| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Departments of Radiology (C.H.C., K.L.W., D.M.S., N.J.P., S.T.K., A.M.S.G., F.G.S.), Electrical Engineering (J.H.L.), Surgery (B.B.H.), and Comparative Medicine (D.M.B.), and the Division of Nephrology (G.C.D., B.D.M.), Stanford University School of Medicine, 300 Pasteur Dr, H-1307, Stanford, CA 94305. From the 2000 RSNA scientific assembly. Received February 5, 2001; revision requested March 20; final revision received August 22; accepted October 10. Supported in part by NIH grants R01-DK48051 and T32-CA90695. Address correspondence to F.G.S. (e-mail: gsommer@stanford.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: High-grade unilateral renal arterial stenoses were surgically created in eight pigs. Direct measurements of renal venous and arterial inulin concentration provided reference standard estimates of single-kidney EF. Pigs were imaged with a 1.5-T imager to estimate EF, renal blood flow, and glomerular filtration rate. A breath-hold inversion-recovery spiral sequence was used to measure T1 of blood in the infrarenal inferior vena cava and renal veins after intravenous administration of gadopentetate dimeglumine, and these data were used to calculate EF. Cine-phase contrast material–enhanced imaging of the renal arteries provided quantitative renal blood flow measurements. Bilateral single-kidney glomerular filtration rate was then determined: glomerular filtration rate = renal blood flow x (1 - hematocrit level) x EF.
RESULTS: A statistically significant linear correlation was found between EF, as determined with MR imaging, and inulin (r = 0.77). As compared with kidneys without renal arterial stenosis, kidneys with renal arterial stenosis showed 50% (0.14/0.28) EF reduction (P < .01) and 59% glomerular filtration rate reduction (P < .01).
CONCLUSION: MR imaging shows promise for in vivo measurement of EF and glomerular filtration rate, which may be useful in assessing the clinical importance of renal arterial stenosis.
© RSNA, 2002
Index terms: Animals • Kidney, function, 81.764 • Kidney, ischemia, 81.764 • Kidney, MR, 81.121412, 81.121413, 81.12143, 81.12144 • Kidney, stenosis or obstruction, 81.481 • Renal arteries, stenosis or obstruction, 961.729
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The purpose of our study was to test whether MR imaging can enable accurate measurement of EF in swine with unilateral renal ischemia and to evaluate the effects of RAS on EF and single-kidney GFR.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Surgery and EFIN
On the morning of the study, each animal was premedicated with 5–7 mg per kilogram of body weight telazol and 0.05 mg/kg atropine intramuscularly, intubated, and mechanically ventilated. General anesthesia was maintained with isoflurane (1%–3%) and oxygen (1.5 L/min) for the duration of the procedure, and the anesthetized animals were continuously monitored with pulse oximetry and temperature probes. A 6-F multiple-side-hole suprapubic catheter was placed in the bladder. Angiographic catheters were placed in the aorta and both renal veins by using the Seldinger technique, with vascular access via the unilateral common femoral artery and bilateral common femoral veins. All pigs underwent volume expansion with 500-mL Lactate Ringers to ensure adequate hydration. The pigs were then given an intravenous bolus of 1 g inulin (Questcor Pharmaceuticals, Hayward, Calif), followed with continuous 14 mL/h infusion of a 9% solution of inulin by using an infusion pump (Ivion, Broomfield, Colo). The left main renal artery was surgically exposed, and a 6-mm inflatable cuff occluder (In Vivo Metrics, Healdsburg, Calif) was applied proximal to the hilar branches. A stenosis was created by reducing vessel diameter by 75%–90%, as judged at subsequent digital subtraction angiography (Fig 1).
|
The time to perform surgery and collect blood samples was approximately 4 hours. The anesthetized animals were then transported to the MR imaging suite. Imaging time was approximately 1 hour, and at completion of MR imaging, the animals were killed. Total anesthesia time was approximately 6 hours.
MR Imaging Technique
The animals were imaged with a 1.5-T system (GE Signa CV/i; GE Medical Systems, Milwaukee, Wis), with a maximum gradient strength of 40 mT/m and a maximum gradient slew rate of 150 mT/m/msec. A quadrature receive coil was positioned around the torso of the pig, centered over the kidneys. The body coil was used for radio-frequency transmission, and all images were acquired during suspended ventilation to minimize motion artifacts. For sequences requiring cardiac gating, a plethysmograph was attached to the pig’s tail.
Anatomic information from the renal vessels and kidneys was obtained with multiphasic 20–40-mL gadopentetate dimeglumine–(intravenously administered Magnevist; Berlex Laboratories, Wayne, NJ) enhanced coronal three-dimensional spoiled gradient-echo sequence (15): 3.6/1.0 (repetition time [TR] msec/echo time [TE] msec); flip angle, 25°; receiver bandwidth, ±125 kHz; fractional echo; one signal acquired; field of view, 32 x 32 x 12 cm; in-plane matrix, 512 x 192; section thickness, 2–3 mm, with zero order filling in z, resulting in section spacing of 1.0–1.5 mm; and imaging time, 30 seconds. The imaging volume, prescribed graphically from a transverse localizer, encompassed the aorta and inferior vena cava (IVC) anteriorly and the kidneys posteriorly. Two phases, arterial and venous, were acquired sequentially by using an imaging delay of 8–16 seconds for the first phase. The aorta, infrarenal IVC, and renal vessels were localized from the three-dimensional spoiled gradient-echo sequence, and multiple (at least two) subsequent sets of T1 relaxation and flow measurements were obtained, as described next.
The T1 relaxation sequence was written at our institution for this project (16) (Fig 2). To minimize pulsatility effects, cardiac gating was used, and the image was acquired during suspended ventilation to minimize artifacts related to respiratory motion. The pulse sequence commenced with a nonselective 180° adiabatic inversion pulse triggered by systolic upstroke. After the inversion pulse, a series of four water-selective section-selective (spectral-spatial) 90° radio-frequency pulse/spiral readouts were executed. The first and third readout pulses were synchronized to systolic triggers (the first was immediately after the inversion pulse), and the second and fourth readouts occurred after a user-defined delay from the systolic trigger. This delay time was chosen to result in approximately evenly spaced readout signals and thus depended on the animal’s heart rate, which was usually 70–110 bpm. For example, for a heart rate of 85 bpm, the R-R interval would be 700 msec, and readout signals would be acquired approximately 14, 364, 714, and 1,064 msec after the inversion pulse. Filling of k-space required 20 interleaved spirals per image and 40 cardiac cycles. The system was tuned to the center frequency of water. TR was every other heartbeat; TE, 6.7 msec; receiver bandwidth, ±100 kHz; one signal acquired; field of view, 30 x 30 cm; matrix, 200 x 200; section thickness, 5 mm; imaging time, 25–40 seconds. For spins replaced within the imaging section between 90° excitations (eg, flowing blood), this sequence yielded four measurements of the signal recovery after inversion, from which T1 could be calculated.
|
|
|
|
|
|
|
|
Flow Quantification
Flow characterization and quantitation with cine-phase contrast-enhanced imaging is based on a relationship between measured phase shifts and flow velocity (18,19). The manufacturer’s two-dimensional segmented k-space cine-phase contrast-enhancement technique with view interpolation was used to acquire data for flow measurement from the renal arteries. Images were acquired during suspended ventilation, and imaging parameters were as follows: 8.1/3.7; bandwidth, ± 32 kHz; one signal acquired; field of view, 24–30 cm; matrix, 256 x 192; section thickness, 5 mm; flip angle, 20°; time frames per cardiac cycle, 20–24; views per frame per cardiac cycle, 4–8; velocity-encoding through-plane, 100–200 cm/sec; imaging time, 25–40 sec. At least one and usually two or three cine-phase contrast-enhanced images per vessel were obtained, and imaging planes were orthogonal to the renal artery being studied. The lower end of the velocity-encoding range was initially chosen and increased if aliasing occurred. The left main renal artery was imaged proximal to the surgical stenosis to avoid dephasing accompanying complex flow distal to the stenosis. Flow through the vessel of interest was computed at an off-line workstation (Sun Microsystems) by using previously described techniques (19–23). Single-kidney GFR was calculated as: RPF = RBF x (1 - H) and GFR = RPF x EFGD, where RPF is single-kidney renal plasma flow and H is hematocrit level from a sample of pig blood.
Renal Volumes
To normalize RBF and GFR to kidney size, renal volumes were determined from MR imaging data. By using a workstation (Advantage Windows; GE Medical Systems), kidneys were segmented from a nephrographic phase of the three-dimensional spoiled gradient-echo sequence. Kidneys were manually traced by one author (C.H.C.), and the software automatically summed voxel volume in the segmented object. The rationale for normalizing RBF and GFR by renal volume is to compensate for differences in size among porcine kidneys (24–26).
Normalized RBF and GFR values from kidneys without RAS (right kidneys) were compared with values from kidneys with RAS (left kidneys) by using a paired t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
MR imaging measures of single-kidney GFR normalized to renal volume were significantly decreased in the ischemic kidneys (0.139 mL/min/mL ± 0.067), as compared with normal kidneys (0.340 mL/min/mL ± 0.085) (P < .01, paired t test) (Fig 6).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The general technique of using renal arterial and venous T1 differences to determine the renal extraction of gadopentetate dimeglumine was described by Dumoulin et al (7) and Katzberg et al (11). More recently, Niendorf et al (8,10) were successful in measuring EF accurately in vivo in an animal model. In the present study, we used a T1 measurement sequence that uses breath holding, peripheral cardiac gating, adiabatic inversion, and interleaved spiral readouts. This differs from standard gradient-echo and segmented echo-planar gradient-echo sequences previously described for measuring T1 in flowing blood (8,27). Some of the potential advantages of the T1 measurement sequence used in the current study include the adiabatic inversion pulse to ensure homogeneous 180° inversion of all spins, spectral-spatial excitation pulses to eliminate off-resonance artifacts from fat, and efficient use of magnetic field gradients and excellent flow characteristics inherent to spiral acquisition.
When the infrarenal IVC was used to measure the T1 of renal arterial input, a statistically significant linear correlation existed between EF measured at MR imaging and EF measured by sampling each vessel directly for inulin concentration (r = 0.77). Both MR imaging and inulin techniques showed a statistically significant reduction of EF in kidneys with RAS (left kidneys), as compared with that in kidneys without RAS (right kidneys). It was noted that although there was reasonable linear correlation between EF measured with MR imaging and inulin concentrations, values obtained with MR imaging were approximately 22% (1.00–0.21/0.27) lower (P = .01). This result may be compared with those in swine without RAS in an investigation by Niendorf et al (8), in which a significant difference between MR imaging and inulin measurements of EF was not observed. This prior study differs from the present study in that blood sampling for inulin and MR imaging occurred simultaneously, whereas in the present study, blood sampling for inulin was completed before MR imaging. It is plausible that the additional anesthesia time between blood sampling and MR imaging was accompanied by declining GFR and filtration fraction in the pigs, leading to lower EF determinations observed during MR imaging than during inulin studies. Any progressive decrease in renal function may have particularly affected the acutely ischemic kidneys. To gain further insight into this possibility, subsequent investigations could either randomize the order of MR imaging with inulin measurement or perform them simultaneously if the MR imaging schedule could accommodate this.
MR imaging EF results were found to be highly variable when aortic T1 measurements were used to characterize renal arterial input; thus, only the infrarenal IVC determinations of T1 were used in all preceding determinations of EF. The reasons for unreliable T1 measurements from aortic data are unclear, although a plausible explanation would be insufficient flow between consecutive 90° readout pulses during the spiral T1 sequence. This may have occurred if two consecutive readouts occurred in the same diastole as a result of (a) prominent delay time between the systolic trigger and the first 90° readout pulse or (b) substantial time difference between aortic systole and peripheral capillary systolic pulse, measured with the plethysmograph attached to the animal’s tail. With sequential readout pulses occurring every 350 msec, blood flow in the vessel being studied needed only to be 14.3 mm/sec to achieve complete refreshment of spins in a 5-mm section, but if flow in diastole were insufficient to replace blood in the imaging section, signal loss from longitudinal magnetization suppression would occur. To gain insight into aortic flow, cine-phase contrast-enhanced imaging of two porcine aortas was performed and indeed showed a flow profile close to zero for much of diastole. This could be responsible for errors in T1 measurement in the aorta, as compared with those in the IVC.
With MR imaging measurements, single-kidney GFR (normalized to renal volume) was significantly decreased in ischemic kidneys (0.139 mL/min/mL ± 0.067) versus nonischemic kidneys (0.340 mL/min/mL ± 0.085) (P < .01). This represents a GFR reduction of 59% in kidneys with acutely created high-grade RAS. While the fact that renal function decreased in the ischemic kidney is not surprising, an interesting observation is that 75%–90% diameter stenosis resulted in mean RBF reduction of only 20% (1.00–1.34/1.67) but EF reduction of 50% (0.14/0.28), as measured with MR imaging. A similar pattern of only mildly reduced renal plasma flow in the face of postischemic acute renal allograft failure in patients was recently observed by Corrigan et al (28). The implication of these results is that reduction of ipsilateral kidney GFR accompanying acute renal arterial obstruction may be predominantly governed by decreased hydrostatic filtration pressure rather than by absolute lack of blood flow. Of note, however, is that most renal arterial occlusive disease is chronic and develops slowly, such that different pathophysiology may be present in the setting of chronic RAS. Published data on MR imaging blood flow in stenotic human renal arteries, in fact, indicates that a more substantial reduction in renal flow accompanies RAS than was seen in our acute swine model (29). Further investigation of the relationship between chronic RAS and single-kidney GFR is needed to address this difference.
MR imaging, CT, catheter angiography, ultrasonography, and nuclear medicine have all been used in the evaluation of RAS. While all these techniques have their roles, none would appear to have potential for providing high-spatial-resolution anatomic detail of the renal arteries and parenchyma, combined with accurate measurement of physiologic parameters, including RBF, EF, and single-kidney GFR, which MR imaging might provide in a single examination. Studies (30,31) have shown MR angiography to be greater than 95% accurate in diagnosing RAS. As noted previously, the present and prior studies show considerable promise for renal EF measurement, exploiting arteriovenous T1 differences caused by renal extraction of gadopentetate dimeglumine. While subsequent measurement of GFR requires knowledge of both EF and RBF, a number of studies (21,26,32,33) have shown cine-phase contrast-enhanced MR imaging to be accurate in measuring RBF, allowing calculation of the critical functional parameter, single-kidney GFR, from in vivo MR imaging data. The ability to make such physiologic measurements with MR imaging, in combination with the widespread availability and favorable safety profile of gadopentetate dimeglumine, make the described technique ideal for noninvasive measurement of GFR.
Practical application: Atherosclerotic RAS is a common disease that may lead to hypertension and/or renal insufficiency. Determining the appropriate management of RAS, however, remains controversial. It is difficult to know whether visualized RAS actually causes hypertension or azotemia in a given individual. It has been noted, for example, that revascularization of radiographically detected RAS leads to significant improvement in renal function in only a minority of cases (1,2) and that balloon angioplasty for hypertension was little better than medical treatment alone in one multicenter trial (34). Knowledge of individual kidney GFR, obtained during a more comprehensive MR imaging examination in which both renal anatomic and physiologic information was evaluated, as has been previously advocated by Ros et al (35), would appear to have great potential for noninvasive renal diagnosis. Ultimately, such a comprehensive renal MR imaging examination could prove useful in accurately selecting those patients with RAS who are most likely to benefit from renal revascularization.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, F.G.S., C.H.C.; study concepts, C.H.C., F.G.S., N.J.P., D.M.S.; study design, C.H.C., F.G.S., N.J.P.; literature research, C.H.C., F.G.S.; experimental studies, all authors; data acquisition, C.H.C., J.H.L., K.L.W., D.M.S., N.J.P., S.T.K., B.B.H., D.M.B., G.C.D., A.M.S.G., F.G.S.; data analysis/interpretation, C.H.C., F.G.S., J.H.L., D.M.B., B.D.M.; statistical analysis, C.H.C.; manuscript preparation, C.H.C.; manuscript definition of intellectual content and revision/review, C.H.C., F.G.S.; manuscript editing, C.H.C., F.G.S., N.J.P.; manuscript final version approval, F.G.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Xia, E. Seckin, Y. Xiang, M. Vranesic, W. B. Mathews, K. Hong, D. A. Bluemke, L. O. Lerman, and Z. Szabo Positron-Emission Tomography Imaging of the Angiotensin II Subtype 1 Receptor in Swine Renal Artery Stenosis Hypertension, February 1, 2008; 51(2): 466 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. K. Park, T. K. Rhee, T. A. Cashen, W. Shin, B. E. Schirf, J. A. Gehl, A. C. Larson, P. V. Prasad, D. Li, T. J. Carroll, et al. Renal Artery Stenosis in Swine: Feasibility of MR Assessment of Renal Function during Percutaneous Transluminal Angioplasty Radiology, July 1, 2007; 244(1): 144 - 150. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. S. Lee, H. Rusinek, L. Bokacheva, A. J. Huang, N. Oesingmann, Q. Chen, M. Kaur, K. Prince, T. Song, E. L. Kramer, et al. Renal function measurements from MR renography and a simplified multicompartmental model Am J Physiol Renal Physiol, May 1, 2007; 292(5): F1548 - F1559. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. V. Prasad Functional MRI of the kidney: tools for translational studies of pathophysiology of renal disease Am J Physiol Renal Physiol, May 1, 2006; 290(5): F958 - F974. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Sommer, E. W. Olcott, L. C. Chow, R. R. Saket, and P. Schraedley-Desmond Measurement of Renal Extraction Fraction with Contrast-enhanced CT Radiology, September 1, 2005; 236(3): 1029 - 1033. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. S. Lee, H. Rusinek, M. E. Noz, P. Lee, M. Raghavan, and E. L. Kramer Dynamic Three-dimensional MR Renography for the Measurement of Single Kidney Function: Initial Experience Radiology, April 1, 2003; 227(1): 289 - 294. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
= value observed in kidney without RAS (right kidneys), 
= value observed in kidney with RAS (left kidneys).