrast-enhanced MIP. In cases in which poor drainage from the renal pelvis meant that no contrast material was observed in the ureters during the 15 min of dynamic scanning, the patient was turned to a prone position to promote mixing of the contrast agent in the collecting system before the acquisition of high-spatial-resolution coronal and sagittal images. The total imaging time for patients with nonobstructed kidneys was typically 45 min; for patients with poorly draining kidneys, the imaging time was typically 1 hr.

View larger version (58K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4 —Maximum intensity projection from delayed 3D volumes acquired in same 3-month-old boy as shown in Figures 1A, 1B, 1C, 1D, 1E, and 1F. This delayed 3D volume was acquired with high spatial resolution (0.8 x 0.8 x 1.0 mm) and was obtained after dynamic volumes. Right kidney has an extrarenal pelvis and persistent fetal folds of right ureter. Latter probably account for slightly longer renal transit time observed on right side.

Each cortex was then segmented using a semiautomatic algorithm based on user-defined intensity thresholds, morphologic erosion, and dilation and region growing steps. In all cases, the same intensity thresholds were used for both kidneys. Binary maps of the resulting volumes were saved to disk. For the isointense volumes, no editing was generally required before performing the segmentation.

The resulting isointense (parenchymal) volumes were used to calculate the volume of each kidney; these volumes were then used to calculate the differential renal function [12, 13]. For each kidney, the cortical volume was then subtracted from the isotropic volume, and the resulting volume was then segmented to yield the medullary volume. For severely obstructed kidneys, it was not possible to reliably segment the medulla; hence only the cortex was segmented in these patients.

The mean signals for the cortical and medullary volumes at each time point were then calculated. The resulting signal values were converted to relative signal using the formula S_t/S₀, where S_t is the signal at time t and S₀ is the mean unenhanced signal (which was calculated as the mean of the three unenhanced volumes). The relative signal has the advantage of removing residual intensity variations resulting from spatial varying coil sensitivities.

Curve Analysis
Figures 1A, 1B, 1C, 1D, 1E, and 1F shows typical images and dynamic MIPs obtained from a patient with renal transit times within the normal range. The relative signal-versus-time curves from the same patient are shown in Figures 2A, and 2B; these illustrate some of the characteristic features seen in these curves. In the cortex, an initial peak corresponding to concentrated contrast material in the vasculature was seen; subsequently, a second peak was seen that is believed to represent the passage of contrast material into the distal tubules [15]. For the medulla, a single peak was observed that corresponds to contrast material in the loop of Henle.

The relative signal-versus-time curves obtained for each patient were quantitated to find the maximum relative signal levels in, and the initial rate of washout from, the cortex and medulla (in the former case, the initial vascular peak was excluded). For the cortex, the washout was calculated for the interval between the distal tubular peak and the end of the dynamic scanning with 15-sec temporal resolution, whereas for the medulla the time between the peak resulting from contrast material in the loop of Henle and the same end point was used. If three or more points were available, the washout rate was calculated using a linear fit; otherwise no fitting was performed.

Two radiologists assessed the curves for the presence or absence of the distal tubular peak; in the cases in which a peak was detected, the time of the distal tubular peak was also recorded. In addition, the radiologists also assessed whether the time point at which the medullary signal exceeded the cortical signal (crossover point) was symmetric or asymmetric for the two kidneys (i.e., compared the curves from the two kidneys). For each patient, a radiologist assessed the renal transit time, and this was used to classify the kidney as being normal (time 245 sec), equivocal (time > 245 but 490 sec), or obstructed (time > 490 sec), according to a previously validated classification [4].

Phantom Studies
To assess the range for which the variation of the relative signal with gadopentetate dimeglumine concentration is linear, a phantom consisting of tubes of saline doped with different concentrations of gadolinium was imaged using the same 3D pulse sequence as used for the patient studies. The signal in each tube was measured in a region of interest positioned in the center of the tube, and the relative signal was calculated using the signal from a tube containing no gadopentetate dimeglumine as the reference (S₀). The results from the phantom study are shown in Figure 3. The results showed that for the sequence used for this study, a flip angle of 30° produced a linear correlation between the relative signal and the concentration of gadopentetate dimeglumine for concentrations up to approximately 1 mmol/L. At higher concentrations, the rate of change of the relative signal with increasing gadopentetate dimeglumine concentration was attenuated.

Statistical Analysis
Repeated measures analyses for the peak signal and rate of washout from both kidneys were analyzed using a means model with SAS Proc Mixed software (version 8) (SAS Institute) providing separate estimates of the means by grade of renal transit time (normal, equivocal, and obstructed). The analysis was performed separately for the cortex and medulla data. A compound symmetry variance-covariance form among the repeated measurements was assumed for each outcome, and robust estimates of the standard errors of parameters were used to perform statistical tests and construct 95% confidence intervals. Student's t tests (paired) were used to compare the three pairwise differences between renal transit time grades for each outcome. The model-based means are unbiased with unbalanced and missing data, so long as the missing data are noninformative (missing at random). Statistical tests were two-sided.

Interobserver reliability for crossover points and for the presence of distal tubular peaks was summarized using the kappa statistic and was calculated separately for left and right kidney data. The kappa statistic is a chance-corrected measure of agreement that equals 0 if agreement among reviewers is equal to what would be expected based on chance alone, and 1 if there is perfect agreement among reviewers. The agreement between the two radiologists as to the temporal position of the distal tubular peak was explored by Bland-Altman analysis, which plots the difference between two measurements against the combined overall mean. The 95% limits of agreement were calculated as described by Bland and Altman [16].

Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
The relative signal-versus-time curves shown in Figures 2A, and 2B exhibit relative signal levels for the cortex and medulla that are within the linear range of the sequence; however, the curves from the calyces are in the nonlinear range. Thus, the curves will underestimate the contrast material concentration in the calyces, and the form of the washout from the calyces will also tend to be "flattened" by the nonlinear response at high concentrations of gadopentetate dimeglumine. No signal dropout artifacts in the calyces were observed, which implies that the combination of hydration and furosemide prevented the concentration of gadopentetate dimeglumine from becoming sufficiently high for T2^* shortening to become the dominant effect. Figure 4 shows one projection from the MIP derived from the 3D sagittal series acquired after the dynamic series. Minimal prominence of the right renal pelvis without significant caliectasis is seen. Multiple persistent fetal folds can be seen in the right ureter, which probably account for the slight delay in the renal transit time on this side.

Differential Renal Function
Fifty-nine imaging examinations (118 kidneys) were included in this study. Nineteen of these (32%) had markedly nonsymmetric differential renal function (defined as one kidney having 40% or less of the total renal function); of these, 11 had a renal transit time in the obstructed category, four were in the equivocal category, and four had a normal renal transit time. Of the four patients with an abnormal value for the differential renal function and a normal renal transit time, two had dysplastic kidneys, one had reflux nephropathy, and one was imaged 6 months after a pyeloplasty (which had removed the obstruction but presumably had not restored the function). The enhancement of the cortex was symmetric in all cases. Differences in the time at which the medulla enhanced were most common in children with an asymmetric differential renal function.

Relative Signal Statistics
We were able to segment the cortex for all 118 kidneys; however, we were unable to segment the medulla in eight (7%) of the 118 kidneys. Of these eight kidneys, six had a renal transit time in the obstructed category and one was in each of the other two categories. The patient with normal renal transit time was imaged after pyeloplasty on a previously obstructed kidney. Six of the eight patients whose medulla could not be segmented also had a nonsymmetric differential renal function. There were two kidneys for which no signal decay was observed and hence for which no washout was calculated; neither of these had a renal transit time in the obstructed category, but both were contralateral to kidneys with a renal transit time in the obstructed category.

For washout in the medulla, there were four patients for whom the medulla could be segmented but in which no washout was observed; thus the washout was calculated for only 106 kidneys. Of these four, three had renal transit times in the obstructed category and three had an abnormal differential renal function.

Table 1 summarizes the peak relative signal values obtained in the cortex and the medulla for the three grades of renal transit time (normal, equivocal, and obstructed). For the cortex, the statistical analysis yielded significant differences between normal and equivocal kidneys (p = 0.005) and between normal and obstructed kidneys (p < 0.0001), whereas the difference between equivocal and obstructed categories was not significant (p = 0.115). For the medulla, the corresponding values were not significantly different for normal versus equivocal (p = 0.064) but were significantly different for normal versus obstructed (p < 0.0001) and equivocal versus obstructed (p = 0.02).

Table 2 summarizes the rate of washout of the relative signal from the cortex and the medulla. For the cortex, the differences were significant between normal and equivocal kidneys (p = 0.0003) and between normal and obstructed kidneys (p < 0.0001). The difference between equivocal and obstructed kidneys was not significant (p = 0.12). For the medulla, the corresponding values were normal versus equivocal, p = 0.0007; normal versus obstructed, p < 0.0002; and equivocal versus obstructed, p = 0.044.

Cortical Peak
For the test of interobserver reliability for determining whether a distal tubular peak was present, all 59 studies were included. For the left kidneys, the observed agreement was 86.4%, compared with a chance agreement of 62.9%, resulting in a kappa value of 0.64 (SE, 0.11). For the right kidneys, the values were 86.4% and 56.7%, respectively, and the kappa value was 0.69 (SE, 0.10). In both cases, the kappa value implies that there is substantial agreement between the observers [17].

The results of the Bland-Altman analysis of the interobserver variability in the temporal position of the distal tubular peak as estimated by the radiologists are shown in Figure 5. The mean difference was 3.6 sec (SD, 8.05 sec), and the 95% limits of agreement were -12.5 to 19.6 sec.

View larger version (14K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5 —Bland-Altman plot summarizes results of interobserver position of distal convoluted tubule peak. Mean difference and limits of agreement can be seen to be good, given temporal resolution of 15 sec used for dynamic studies.

Crossover Point
For the crossover points, 56 studies were assessed. The observed agreement was 87.5%, the chance agreement was 51.4%, and the kappa value was 0.74 (SE, 0.09). This kappa value also indicates that substantial agreement existed between the observers. To study the correlation between the renal transit time and the crossover point, patients in whom the medullary signal was absent from one or both kidneys and patients with equivocal renal transit times were excluded, leaving 33 studies. Among these 33, four had delayed crossover points; of these, two (50%) had a renal transit time in the obstructed range. Among the 29 with symmetric crossover points, four (13.8%) had a renal transit time in the obstructed range (p = 0.14, Fisher's exact test).

Serial Imaging
Figures 6A, 6B, 7A, and 7B show images and curves, respectively, obtained on an infant boy at 3 and 9 months old and show the ability of dynamic MR urography to document the evolution of obstruction. In Figures 6A, and 6B, significant improvement can be seen in the degree of hydroureteronephrosis as the child grows. At 9 months old, the right ureter is less dilated and tortuous than in the scan obtained when the child was 3 months old. Signal-intensity-versus-time curves at the two different ages also support the improvement in hydroureteronephrosis by showing improvement in the curve morphology. At 9 months the corticomedullary crossover point is nearly symmetric, suggesting that the intratubular pressure is returning to normal.

View larger version (55K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6A —Delayed contrast-enhanced maximum intensity projections (MIPs). Infant boy who was studied at 3 months (A) and 9 months (B) old. Renal transit times were 100 sec on left and 350 sec on right in first study. In second study, corresponding times were 90 sec on left and 225 sec on right. MIPs show improvement in degree of hydroureteronephrosis, with elongation and decreased tortuosity of ureter.

View larger version (54K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6B —Delayed contrast-enhanced maximum intensity projections (MIPs). Infant boy who was studied at 3 months (A) and 9 months (B) old. Renal transit times were 100 sec on left and 350 sec on right in first study. In second study, corresponding times were 90 sec on left and 225 sec on right. MIPs show improvement in degree of hydroureteronephrosis, with elongation and decreased tortuosity of ureter.

View larger version (28K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7A —Relative signal-versus-time curves for two studies from patient shown in Figures 6A, and 6B. Initial study when infant was 3 months old (A) shows delayed crossover point. By time of second study when infant was 9 months old (B), crossover point has become symmetric, suggesting that intracortical pressure has returned to normal. Distal cortical peak is preserved in both studies, as well as concentration of contrast material in medulla and loop of Henle. Note mild decrease in amplitude of cortex on right, indicating slight impairment in ability of this kidney to extract and concentrate contrast material from vasculature.

View larger version (29K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7B —Relative signal-versus-time curves for two studies from patient shown in Figures 6A, and 6B. Initial study when infant was 3 months old (A) shows delayed crossover point. By time of second study when infant was 9 months old (B), crossover point has become symmetric, suggesting that intracortical pressure has returned to normal. Distal cortical peak is preserved in both studies, as well as concentration of contrast material in medulla and loop of Henle. Note mild decrease in amplitude of cortex on right, indicating slight impairment in ability of this kidney to extract and concentrate contrast material from vasculature.

Top Abstract Introduction Materials and Methods Results Discussion References

MR Urography
The use of hydration and furosemide to improve the quality of MR urography has been previously documented [11, 12]. Furosemide increases both the volume and the flow of urine, resulting in dilution of the gadopentetate dimeglumine in the collecting system and a rapid and uniform distribution of the gadopentetate dimeglumine in the collecting system and urinary tract. The gadopentetate dimeglumine affects both the T1 and T2 relaxation times, and the in vivo concentrations can be calculated exactly by measuring the relaxation times. However, a dynamic measurement of the relaxation times is challenging in a clinical environment and would compromise the temporal resolution. Several techniques have been developed using additional unenhanced and contrast-enhanced scanning to calculate the relaxation time, but these are also difficult to apply to routine clinical studies [21].

We chose to use the ratio of the unenhanced and contrast-enhanced signals from a 3D sequence and validated this approach using phantom studies. The use of a 3D sequence removes most slice profile effects that complicate the analysis of 2D dynamic series [9, 21]. To obtain high temporal resolution while minimizing the T2^* shortening effect of gadopentetate dimeglumine, a short TR/TE gradient-echo sequence with optimized fat suppression was used. The TE of 1.5 sec used in this study meant that the duration of the radiofrequency pulse is very short; consequently, the maximum radiofrequency power available, rather than the specific absorption rate, determined the maximum attainable flip angle. We found that a flip angle of 30° could be routinely obtained but that higher flip angles were not possible. A higher flip angle would provide a greater range over which a linear response to gadopentetate dimeglumine concentration could be obtained [22], but at the cost of a reduced signal-to-noise ratio. In studies in which it is necessary to estimate the concentration of gadopentetate dimeglumine in the collecting system or in the vasculature, either a lower dose of contrast material or a greater flip angle will probably be required.

A standard dose of gadopentetate dimeglumine was used in this study because we were concerned only with the signals in the renal parenchyma, and these appear to fall within the linear range. The high concentration of contrast material in the collecting system means that its transverse relaxation times will be rather short, which will degrade the point-spread function of the voxels in the collecting system. The degraded point-spread function, combined with the high signal intensity in the collecting system, will result in some contamination of the parenchyma by signal from the collecting system.

A full or rapidly filling bladder, or a bladder with poor compliance, may affect upper tract emptying and lead to false-positive results for obstruction [23]. Because the patients in this study were all sedated and could not be asked to void, a bladder catheter was placed for all studies to mitigate these problems. The catheter also ensures that reflux does not distort the results of the study.

The MIPs derived from the dynamic volume corresponding to the first passage of the gadopentetate dimeglumine permit an evaluation of the major arteries, and the venous system is generally well seen in the subsequent volume. The arterial MIP allowed the basic architecture of the renal arteries, and the presence of any accessory renal arteries, to be assessed.

Renal Volume
The differential renal function as measured by diuresis renal scintigraphy is based on the integration of the tracer curve over a range of time points at which the tracer is assumed to be predominately in the parenchyma. Because diuresis renal scintigraphy measurements are based on projection images of the whole kidney, they measure the activity in the whole kidney. Most techniques developed for measuring the differential renal function with MRI have attempted to duplicate this approach by combining the area under the time-intensity curve obtained from either a single slice [10] or a few slices [9] with a separate volume measurement.

Our approach is slightly different. Because the 3D volumes used in this study cover the full extent of both kidneys, we could follow the uptake of the contrast material in each kidney. We made the assumption that voxels represent either functional or nonfunctional tissue and that, by summing the voxels that show a significant uptake of contrast material, one can calculate the functional volume of each kidney and hence the split renal function. The dynamic series were visually inspected to determine the volumes in which contrast material is first seen in the collecting system of each kidney; the volume before this is then used for the calculation of the functional volume of each kidney. In this way, possible differences between the two kidneys are taken into account, and it is not necessary to assume a particular time point, or range of times, for the calculation. In previous studies, we have shown that this agrees well with the differential renal function calculated using nuclear medicine studies [11, 12].

For the dose of contrast material, and pulse sequence parameters, used in this study, the segmentation of the kidneys at this homogeneous enhancement phase was straightforward. The operator was required only to place a seed point and adjust the threshold limits; no manual editing was generally required, and the calculation of the volume typically required less than a minute per kidney. The methodology presented here accurately assumes a clear distinction between functioning and nonfunctioning tissue. It does account for the effects of cortical scarring or unusual morphology, such as that seen in polycystic kidneys.

Time-Intensity Curves
The initial part of the relative signal-versus-time curves derived for the medulla and cortex showed changes in the relative signal that corresponded to the initial arrival of the contrast agent in the cortical blood vessels. The changes at subsequent time points reflected the glomerular filtration and tubular concentration of the contrast medium. Because of the vascular nature of the cortex and the high concentration of the contrast agent during its initial transit through the kidney, an initial peak in the signal is clearly seen in the cortex. The transit of the contrast material to the proximal tubules is too fast to be clearly shown with the temporal resolution used in this study; however, the delayed arrival of contrast material in the distal tubules can be seen in normal kidneys.

In the medulla, the vascular signal is weaker, reflecting the lower vascularity of the medulla. The curves from the medulla are dominated by the peak corresponding to the arrival of contrast material in the loop of Henle. Multiparametric curve fitting has been used to describe the passage of the contrast material into each of these compartments for electron beam CT data acquired with high temporal resolution [15]; however, this was not possible with the temporal resolution used for this study.

The interobserver analysis for the presence and position of delayed cortical peak corresponding to contrast material in the distal tubules indicated that the detection of this peak is excellent, implying that it can be used as a marker for normal cortical renal function. The intensity of the cortical and medullary peaks and the rate of washout of contrast material from these compartments all showed statistically significant variations between kidneys classified as normal and those classified as obstructed by the renal transit time (Tables 1 and 2).

Differentiating between normal and equivocal kidneys, and between equivocal and obstructed systems, was more problematic because the categories tend to overlap; in such cases, follow-up studies were typically performed (see Figs. 6A, 6B, 7A, and 7B). In eight kidneys, we were unable to segment the medulla; six of these kidneys had a renal transit time in the obstructed range. In these cases, the medulla typically formed a thin rind along the inner edge of the cortex, making reliable segmentation difficult. All these kidneys had decreased differential renal function with loss of renal volume, most particularly involving the renal medulla. The remaining kidney had a normal renal transit time but the study was performed after successful pyeloplasty.

In a small study of 11 adults, Katzberg et al. [20] reported that the point at which the relative signal from the medulla exceeds that from the cortex (the crossover point) could differentiate between normal and obstructed kidneys. They suggested that the delay in rise of the medullary signal intensity curve may be secondary to increased intratubular hydrostatic pressure resulting from obstruction. In this study, we compared the crossover point in patients with renal transit times in the normal and obstructed ranges to see if an association existed between the crossover point and the renal transit time. In our study, the relationship between the crossover point and the renal transit time was not statistically significant; however, the number of patients with renal transit times in the obstructed range and for whom the medulla could be segmented was small. A further study with a larger patient population will be required to investigate the correlation between the crossover point and obstruction in more detail. If delay in the crossover point indicates increased intratubular pressure, it may also be predictive of patients at risk for progressive loss of renal function and hence identify patients with hydronephrosis who may benefit from early surgery. A randomized prospective trial is needed to establish this hypothesis. Higher temporal resolution may also permit a more precise definition of the crossover point.

By combining the various indexes (differential renal function, renal transit time, peak relative signals, washout rates, crossover point) provided by MR urography with the anatomic information available from the same MRI examination, it is possible to provide an accurate diagnosis of obstruction. Using this technique, we can quantitatively evaluate the ability of the cortex and medulla of each kidney to both concentrate and excrete the contrast agent. Subtle changes in renal function can be analyzed, and this technique may provide insights into renal pathophysiology. Not all patients with obstruction develop progressive loss of renal function. Further analysis of these parameters may allow better delineation of those patients who will benefit from early surgery.

Future Developments
Several techniques for measuring the glomerular filtration rate using MRI have been reported. Most of these involve the measurement of relaxation times and are rather impractical for routine clinical use [20, 24, 25]. Recently, the application of a Rutland-Patlak analysis to the data from dynamic MR urography has been shown to provide an accurate estimation of the glomerular filtration rate in adults [22]. This approach requires the measurement of an arterial input function.

For the sequence used in our study, the values measured in arterial blood are likely to be in the nonlinear range and hence distorted. Furthermore, the first passage of the contrast agent and the initial wash-in of the contrast material into the kidney were somewhat undersampled because of the limited temporal resolution. By shortening the time required for each time point using parallel imaging and a slower infusion, or a lower dose of contrast material, it should be possible to extend the technique described here to include a measurement of single-kidney glomerular filtration rate.

Limitations
The technique we have described requires patient preparation in the form of hydration and the placement of a bladder catheter; it also requires the administration of furosemide and gadopentetate dimeglumine. For children in the age range reported in this article, sedation is also required; however, in older children we have been able to acquire data without using sedation, although in these patients the motion of the kidneys is more problematic. Using a navigator to monitor the position of the diaphragm and using this to phase-correct the data can result in the acquisition of high-quality static images during normal respiration. For dynamic imaging, asking the patients to perform shallow breathing using the belly, rather than the thorax, generally produces good results. The examination generally requires a 1-hr slot of scanner time, and the hydration and sedation require extra nursing time before scanning. The processing to produce the curves and differential renal function is done offline on a PC and typically requires an additional 20-30 min.

The main problem for MR urography has been cost; however, for the technique presented here, the cost is offset by the fact that a single MRI examination provides anatomic and functional information and an assessment of the vasculature. Obtaining a similar amount of information would require the use of several techniques, and the quality of the information obtained is likely to be inferior to that obtained with MRI. The rationale for using MRI will be further improved if an accurate measurement of glomerular filtration rate can be incorporated into the methodology presented in this article.

References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. O'Reilly PH. Obstructive uropathy. Q J Nucl Med 2002; 46:295 -303[Medline]
2. Wen JG, Frøkær J, Jørgensen TM, Djurhuus JC. Obstructive nephropathy: an update of the experimental research. Urol Res 1998; 27:29 -39
3. Peters CM. The long-term follow-up of prenatally detected severe bilateral newborn hydronephrosis initially managed non-operatively. (editorial) J Urol 2002;168 : 1121-1122[CrossRef][Medline]
4. Jones RA, Perez-Brayfield MR, Kirsch AJ, Grattan-Smith JD. Renal transit time with MR urography in children. Radiology2004; 233:41 -50[Abstract/Free Full Text]
5. Houben C, Wischermann A, Borner G, Slany F. Outcome analysis of pyeloplasty in children. Pediatr Surg Intl2000; 16:189 -193
6. Koff S, Campbell KD. The nonoperative management of unilateral neonatal hydronephrosis: natural history of poorly functioning kidneys. J Urol 1994; 152:593 -595
7. Wåhlin N, Magnusson A, Persson AE, Läckgren G, Stenberg A. Pressure flow measurement of hydronephrosis in children: a new approach to definition and quantification of obstruction. J Urol2001; 166:1842 -1847[CrossRef][Medline]
8. O'Reilly P, Aurell M, Britton K, Kletter K, Rosenthal L, Testa T. Consensus on diuresis renography for investigating the dilated upper urinary tract. J Nucl Med 1996;37 : 1872-1876[Free Full Text]
9. Taylor J, Summers PE, Keevil SF, et al. Magnetic resonance renography: optimisation of pulse sequence parameters and Gd-DTPA dose, and comparison with radionuclide renography. Magn Reson Imaging 1997; 15:637 -649[CrossRef][Medline]
10. Rohrschneider WK, Haufe S, Wiesel M, et al. Functional and morphologic evaluation of congenital urinary tract dilatation by using combined static-dynamic MR urography: findings in kidneys with a single collecting system. Radiology 2002;224 : 683-694[Abstract/Free Full Text]
11. Teh HS, Ang ES, Weng WC, et al. MR renography using a dynamic gradient-echo sequence and low-dose gadopentetate dimeglumine as an alternative to radionuclide renography. AJR2003; 181:441 -450[Abstract/Free Full Text]
12. Grattan-Smith JD, Perez-Brayfield MR, Jones RA, et al. MR imaging of kidneys: functional evaluation using F-15 perfusion imaging. Pediatr Radiol 2003;33 : 293-304[Medline]
13. Perez-Brayfield MR, Kirsch AJ, Jones RA, Grattan-Smith JD. A prospective study comparing ultrasound, nuclear scintigraphy and dynamic contrast enhanced magnetic resonance imaging in the evaluation of hydronephrosis. J Urol 2003;170 : 1330-1334[CrossRef][Medline]
14. Brown SC, Upsdell SM, O'Reilly PH. The importance of renal function in the interpretation of diuresis renography. Br J Urol 1992; 69:121 -125[Medline]
15. Krier JD, Ritman EL, Bajzer Z, Romero JC, Lerman A, Lerman LO. Noninvasive measurement of concurrent single-kidney perfusion, glomerular filtration, and tubular function. Am J Physiol Renal Physiol 2001; 281:F630 -F638[Abstract/Free Full Text]
16. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8 : 135-160[Abstract/Free Full Text]
17. Landis J, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33 : 159-174[CrossRef][Medline]
18. Kikinis R, von Schulthess GK, Jager P, et al. Normal and hydronephrotic kidney: evaluation of renal function with contrast-enhanced MR imaging. Radiology 1987;165 : 837-842[Abstract/Free Full Text]
19. Semelka RC, Hricak H, Tomei E, Floth A, Stoller M. Obstructive nephropathy: evaluation with dynamic Gd-DTPA-enhanced MR imaging. Radiology 1990;175 : 797-803[Abstract/Free Full Text]
20. Katzberg RW, Buonocore MH, Ivanovic M, et al. Functional, dynamic, and anatomic MR urography: feasibility and preliminary findings. Acad Radiol 2001;8 : 1083-1099[CrossRef][Medline]
21. Brookes JA, Redpath TW, Gilbert FJ, Murray AD, Staff RT. Accuracy of T1 measurement in dynamic contrast-enhanced breast MRI using two- and three-dimensional variable flip angle fast low-angle shot. J Magn Reson Imaging 1999; 9:163 -171[CrossRef][Medline]
22. Hackstein N, Heckrodt J, Rau WS. Measurement of single-kidney glomerular filtration rate using a contrast-enhanced dynamic gradient-echo sequence and the Rutland-Patlak technique. J Magn Reson Imaging 2003; 18:714 -725[CrossRef][Medline]
23. O'Reilly PH. Standardization of the renogram technique for investigating the dilated upper urinary tract and assessing the results of surgery. BJU Int 2003;91 : 239-243[CrossRef][Medline]
24. Dumoulin CL, Buonocore MH, Opsahl LR, et al. Noninvasive measurement of renal hemodynamic functions using gadolinium enhanced magnetic resonance imaging. Magn Reson Med 1994;32 : 370-378[Medline]
25. Niendorf ER, Grist T, Lee FT Jr, Brazy PC, Santyr GE. Rapid in vivo measurement of single-kidney extraction fraction and glomerular filtration rate with MR imaging. Radiology 1998;206 : 791-798[Abstract/Free Full Text]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				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]

				J. R. Leyendecker, C. E. Barnes, and R. J. Zagoria MR Urography: Techniques and Clinical Applications RadioGraphics, January 1, 2008; 28(1): 23 - 46. [Abstract] [Full Text] [PDF]

				A. Boubaker, J. O. Prior, J.-Y. Meuwly, and A. Bischof-Delaloye Radionuclide Investigations of the Urinary Tract in the Era of Multimodality Imaging J. Nucl. Med., November 1, 2006; 47(11): 1819 - 1836. [Abstract] [Full Text] [PDF]

B. B. McDaniel, R. A. Jones, H. Scherz, A. J. Kirsch, S. B. Little, and J. D. Grattan-Smith Dynamic Contrast-Enhanced MR Urography in the Evaluation of Pediatric Hydronephrosis: Part 2, Anatomic and Functional Assessment of Uteropelvic Junction Obstruction Am. J. Roentgenol., December 1, 2005; 185(6): 1608 - 1614. [Abstract] [Full Text] [PDF]