DOI: 10.1148/radiol.2461062129
(Radiology 2008;246:241-248.)
© RSNA, 2008
Individual Kidney Blood Flow Measured with Contrast-enhanced First-Pass Perfusion MR Imaging1
Diego R. Martin, MD, PhD,
Puneet Sharma, PhD,
Khalil Salman, MD,
Richard A. Jones, PhD,
J. Damien Grattan-Smith, MD, MBBS,
Hui Mao, PhD,
Thomas C. Lauenstein, MD,
Bobbie K. Burrow, RT,
Dana L. Tudorascu, and
John R. Votaw, PhD
1 From the Department of Radiology, Emory University School of Medicine, Building A, AT-622, 1365 Clifton Rd NE, Atlanta, GA 30322 (D.R.M., P.S., K.S., R.A.J., J.D.G., H.M., T.C.L., B.K.B., D.L.T., J.R.V.); and Department of Radiology, Children's Hospital of Atlanta, Atlanta, Ga (R.A.J., J.D.G.). Received December 14, 2006; revision requested January 15, 2007; revision received March 22; final version accepted May 14.
Address correspondence to D.R.M. (e-mail: dmartin{at}emory.edu).
 |
ABSTRACT
|
|---|
The study design was HIPAA-compliant and approved by the Institutional Review Board, with all participants providing signed informed consent prior to the study. The purpose of this study was to prospectively evaluate the feasibility of determining renal blood flow (RBF) by using a technique based on intravenous administration of gadolinium chelate and evaluation of first-pass gadolinium chelate perfusion by using highly accelerated three-dimensional (3D) gradient-echo magnetic resonance (MR) imaging of the kidney in freely breathing subjects. Flow is determined with Kety-Schmidt formalism by modeling the uptake of gadolinium chelate in the kidney prior to its leaving through the venous system. Validation of the gadolinium chelate perfusion technique is based on comparison of values determined for participants with phase-contrast gradient-echo imaging. The model fit to the measured data is excellent over the first 7–8 seconds of gadolinium chelate uptake and diverges after its appearance in the renal vein. The perfusion data analysis technique showed less than 10% interobserver variation. The average difference between phase-contrast and gadolinium chelate perfusion measurements was 0.08 mL/sec (95% confidence interval: –3.73, 3.58) for left and right kidneys. This study demonstrates feasibility of the gadolinium chelate perfusion method for RBF measurement and discusses potential applications.
Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/246/1/241/DC1
© RSNA, 2008
|
INTRODUCTION
|
|---|
There have been development and evolving validation of gadolinium chelate perfusion magnetic resonance (MR) imaging techniques for the evaluation of glomerular filtration rate, assuming its behavior as a filtered agent without active excretion or uptake from the renal tubules (1–8). It would be important if renal blood flow (RBF) could additionally be extracted from the same gadolinium-based renal perfusion studies. The current standard for measurement of RBF is by using MR imaging on the basis of phase-contrast imaging (9–12). This method requires several additional steps prior to the phase-contrast acquisition in order to achieve and accurately determine renal artery positioning. Further, phase-contrast imaging must be performed separately for each kidney.
Phase-contrast imaging is limited generally to subjects that have normal renal and vascular anatomy with solitary renal arteries; accessory renal arteries require separate measurements or may be too small for accurate phase-contrast assessment. Moreover, highly angled or tortuous renal arteries are technically difficult to assess with phase-contrast imaging owing to the need to place imaging planes orthogonal to the vascular flow vector. A perfusion method for RBF measurement could provide an alternative that may be more robust for certain clinical applications, even in the setting of anomalous vascular anatomy, or unusual renal morphology and positioning, including patients with accessory renal arteries and posttransplant or posttraumatic kidneys.
The purpose of our study was to prospectively evaluate the feasibility of determining RBF by using a technique derived from intravenous administration of gadolinium chelate and evaluation of first-pass perfusion by using highly accelerated three-dimensional (3D) gradient-echo MR imaging of the kidney in freely breathing subjects.
|
MATERIALS AND METHODS
|
|---|
Participants
The study design was Health Insurance Portability and Accountability Act–compliant and approved by our Institutional Review Board, with all participants providing signed informed consent prior to the study. Twenty-nine participants (10 women, 19 men; mean age, 41.5 years) were included in this study. The inclusion criteria were normal renal arterial anatomy and solitary renal arteries free of stenotic lesions, determined by using angiographic imaging acquired at the time of the study. Participants (n = 5) were excluded if accessory renal arteries were identified or if primary renal arterial anatomy could not be clearly determined (see Imaging Techniques).
Nine volunteer participants had no history of renal disease or hypertension. Other participants had renal hypertension (n = 6), ureteropelvic junction obstruction (n = 6), autosomal dominant polycystic renal disease (n = 5), and ureteric obstruction from other causes (n = 2). One additional participant underwent prior unilateral nephrectomy 4 years previously but remained free of metastatic disease. Overall, we separately analyzed 56 kidneys (27 right, 29 left). All subjects were allowed to consume fluids ad libitum prior to arriving at the scanner facilities and were provided an additional 300 mL of water to consume 30 minutes prior to imaging to ensure a well-hydrated status.
Imaging Techniques
Imaging was performed on a 1.5-T imager (Intera; Philips, Best, the Netherlands) with gradient strength of 40 mT/m and slew rate of 200 T/m/sec. Imaging was performed by using a four-element flexible phased-array body-wrap coil.
Gadolinium chelate phantom imaging.—Since the two-compartment model assumes a linear relationship between gadolinium chelate concentration and MR signal intensity (SI), the SI behavior of the renal perfusion imaging sequence was assessed with 17 gadolinium chelate–doped plasma phantoms. Freshly frozen human plasma was obtained from the hematology division at Emory University Hospital (Atlanta, Ga). The phantoms were prepared individually in 50-mL plastic screw-cap sealed tubes with gadolinium chelate (gadopentetate dimeglumine, Magnevist; Berlex Laboratories, Wayne, NJ) concentrations varying from 0 to 30 mmol/L. Within 2 days of preparation, the phantoms were arranged on a styrofoam holder and imaged at room temperature by using a 3D spoiled gradient-echo MR imager with sequence parameters reproducing those used for in vivo renal perfusion imaging (see Single-Kidney RBF-Gadolinium–enhanced Renal Perfusion Imaging).
To simulate the steady-state conditions used in vivo, the number of dynamic images was set to 10, and subsequent SI measurements were performed from the fifth to the 10th dynamic image. Circular regions of interest (ROIs, 60 pixels) were drawn on the short-axis image of each tube (P.S.) and the mean SI was noted. The SI values were normalized relative to the 0-mmol/L gadolinium-chelate–plasma sample, as performed with in vivo imaging described below. The relationship was analyzed for SI–gadolinium-chelate concentration linearity. This study assumes the gadolinium-chelate relaxivity (r1) is not dependent on tissue type (8). Hence, SI-to-gadolinium-chelate concentration linearity is the only criterion that ensured the direct use of MR SI in perfusion kinetic models, such as the one implemented in this investigation.
Single-kidney RBF-phase-contrast imaging.—Multisection abdominal scout imaging was performed in the transverse and coronal planes to include both kidneys by using a steady-state balanced fast field echo technique (repetition time msec/echo time msec, 3.3/1.3; field of view, 340 mm2; matrix, 256; flip angle, 60°; 30 sections; section thickness, 6 mm; section gap, 0 mm; sensitivity encoding factor of two) to localize the renal arteries. Phase-contrast imaging was oriented perpendicular to the vessel of interest, following the localization in two planes.
The imaging technique was a gradient-echo/echo-planar hybrid imaging sequence gated to a pulse oximeter placed on the subject's left index finger. Image parameters were as follows: 24/8.6; field of view, 200 mm2; matrix, 176 (reconstructed to 256); flip angle, 35°; echo-planar imaging factor of seven; section thickness, 6 mm; temporal resolution, approximately 50 msec; 12–16 phases; velocity encoding value, 60 cm/sec; and bandwidth, 98 Hz per pixel. The breath-hold time was typically 20 seconds. The images were transferred off-line for analysis on a dedicated workstation (View Forum; Philips Medical Systems, Best, the Netherlands). ROIs were placed (P.S., K.S.) to encompass the entire renal artery cross section on each phase-contrast image, while manually correcting for motion and pulsatility. Flow was automatically calculated for each kidney from the average velocity multiplied by the area of the ROI.
Single-kidney RBF-gadolinium–enhanced renal perfusion imaging.— In each patient, 0.1 mmol/kg gadolinium-chelate was diluted in 60 mL normal saline and injected at a rate of 0.6 mL/sec by using a power injector (Mallinckrodt, St. Louis, Mo). The lowest possible dose rate was implemented to minimize nonlinear T1 and T2* effects (see "Gadolinium-Chelate Phantoms" in Results). Renal perfusion imaging was performed during the first pass by using a coronal 3D gradient-echo MR imaging technique (THRIVE; Philips Medical Systems, Best, the Netherlands) with fat saturation and centric-radial k-space acquisition, with the following parameters: 3.7/1.7; field of view, 430 mm2; matrix, 96 (reconstructed to 256); flip angle, 30°; sections, 30; section thickness, 2.8 mm; k-lines per segment, 120; and sensitivity encoding factor of three. These parameters resulted in an acquisition time of 0.9 second per dynamic image (Fig 1).
View larger version (87K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1: Shows coronal MR image of dynamic 3D gradient-echo RBF perfusion sequence (3.7/1.7; field of view, 430 mm2; matrix, 96 [reconstructed to 256]; flip angle, 30°; sections, 30; and section thickness, 2.8 mm). ROI tracing for each kidney represents volumetric binary kidney mask created semiautomatically by means of threshold and erosion/dilation steps. Overlay shows final analytic step before mask is propagated through appropriate dynamic series while manually correcting for kidney motion; this process is repeated for each section. Software is configured to sum number of voxels and SI in mask, which can subsequently help determine kidney volume and gadolinium-chelate perfusion (Equation). Similar mask (not shown) was used to help determine aortic SI.
|
|
The imaged volume was oriented to include both kidneys entirely, as determined from scout images, to ensure evaluation of all perfused renal parenchyma. Imaging was performed during normal breathing and commenced 5 seconds after the initiation of the contrast agent administration. This provided a sufficient number of images (more than three) prior to contrast agent arrival to serve as a measure of background unenhanced SI. Relevant data for the single-kidney RBF measurement was recorded during the first 10 seconds after contrast agent arrival in the kidney.
Image Analysis
The dynamic 3D gradient-echo images were transferred off-line for analysis by using software (Analyze 5.0; Mayo Clinic, Cleveland, Ohio) with a previously described method (3). Each total perfused kidney volume (cortex and medulla) was segmented by using a semiautomatic algorithm on the basis of user-defined SI threshold levels (J.R.V., K.S., P.S.), morphologic erosion and/or dilation, and region growing steps. Renal pelvis, pelvic vessels, and adjacent soft tissues were excluded from the renal segments.
Additionally, a dynamic SI mask was created in the descending aorta to serve as the input function. These segmented binary masks of the kidneys and aorta were applied to the images at each time point and adjusted manually for position changes related to respiration. The output from the perfusion masks was a time course of SI in the aorta and each kidney. Summing the number of pixels in the kidney mask and multiplying by the voxel size provided the perfused kidney volume, V.
The mean SI for the renal volumes was calculated for each 0.9-second image. Relative SI values were determined by using the formula (St – S0)/S0, where St is the signal at time t and S0 is the mean unenhanced signal, calculated from the mean of at least three unenhanced images. The relative SI is used to isolate the signal contribution from the contrast agent perfusion and to help eliminate residual signal variations that may result from imperfect coil spatial sensitivity corrections or other field inhomogeneities.
Calculation of Single-Kidney RBF
The Kety-Schmidt (13) integral (derived from Fick's first law of diffusion [14]) states that the rate of uptake of an indicator (in this case, gadolinium chelate) in an organ is equal to the blood flow through the organ multiplied by the arteriovenous difference in the concentration of the indicator. This formalism was used to model the uptake of gadolinium-chelate in the kidneys (calculations performed by J.R.V., K.S., P.S.). In this experiment, data collection was stopped before the indicator appeared in the venous drain from the kidney, so the venous concentration is zero. Hence, the Kety-Schmidt integral reduces to:
where CT is the measured gadolinium-chelate concentration in the whole kidney, F is the blood flow to the kidney, Ca is the measured gadolinium-chelate concentration in the artery supplying the kidney, and td is the difference between times when the gadolinium-chelate is measured in the artery and the kidney. It is recognized that in application, V is the volume of the kidney ROI and CT is the average pixel SI inside this region. This leads to VCT being proportional to the total amount of gadolinium-chelate in the kidney. Hence, when determining the kidney regions, any errors were purposefully made to the too-large side so that the estimation of the total amount gadolinium-chelate in the kidney would not be affected.
On the other hand, the right side of the equation requires the gadolinium-chelate concentration in the artery. Hence, the artery ROI was purposefully made smaller than the edge of the vessel. The integral was calculated by using the trapezoidal rule and the relative aortic SI. F was estimated by adjusting F and td to minimize the least-squares difference between the measured kidney gadolinium-chelate signal, CT(t), and the calculated kidney tissue concentration (right side of Equation). The minimization was performed by using the Powell method (15). The data set spanned the range after the appearance of gadolinium-chelate in the kidneys until it appears in the ureter or vein. The only assumptions necessary for applying this method are that the corrected MR SI (see above) is linearly proportional to the gadolinium-chelate concentration in the aorta and kidney and that the analysis stops before any gadolinium-chelate leaves the kidney.
Statistical Analysis
The RBF data (milliliters per second) from gadolinium perfusion and phase-contrast measurements were analyzed to determine statistical differences. All statistical calculations were performed by using software (SAS, version 9.1.3; SAS Institute, Cary, NC). To compare the RBF measurements with the phase-contrast imaging measurements, a Bland-Altman analysis was performed. The goal of the Bland-Altman analysis was to determine if the RBF measurements agree, on average, with the phase-contrast measurements. The comparison was performed separately for each kidney. A Pearson correlation test was used to determine the relationship between the differences of and the average between the two methods (bias), with a P value of less than .05 indicating a significant difference.
Interobserver variation was evaluated by having four different observers (D.R.M., K.S., P.S., T.C.L.) perform the image postprocessing on all of the subjects with bilateral kidneys present, including signal threshold and segmentation steps. The observers performed this task after demonstrating familiarity with the software, executed all tasks independent of the other study members, and were blinded to the results of others. Final results were expressed as an average of the four individual results, and interobserver variability was measured as an average standard deviation among all subjects.
|
RESULTS
|
|---|
Gadolinium Chelate Phantoms
The MR SI rose linearly until reaching approximately 3.5 mmol/L gadolinium-chelate concentration (approximately 15 MR SI units), after which the curve exhibited nonlinear traits (Fig 2). Beyond this threshold, the amount of gadolinium-chelate cannot be confidently predicted from MR SIs with these sequence parameters. Even though this demonstration was performed in gadolinium-chelate-doped plasma phantoms, we have assumed gadolinium-chelate relaxivity was independent of tissue type, making the linear relationship equivalent in plasma and kidney. This enabled the direct use of relative MR SI in place of gadolinium-chelate concentration in the kinetic model (Equation).
View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2: Graph shows MR SI response of 3D gradient-echo perfusion sequence as function of gadolinium-chelate concentration, as determined with array of gadolinium chelate–doped plasma phantoms. SIs were measured after fifth dynamic scan to ensure system was in steady-state, as expected with in vivo conditions. Note apparent region of linearity between MR SI and gadolinium-chelate concentration below approximately 3.5 mmol/L, allowing simple use of physiologic models, such as two-compartment model, for describing changes in tracer concentrations. Beyond linearity threshold, no predictable relationship exists between SI and gadolinium chelate concentration; kinetic model cannot be used. This phenomenon occurs primarily because of T1 and T2* effects of imaged tissue and on imaging sequence properties.
|
|
Perfusion Image Analysis
Three-dimensional perfusion images were acquired in all individuals without complication (Fig 3a) and data were fit numerically (Equation), with all fits providing a correlation above 0.95 (Fig 3b). The image data diverges from the calculated curve fit after the appearance of contrast agent in the renal vein. The relative MR SI (y-axis) was calculated similarly to the plot shown in Figure 2. Therefore, the measured SI values are expected to be well below the region of SI-to-gadolinium-chelate nonlinearity.
View larger version (48K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a: (a) Shows coronal maximum intensity projection MR series of contrast-enhanced 3D gradient-echo 0.9-second images (3.7/1.7; field of view, 430 mm2; matrix, 96 [reconstructed to 256]; flip angle, 30°; sections, 30; section thickness, 2.8 mm) acquired in freely breathing subject just before arrival of gadolinium chelate in aorta and kidneys (image 1), followed by sample points (images 2–4) within first 7 seconds of its arrival, showing progressive kidney enhancement. (b) Relative MR SI from left kidney from each 0.9-second scan is shown in relation to acquisition time. Points indicated by numbered arrows 1–4 indicate acquisition time and measured renal SI of corresponding image shown in a. Model points are also calculated (Equation). Note that divergence occurs 8 seconds after contrast agent arrival and corresponds to renal vein filling. K = kidney.
|
|
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b: (a) Shows coronal maximum intensity projection MR series of contrast-enhanced 3D gradient-echo 0.9-second images (3.7/1.7; field of view, 430 mm2; matrix, 96 [reconstructed to 256]; flip angle, 30°; sections, 30; section thickness, 2.8 mm) acquired in freely breathing subject just before arrival of gadolinium chelate in aorta and kidneys (image 1), followed by sample points (images 2–4) within first 7 seconds of its arrival, showing progressive kidney enhancement. (b) Relative MR SI from left kidney from each 0.9-second scan is shown in relation to acquisition time. Points indicated by numbered arrows 1–4 indicate acquisition time and measured renal SI of corresponding image shown in a. Model points are also calculated (Equation). Note that divergence occurs 8 seconds after contrast agent arrival and corresponds to renal vein filling. K = kidney.
|
|
Phase-Contrast Perfusion Comparison and Image Postprocessing Reproducibility
The average difference between both measurement techniques for right and left kidneys together (n = 56) was 0.08 mL/sec ± 1.8 (standard deviation). The Bland-Altman analysis shows good average agreement between RBF and phase-contrast measurements. In the left kidney, the mean bias (average of differences between RBF and phase-contrast imaging) was 0.32 mL/sec (95% confidence interval: –0.35, 0.97) for the true mean bias. For the right kidney, the mean bias was –0.18 mL/sec (95% confidence interval: –0.92, 0.56) for true mean bias. Given the 95% confidence intervals, one does not reject the null hypothesis that the true mean bias is zero for either the left or right kidney.
The Bland-Altman analysis was performed individually for the left and right kidney (Fig 4). The limits of agreement, defined as mean bias ± 1.96 multiplied by standard deviation, were determined for the left (–3.1 mL/sec, 3.72 mL/sec) and right kidney (–3.9 mL/sec, 3.56 mL/sec). No discernible patterns between the differences and the mean of the two methods are evident in either the left or right kidney. This dependence was assessed statistically by testing that the correlation between the difference and the mean is zero. We did not reject this null hypothesis for either kidney (P > .05).
For a measure of analysis reproducibility, the average standard deviation of single-kidney RBF measurements among observers was 0.57, which is less than 10% of the average calculated RBF (6.44 mL/sec).
|
DISCUSSION
|
|---|
The measurement of RBF by using first-pass perfusion showed interobserver reproducibility is on average within 10% of the average measured flow, which we consider to be an acceptable degree of precision. Comparison with the phase-contrast measurement of RBF helped show 95% confidence intervals for the left (–3.1 mL/sec, 3.72 mL/sec) and right (–3.9 mL/sec, 3.56 mL/sec) kidneys. Such performance may be adequate for certain clinical applications. Proposed examples include investigations involving left-right perfusion comparisons, such as unilateral collecting system obstruction (16) or longitudinal evaluation to assess progression or improvement of disease, such as before and after angioplasty for renal arterial stenosis, or longitudinal evaluation of posttransplant renal function (17).
An important difference between perfusion and phase-contrast determination of RBF is that phase-contrast imaging measures macrovascular flow at the level of the major arterial trunk with one vessel per image, while perfusion imaging helps to evaluate flow rate at the level of the microvasculature within the tissue parenchyma, acquiring this data from both kidneys simultaneously. This distinction leads to the capacity to measure RBF in terms of flow volume of each kidney, a measure that may yield insights into disease processes that affect the kidney regionally or allow detection of alterations in the number or flow regulation of glomerular filtration units.
The phase-contrast imaging technique has spatial resolution limitations that may be problematic when evaluating small-caliber vessels, including accessory renal arteries, while perfusion imaging is insensitive to this factor. In addition, the phase-contrast technique demands accurate anatomic placement orthogonal to the direction of blood flow to avoid underestimation of flow. This is not always easily achieved; for example, in the transplanted kidney with renal arteries that typically are tortuous.
The first-pass perfusion technique is not dependent on image plane placement and only requires that the entire kidney be included in the image volume. Furthermore, first-pass perfusion has no effective limits on sensitivity range for flow rate, while phase-contrast imaging requires optimally adjusted velocity-encoding parameters to avoid aliasing and underestimation of RBF.
Phase-contrast imaging correlates well when tested against another accepted standard of RBF assessment, para-amino hippuric acid clearance (18,19). It has been noted that comparison between phase-contrast imaging and para-amino hippuric acid clearance has shown correlation within ± 2 mL/sec, with 95% confidence interval that is similar to the correlation between the phase-contrast and first-pass gadolinium-chelate perfusion techniques determined in our study (18,19). However, para-amino hippuric acid clearance is a cumbersome test that is not routinely available for clinical use, making imaging tests (either phase-contrast or first-pass perfusion) attractive alternatives. It is further emphasized that the participants used in this study were selected on the basis of having simple renal vascular anatomy, representing optimal conditions for successful implementation of phase-contrast RBF measurements as a reference standard.
Motion due to respiration was corrected manually in this study by applying ROIs on the segmented images individually for each section at each time. Although this is an additional postprocessing step that is potentially time consuming, applying this technique was necessary for only a small number of times (approximately 10 seconds). We are not aware of commercially available postprocessing software that can deal with renal motion for this application, but this is an area of continued development. A specific test to determine the effect of respiration-induced kidney motion on the RBF perfusion analysis was beyond the scope of this feasibility study. We specifically instructed our subjects to breathe freely in order to emphasize and demonstrate the potential to apply the highly accelerated subsecond time-resolved perfusion technique even in patients who cannot follow breathing instructions, such as sedated patients or young children.
We applied the Kety-Schmidt integral to calculate RBF by assuming inflow concentration of gadolinium-chelate is equal to the concentration in the descending aorta and that there is no outflow of gadolinium-chelate from the kidney. This assumption is true until gadolinium-chelate leaves the kidney via the renal vein or the urinary collecting system, at which point we would have to consider a more sophisticated model. Urinary excretion is relatively slow (occurring in minutes), but venous discharge is rapid (occurring in seconds). Support for application of the Kety-Schmidt integral comes from our data showing a good fit between the predicted model and measured data for approximately the first 10 seconds following gadolinium-chelate arrival at the kidney and that there is no visual evidence from the images that gadolinium-chelate has entered the vein or ureter less than 10 seconds after it enters the kidney.
Estimation of RBF by using the proposed method requires determination of the total amount of gadolinium-chelate in the kidney and its concentration in the artery. By making sure the kidney ROI is at least as large as the perfused portion of the kidney and that the artery ROI is smaller than the artery, the analysis is relatively insensitive to the ROI placement. Interobserver variability of less than 10% suggests that ROI definition is not a major source of uncertainty.
Another assumption is that the SI changes linearly with changes in gadolinium-chelate concentration in the perfusion images. This allows a predictable conversion between the measured parameter (MR SI) and the unknown parameter (gadolinium chelate concentration). The primary objective for introducing the gadolinium chelate–doped phantom component of the study is to evaluate the relationship between gadolinium concentration and SI for the imaging technique used for renal perfusion imaging. After establishing that there is a useful range of contrast agent concentration that fulfills the linearity requirement between that concentration and SI, proportionality of SI measurements can be used in place of direct measurements of those concentrations. The constant of proportionality is dependent on the gadolinium chelate relaxivity (r1) and the specific MR sequence parameters.
Since we have assumed that relaxivity is independent of tissue type (8), determining the precise value of this scaling constant is not necessary because this value conveniently cancels itself during the analysis (Equation). An alternative method of determining gadolinium chelate concentration is to measure T1 at each time. Though possible (20–32), one would have to sacrifice temporal resolution for accurate T1 measurements and determine the tissue-specific relaxivity value for the contrast agent. Demonstrating MR SI-gadolinium-chelate concentration linearity bypasses the T1-determination step, with the trade-off that this linearity is sequence and hardware dependent.
The limitations of the perfusion technique presented here include factors that result in diminished signal-to-noise ratio including low gadolinium chelate dose administration rate and high 3D gradient-echo acquisition rate. Low gadolinium chelate dose administration rate is used to minimize potential saturation effects at higher concentrations of gadolinium-chelate (Appendix E1 http://radiology.rsnajnls.org/cgi/content/full/246/1/241/DC1). The necessarily high image acquisition rate will further limit resolution and signal-to-noise ratio per image pixel. However, we are measuring a macroscopic parameter, total RBF, and only need to accurately determine the total amount of gadolinium-chelate in the kidney and its average concentration in the artery. Hence, the trade-off between signal-to-noise ratio and temporal resolution for this application is biased toward the temporal resolution side. This argument holds further when considering the balance between spatial and temporal resolution.
Furthermore, although it may appear that the images are signal-to-noise ratio–challenged, the observation that we are able to achieve high interobserver analysis precision helps to support our conclusion that the images are of adequate quality and are interpretable for the analysis methods employed in this study. Image artifacts may occur related to use of sensitivity encoding acceleration and careful attention is required to ensure that the patient is fully included within the imaging field of view. If necessary, the patient's arms may be placed over their head.
In summary, we have evaluated the feasibility of quantifying RBF by using high-frequency time-resolved 3D gradient-echo MR imaging during first-pass gadolinium chelate perfusion in the kidney. By using a simple kinetic model, measurements correlated well with phase-contrast imaging. This technique allows quantitative assessment of the perfusion to each kidney in terms of blood volume per unit of time per unit of renal volume.
|
ADVANCES IN KNOWLEDGE
|
|---|
- It is feasible to use intravenously administered gadolinium-chelate first-pass perfusion for estimating single-kidney blood flow.
- The method presented here provides simultaneous assessment of the renal parenchymal volume and allows calculation of the perfusion to each kidney in terms of blood volume per unit of time per unit of renal volume.
|
IMPLICATIONS FOR PATIENT CARE
|
|---|
- The accelerated MR three-dimensional volumetric imaging technique used in this study facilitates imaging of freely breathing patients, such as is required for the imaging of children.
- This study's method for single-kidney blood flow evaluation has potential to supplement routine MR imaging to provide a fusion of structural and functional analyses of renal disease from a single examination.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank John Carew, PhD, for his contributions to the statistical analysis of our data.
|
FOOTNOTES
|
|---|
Abbreviations: RBF = renal blood flow • ROI = region of interest • SI = signal intensity • 3D = three-dimensional
Guarantors of integrity of entire study, D.R.M., P.S., K.S., H.M., T.C.L.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, D.R.M., K.S., R.A.J., J.D.G., H.M., T.C.L.; clinical studies, D.R.M., P.S., K.S., T.C.L., B.K.B.; experimental studies, D.R.M., P.S., K.S., R.A.J., H.M., T.C.L., B.K.B., J.R.V.; statistical analysis, D.R.M., T.C.L., D.L.T.; and manuscript editing, D.R.M., P.S., J.D.G., H.M., T.C.L., D.L.T., J.R.V.
Authors stated no financial relationship to disclose.
|
References
|
|---|
- Holechek MJ. Glomerular filtration: an overview. Nephrol Nurs J 2003;30:285–290; quiz 291–292. [Medline]
- McDaniel BB, Jones RA, Scherz H, Kirsch AJ, Little SB, Grattan-Smith JD. Dynamic contrast-enhanced MR urography in the evaluation of pediatric hydronephrosis. II. Anatomic and functional assessment of uteropelvic junction obstruction. AJR Am J Roentgenol 2005;185:1608–1614.
- Jones RA, Easley K, Little SB, Scherz H, Kirsch AJ, Grattan-Smith JD. Dynamic contrast-enhanced MR urography in the evaluation of pediatric hydronephrosis. I. Functional assessment. AJR Am J Roentgenol 2005;185:1598–1607.
- Huang AJ, Lee VS, Rusinek H. Functional renal MR imaging. Magn Reson Imaging Clin N Am 2004;12:469–486. [CrossRef][Medline]
- Lee VS, Rusinek H, Noz ME, Lee P, Raghavan M, Kramer EL. Dynamic three-dimensional MR renography for the measurement of single kidney function: initial experience. Radiology 2003;227:289–294. [Abstract/Free Full Text]
- Niendorf ER, Grist TM, Frayne R, Brazy PC, Santyr GE. Rapid measurement of Gd-DTPA extraction fraction in a dialysis system using echo-planar imaging. Med Phys 1997;24:1907–1913. [CrossRef][Medline]
- 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 plot technique. J Magn Reson Imaging 2003;18:714–725. [CrossRef][Medline]
- Hackstein N, Kooijman H, Tomaselli S, Rau WS. Glomerular filtration rate measured using the Patlak plot technique and contrast-enhanced dynamic MRI with different amounts of gadolinium-DTPA. J Magn Reson Imaging 2005;22:406–414. [CrossRef][Medline]
- Park JB, Santos JM, Hargreaves BA, et al. Rapid measurement of renal artery blood flow with ungated spiral phase-contrast MRI. J Magn Reson Imaging 2005;21:590–595. [CrossRef][Medline]
- Bax L, Bakker CJ, Klein WM, Blanken N, Beutler JJ, Mali WP. Renal blood flow measurements with use of phase-contrast magnetic resonance imaging: normal values and reproducibility. J Vasc Interv Radiol 2005;16:807–814. [Medline]
- Thompson RB, McVeigh ER. Flow-gated phase-contrast MRI using radial acquisitions. Magn Reson Med 2004;52:598–604. [CrossRef][Medline]
- de Haan MW, van Engelshoven JM, Houben AJ, et al. Phase-contrast magnetic resonance flow quantification in renal arteries: comparison with 133Xenon washout measurements. Hypertension 2003;41:114–118. [Abstract/Free Full Text]
- Kety SS, Schmidt CE. The nitrous oxide method for the quantitative determination of cerebral blood flow in man: theory, procedure and normal values. J Clin Invest 1948;27:476–483. [Medline]
- Fick A. Ueber diffusion. Ann Physik 1855;170:59–86.
- Press W, Flannery B, Teukolsky S, Vetterling W. Numerical recipes in C: the art of scientific computing. New York, NY: Cambridge University Press, 1988.
- Mustonen S, Ala-Houhala IO, Vehkalahti P, Laippala P, Tammela TL. Kidney ultrasound and Doppler ultrasound findings during and after acute urinary retention. Eur J Ultrasound 2001;12:189–196. [CrossRef][Medline]
- Ardalan MR, Tarzamni MK. Renal allograft hemodynamic and diameter changes after living donor transplantation. Transplant Proc 2006;38:388–389. [CrossRef][Medline]
- Myers BD, Sommer FG, Li K, et al. Determination of blood flow to the transplanted kidney: a novel application of phase-contrast, cine magnetic resonance imaging. Transplantation 1994;57:1445–1450.
- Wolf RL, King BF, Torres VE, Wilson DM, Ehman RL. Measurement of normal renal artery blood flow: cine phase-contrast MR imaging vs clearance of p-aminohippurate. AJR Am J Roentgenol 1993;161:995–1002. [Abstract/Free Full Text]
- Buonocore MH, Katzberg RW. Estimation of extraction fraction (EF) and glomerular filtration rate (GFR) using MRI: considerations derived from a new Gd-chelate biodistribution model simulation. IEEE Trans Med Imaging 2005;24:651–666. [CrossRef][Medline]
- O'Byrne E, Pellas T, Laurent D. Qualitative and quantitative in vivo assessment of articular cartilage using magnetic resonance imaging. Novartis Found Symp 2003;249:190–198. [Medline]
- Lyng H, Dahle GA, Kaalhus O, Skretting A, Rofstad EK. Measurement of perfusion rate in human melanoma xenografts by contrast-enhanced magnetic resonance imaging. Magn Reson Med 1998;40:89–98. [CrossRef][Medline]
- Ibrahim MA, Emerson JF, Cotman CW. Magnetic resonance imaging relaxation times and gadolinium-DTPA relaxivity values in human cerebrospinal fluid. Invest Radiol 1998;33:153–162. [CrossRef][Medline]
- Morkenborg J, Taagehoj JF, Vaever PN, Frokiaer J, Djurhuus JC, Stodkilde-Jorgensen H. In vivo measurement of T1 and T2 relaxivity in the kidney cortex of the pig: based on a two-compartment steady-state model. Magn Reson Imaging 1998;16:933–942. [CrossRef][Medline]
- Tofts PS. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson Imaging 1997;7:91–101. [Medline]
- Brookes JA, Redpath TW, Gilbert FJ, Needham G, Murray AD. Measurement of spin-lattice relaxation times with FLASH for dynamic MRI of the breast. Br J Radiol 1996;69:206–214. [Abstract/Free Full Text]
- Donahue KM, Weisskoff RM, Chesler DA, et al. Improving MR quantification of regional blood volume with intravascular T1 contrast agents: accuracy, precision, and water exchange. Magn Reson Med 1996;36:858–867. [Medline]
- Patel PP, Koppenhafer SL, Scholz TD. Measurement of kinetic perfusion parameters of gadoteridol in intact myocardium: effects of ischemia/reperfusion and coronary vasodilation. Magn Reson Imaging 1995;13:799–806. [CrossRef][Medline]
- Donahue KM, Weisskoff RM, Parmelee DJ, et al. Dynamic Gd-DTPA enhanced MRI measurement of tissue cell volume fraction. Magn Reson Med 1995;34:423–432. [Medline]
- Nelson KL, Runge VM. Basic principles of MR contrast. Top Magn Reson Imaging 1995;7:124–136. [Medline]
- 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]
- Niendorf ER, Grist TM, 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]
- Kuperman VY, Karczmar GS, Blomley MJ, Lewis MZ, Lubich LM, Lipton MJ. Differentiating between T1 and T2* changes caused by gadopentetate dimeglumine in the kidney by using a double-echo dynamic MR imaging sequence. J Magn Reson Imaging 1996;6:764–768. [CrossRef][Medline]
TOP
ABSTRACT
INTRODUCTION
