Radiology Imaging of Renal Structure and Function by Computed Tomography, Magnetic Resonance Imaging, and Ultrasound

Radiology Imaging of Renal Structure and Function by Computed Tomography, Magnetic Resonance Imaging, and Ultrasound Nicolas Grenier, MD,* Emilio Quaia, MD,† Pottumarthi V. Prasad, PhD,‡ and Laurent Juillard, MD, PhD§ Radiological techniques are now able to provide morphologic, functional, and structural information relative to kidney diseases. Many of these approaches have been proposed experimentally, but validation studies in patients still remain mandatory. Contrast-enhanced ultrasound allows for the measurement of perfusion parameters. Multidetector computed tomography and magnetic resonance imaging make it possible to measure the differential function of filtration. Measurement of absolute glomerular filtration rate is still under development. Finally, magnetic resonance imaging is also able to provide information on the level of intrarenal oxygenation by the use of blood oxygenation level-dependent sequences and on cell density and water exchanges by the use of diffusion-sensitive acquisitions. Semin Nucl Med 41:45-60 © 2011 Elsevier Inc. All rights reserved.

N

uclear medicine remains the reference for the evaluation of quantitative functional parameters. However, radiological techniques are now able to compete in some areas of kidney function evaluation. They have the advantage of providing morphologic and functional data in the same examination as well as structural information related to kidney diseases. Ultrasound (US) with intravascular microbubbles provides access to perfusion parameters. Multidetector computed tomography (CT) and magnetic resonance imaging (MRI) make possible to measure relative or absolute function of filtration, with each having different advantages and drawbacks. However, validation studies in patients still remain mandatory. Finally, MRI, because of flexibility, is also able to provide information on the level of intrarenal oxygenation using blood oxygenation level-dependent (BOLD) sequences and on cell density and water exchanges by the use

*Service d’Imagerie Diagnostique et Interventionnelle de l’Adulte, Groupe Hospitalier Pellegrin, Bordeaux-Cedex–France. †Department of Radiology, Cattinara Hospital, University of Trieste, Trieste, Italy. ‡Radiology Department/Center for Advanced Imaging, Evanston Hospital, Evanston, IL. §Département de Néphrologie, Hôpital Edouard Herriot, Lyon, France. Address reprint requests to Nicolas Grenier, Service d’Imagerie Diagnostique et Interventionnelle de l’Adulte, Groupe Hospitalier Pellegrin, Place Amélie Raba-Léon, 33076 Bordeaux-Cedex-France. E-mail: nicolas. [email protected]

0001-2998/11/$-see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1053/j.semnuclmed.2010.09.001

of diffusion-sensitive acquisitions. This review will cover the potentials and limitations of these techniques.

Assessment of Renal Perfusion by Contrast-Enhanced US Color and power Doppler are limited by the low sensitivity to low-velocity flow in smaller vessels (⬍2 mm in diameter). Contrast-enhanced ultrasound (CEUS) with microbubble contrast agents1-3 has recently been proposed as a new imaging modality to quantify tissue perfusion.4,5 CEUS presents several advantages, including inexpensiveness, portability, availability, lack of restrictions in performing serial examinations at short intervals, and absence of exposure to radiation or nuclear tracers. Microbubble contrast agents for US have a diameter from 2 to 6 ␮m. The microbubble shell may be stiff (eg, denaturated albumin) or flexible (phospholipids) and has a thickness of 10 to 200 nm. New-generation microbubbles are filled with a high-molecular-weight gas (eg, perfluorocarbon or sulfur hexafluoride) with low solubility in the bloodstream. Microbubbles have a pure intravascular distribution, even though some agents present a postvascular hepato- and/or spleno-specific phase from 2 to 5 minutes after intravenous injection.6,7 The microbubble gas content is exhaled via the lungs from 10 to 15 minutes after injection whereas the components of the shell are metabolized or filtered by the kidney and eliminated by the liver. Adverse 45

N. Grenier et al

46 reactions in humans are rare, usually transient, and of mild intensity.8-13 Hypotensive reactions have been observed after microbubble injection, and some deaths were reported in cardiac patients. At resonant frequency (fo), the microbubble radial oscillation becomes efficient and exaggerated, and the scattering cross section of a microbubble is no longer simply dependent on microbubble size; it can reach peak values a thousand times greater compared with values at off-resonance. The insonation power is usually expressed by the mechanical index defined as p–/公fc, where p– is the largest peak negative pressure and fc is the center frequency of the pulse. When acoustic pressures at or near the resonant frequency are sufficiently high, nonlinear microbubble oscillation develops, producing harmonic frequencies. These frequencies allow us to distinguish microbubble signal from tissue clutter by using specialized contrast-specific US techniques. Pulse inversion is the best-known phase modulation technique. Vascular recognition imaging combines Doppler information with phase analysis and involves the transmission of 4, alternately inverted, pulses along each scan line. Cadence contrast pulse sequencing works by interrogating each scan line several times with pulses having various amplitudes and phases. Both harmonic and nonlinear fundamental signals from microbubbles are represented in a grayscale or color map suppressing the linear fundamental echoes from stationary tissues. The intravenous injection of microbubble-based contrast agents allows the estimation of relative blood flow and fractional blood volume of the microvasculature in a region-ofinterest (ROI). Determination of the degree of tissue contrast enhancement relies on the accurate distinction between microbubble backscatter signals and stationary tissue background. The fundamental assumption is the linear relation between video-intensity and microbubble concentration up to the achievement of a plateau phase.14 After the achievement of the plateau phase, the concentration of microbubbles increases whereas videointensity remains constant, and at even greater concentrations, the video intensity actually decreases because of attenuation of the US beam by the microbubbles themselves. Quantitative analysis of tissue perfusion using CEUS is still limited by acoustic shadowing because of the inadequate compensation for microbubble attenuation, and tissue attenuation correction algorithms or mathematical models estimating microbubble attenuation have been proposed.15,16 Backscatter signals from microbubbles are processed into pixels of brightness for the video presentation in the US system to express echo-power values that reflect instantaneous in situ concentration of microbubble contrast agents. Logarithmic compression is used for displaying signals with a large dynamic range on US monitors. Time-intensity curves may be calculated by positioning a manually defined or automatically copied ROI over a parenchymal region and by correlating the measured video intensity (0-255 grayscale levels) with time in seconds.15,16 On the basis of the mathematics of logarithms the mean video intensity in a ROI is as follows:

mean VI ⫽

1 N

⫻

N

10 log 10 兺 j⫽1

冉 冊 Ij

Iref

Where ⌸ indicates the product of the intensity of all the pixels in the ROI, N is the pixel number within the ROI, Ij is the acoustic intensity, and Iref is an intensity reference level determined through equipment gain.17 The direct visual assessment of the degree of log-compressed video-intensity is the least accurate method because the background tissue signal varies widely within the US sector as the result of inhomogeneities of acoustic power and to differences in the attenuation and absorption of US energy by tissue. A practical approach consists of collecting scanconverted video data, log-compressed and palletized as grayscale or color-coded 8-bit data, in the form of DICOM (ie, Digital Imaging and Communications in Medicine) files. Here, however, proper linearization needs to be applied before curve-fitting and analysis to reverse the effects of logcompression and possibly nonlinear palette rendering. The quantification of echo-signal intensity after antilogarithmic transformation is the most accurate method and eliminates the influence of logarithmic compression, color maps, postprocessing curves, and techniques for edge enhancement on the input signal mapping for video presentation. Software packages access the raw radio frequency data before application of nonlinear modifications and allow antilogging of the grayscale signals, as well as image alignment, signal averaging, and background subtraction. Parametric images may be obtained after background subtraction in a pixel-by-pixel evaluation from the analysis of harmonic grayscale imaging data through the use of dedicated software packages for the automated color-coded depiction of the different kinetic parameters.18,19 The achievement of a steady-state microbubble concentration in the peripheral circulation is essential in tissue perfusion studies.20 When microbubbles are administered as a constant infusion, the steady state is achieved after 2 to 3 minutes. This is obtained by dedicated microbubble injectors usually equipped with a rotating syringe to avoid microbubble sedimentation. At steady state, the inflow and outflow of microbubbles in any microcirculatory unit is constant, proportional to the fractional blood volume of that unit, and dependent only on the flow rate of microbubbles. Local tissue perfusion may be calculated by analyzing the replenishment kinetics of the volume of microbubbles after their destruction by initial high-transmit-power insonation.21,22 After microbubble destruction the system is switched to a low transmit power so the refill rate of microbubbles returning to the imaged volume can be monitored (Fig. 1). One of the main limitations of the technique is that perfusion data are acquired from a single tissue plane, a situation unlikely to accurately reflect global perfusion of the tissue or organ under consideration. The 3-dimensional volumetric hemodynamics of the reperfusion are complex and have not yet been fully modeled. The opposite approach, known as diminution kinetics,23 can also be implemented, ie, by observing the rate of decay of microbubbles exposed to a high-intensity beam, but this method cannot be performed in real time.

Radiology imaging of renal structure and function by CT, MRI, and US

47

Figure 1 Progressive refilling of renal parenchyma at low acoustic power insonation after destruction of microbubbles. After image alignment and background subtraction, the video-intensity in grayscale may be quantified by manually defined ROIs positioned in a region of renal parenchyma. A preliminary high acoustic power train of US pulses (A) destroys all the microbubbles comprised in the imaged slice of renal parenchyma. The progressive refilling of renal parenchyma by microbubbles is monitored at low acoustic power (B-D). The progressive greater echo-signal videointensity of renal parenchyma is determined by the progressive accumulation of microbubbles in the imaged slice. (Color version of figure is available online).

The analysis of the refilling process (Fig. 2) provides semiquantitative data related to tissue perfusion values.20,24,25 These include time to peak intensity, slope of the first ascending tract (wash-in) of the curve (␤ ⬃ blood flow velocity), the maximum amplitude of the refilling curve (A ⬃fractional blood volume), the area under the curve (⬃blood volume), and the mean transit time. According to the central volume theorem, tissue perfusion may be calculated as the ratio of the fractional vascular volume to the mean transit time.26 The prod-

uct of A (cm3/g of tissue) ⫻ ␤ (s⫺1), corresponding to the area under the curve, relates with the perfusion. Unfortunately, this cannot be converted to true perfusion because the volume of the perfused tissue is not known; the beam thickness and shape are complex and dependent on machine settings (eg, the focus position) and patient variables (eg, attenuation). Different mathematical models have been proposed to fit the refilling kinetics.27-29 In the first model, the exponential behavior is a consequence of microbubble diffusion and the refill curve has a rising exponential form described by the equation: Signal intensity ⫽ A(1 ⫺ e⫺␤t).

Figure 2 Quantification of renal perfusion through CEUS after the injection of sulfur hexafluoride-filled microbubbles and at low acoustic power insonation. The difference in the refilling kinetic of renal parenchyma is caused by renal artery stenosis. The normally perfused kidney (arrows) presents a higher slope of the first ascending tract of the curve and a greater value of grayscale intensity at the plateau phase in comparison with the kidney, which presents a tight renal artery stenosis (black dots).

Unfortunately, the exponential function does not take into account the fluid dynamics, and it fails to predict the experimental results if the percentage of microbubble destruction in the vessels feeding the ROI is not null. Further limitations of this model include the assumption of a constant concentration of microbubbles entering the ROI immediately after the destruction pulse and the neglect of the different directions of the vessels inside the examined ROI. Potdevin et al21,22 developed a dual model with a crossplane and an in-plane vascularity for both the renal medullary and cortical regions, in which the whole refilling process is a weighted average of all possible elemental refill curves. Krix et al27 proposed a multivessel model that neglects diffusion and does not present exponential features but assumes and takes into account a particular geometry of the vessels in

N. Grenier et al

48

modalities, contrast in MRI can be modified in a variety of ways, leading to efficient tissue characterization paradigms. MRI has a further advantage in that the methodology is equally applicable to small-animal models and hence allows for translation of results from preclinical to human applications. Here, we will review 2 endogenous contrast mechanisms unique to MRI as applied to the kidney that would be of interest to the nephrourology community. The 2 techniques discussed will be diffusion and BOLD MRI.

Diffusion MRI Figure 3 Renal parenchyma perfusion defect (arrow) after the injection of microbubbles in a New Zealand white rabbit. In all animals, the abdominal aorta and both kidneys were surgically exposed after midline laparotomy. The blood flow was stopped by manual compression of the subrenal abdominal aorta for approximately 5 seconds while 2 mL of polyvinyl alcohol embolizing particles (150-250 ␮m in diameter) was injected directly into the suprarenal abdominal aorta, below the level of the superior mesenteric artery and coeliac trunk, through a 22-gauge needle. This procedure caused random occlusion of one or more renal segmental arteries, with the production of focal ischemic renal perfusion defects.

the ROI. In all these mathematical models the time-intensity curves frequently present a wide data dispersion during both the ascending and the second plateau phase of the curve.30 Lucidarme et al31 proposed a model described by a sigmoid function that is based on the assumption that microbubble destruction actually occurs in the feeding vessels which reach the ROI. According to a further model proposed by Quaia et al,32 the refilling kinetics depends on the distribution of vessel transit times and flows in the kidney, resulting in a piecewise linear function in which the transit times are the times that separate the linear tracts, and the slopes are directly related to the flows. CEUS quantitation of tissue perfusion has been applied in kidneys28,32,33 and renal transplants.34-36 In patients with renal artery stenosis (Fig. 2) or other vascular abnormalities, the refilling curve revealed an increased time-to-peak and reduced slope of the wash-in tract and maximum amplitude.32,34-36 Animal models were initially used to assess the capabilities of CEUS in the detection of renal perfusion defects,37-41 which appear as single or multiple focal wedge-shaped areas of absent, diminished, or delayed contrast enhancement in comparison with the adjacent renal parenchyma.42,43 The identification of small renal perfusion defects in the subcapsular renal region is penalized by the limited spatial resolution of CEUS, which cannot identify renal perfusion defects smaller than 5 mm (Fig. 3)

Bold and Diffusion Magnetic Resonance Imaging Magnetic resonance imaging (MRI) is already well established as a diagnostic imaging modality providing exquisite softtissue contrast and anatomic detail. Tremendous advances have been made during the last 2 decades in improved image quality and spatial coverage. Unlike other diagnostic imaging

Nuclear magnetic resonance (NMR) is the most noninvasive method to estimate self-diffusion coefficient and can be useful in characterizing tissue.44,45 Diffusion MRI is well developed and used in routine clinical practice in the brain mainly for early detection of ischemic stroke lesions. Although application to human kidneys has been shown to be feasible for more than a decade,46,47 only recently are studies in which the authors focus on specific clinical indications beginning to emerge.48 Diffusion MRI sequences involve strong motion sensitizing gradient pulses to be sensitive to Brownian motion of water molecules.48 By obtaining measurements with different diffusion sensitivities (b values), one can estimate the so-called apparent diffusion coefficient (ADC). The word “apparent” is used because the actual (real) diffusion coefficient is almost impossible to measure under in vivo conditions because of the strong influence of macroscopic movement of the tissue, for example, breathing, blood pulsation, blood flow, and peristaltic movements. ADC measurement in structured tissue would inherently depend on the microscopic structural makeup and hence could be a very sensitive indicator of change.48 Because diffusion-weighted images are highly sensitive to motion, any type of voluntary and involuntary motion in the subject could adversely affect the measurements. Use of single-shot imaging in combination with breath-holding is a good option but is not very practical especially in patients. Measurements during free breathing by the use of multiple averages appears to be a more practical approach for routine clinical use.49 Although Brownian motion is truly random directional motion in fluids, when examining water within structured matter, such as fibers there may be significant difference in the ADC measurements depending on the direction. This difference has led to anisotropic diffusion measurements50 in tissue and applications of fiber tracking.51 Anisotropic diffusion can be characterized by the term fractional anisotropy, which describes the degree of anisoptropy, ie, zero in free fluid and 1 in limiting cases of free diffusion only in one direction and none in the others. On the basis of diffusion tensor imaging, the directional vectors can be derived and can be used to map fiber tracks.51 In the kidneys, even though there are no fibers, the medulla shows anisotropy mainly as the result of the parallel orientation of the tubules and vasa recta.47,48 This leads to a radial orientation of diffusion within the medulla in human kidneys, for example,

Radiology imaging of renal structure and function by CT, MRI, and US

49

Intrarenal Oxygenation as Evaluated by BOLD MRI

Figure 4 Apparent diffusion coefficient values for single kidney at different chronic renal failure stages. Reproduced with the permission from Springer Science and Business Media.

Applications of Diffusion MRI in the Kidney In the brain, the most interesting application of diffusion MRI is to detect early changes after ischemic stroke.52 In other organs, a more compelling application is the changes in tissue stiffness associated with development of fibrosis. In the kidney, there is evidence that variations based on clinical grade of fibrosis exist with diffusion measurements (Fig. 4).53 The key advantage of noninvasive MRI-based assessment is the potential for avoiding biopsy, which remains the gold standard for the evaluation of fibrosis. Apart from being invasive, biopsy-based assessments are limited by sampling errors and potential heterogeneity in spatial distributions.54 Figure 5 illustrates that renal medulla exhibits anisotropy and the contrast on the fractional anisotropy maps is much higher in differentiating cortex and the medulla compared with ADC maps. The directionality is consistent with the known radial orientation of renal tubules and vasa recta within the medulla. These measurements are now being used to evaluate differences in chronic renal diseases, such as diabetic nephropathy.55

Scientific and clinical communities are increasingly recognizing the significance of renal oxygenation status.56-58 In most organs, regional oxygen tension (PO2) closely follows the level of regional blood flow because oxygen consumption is relatively constant. However, this is not true in the kidney, where active tubular reabsorption demands more oxygen consumption whenever filtration and blood flow increase together.59 Over a wide range of normal blood flows the renal arteriovenous oxygen difference is remarkably constant. For the purposes of function and oxygen supply, the mammalian kidney can be considered made of 2 separate organs, cortex and medulla.59 Although the flow of blood to the renal cortex normally supplies oxygen in greater quantities than its metabolic needs, blood flow to the renal medulla is relatively low. In addition, oxygen diffuses from the arterial to venous vasa recta, and the process of generating an osmotic difference by active reabsorption of sodium requires large amount of oxygen. In combination, the kidney especially the renal medulla functions in a hypoxic mileu. This medullary hypoxia has consequences to renal physiology and pathophysiology,56 and hence evaluation of in vivo renal oxygenation is important. BOLD MRI has been used extensively in organs, such as the brain.60-62 The BOLD MRI technique exploits the fact that the magnetic properties of hemoglobin vary, depending on whether it is in the oxygenated or deoxygenated form. This affects the T2* relaxation time of the neighboring water molecules and in turn influences the MRI signal on T2*-weighted images. The rate of spin dephasing R2* (⫽1/T2*) is closely related to the tissue content of deoxyhemoglobin. Because the oxygen tension (PO2) of capillary blood is thought to be in equilibrium with the surrounding tissue, changes estimated by BOLD MRI can be interpreted as changes in tissue PO2.63-65 A strong correspondence has been demonstrated between renal BOLD MRI measurements in humans64,65 and rodents,66 with earlier animal data obtained by the use of inva-

Figure 5 Diffusion tensor imaging in a healthy human kidney at 1.5 T. The b0 image shows normal anatomical corticomedullary differentiation. The MD map shows slightly lower ADC in the medulla than in the surrounding cortex. Dramatic contrast between cortex and medulla (circles) appears in the FA map, where the anisotropic microstructure of the medullary tubules induces high FA values. b0: T2-weighted reference image; MD, mean diffusivity (ADC map); FA, fractional anisotropy map. Image provided by Prof Chris Flask.

N. Grenier et al

50

Figure 6 Top panel shows the anatomic image of the kidneys ; middle is the R2* map at baseline; and bottom is the R2* map after administration of furosemide 20 mg intravenously. The R2* maps are shown in pseudocolor along with a color bar showing the range of R2* values (2-50 s⫺1) represented. Greater R2* values imply greater levels of hypoxia or lower levels of blood and hence tissue oxygenation.65 Note that the medulla (M) turns more blue in the postfurosemide R2* map, implying an improved oxygenation level approaching that of the cortex (C).

sive microelectrodes.67 This was done by demonstrating changes after the administration of furosemide and comparing them with those after acetazolamide. Furosemide, a loop diuretic, reduces the reabsorptive work in the thick ascending limbs and hence improves medullary oxygenation (Fig. 6), whereas acetazolamide results in similar diuretic effect but acts on the cortical portion of the nephron and shows minimal change in the medullary oxygenation. Applications of BOLD MRI to the Kidney A significant portion of the work to date has been determined by the hypothesis that the compromise of endogenous protective molecular mechanisms fail to maintain medullary oxygenation status and lead to manifestation of disease and/or disease progression. It has been demonstrated use of the BOLD MRI technique that age and diabetes (both recognized as predisposing factors for acute renal failure) are associated with reduced prostaglandin production, a hypothesized protective mechanism.66,68 This was done by demonstrating differences in response to waterload between different age groups and between diabetic and age-matched control patients. Further it was demonstrated that young subjects who showed a significant improvement in renal medullary PO2 failed to do so when pretreated with ibuprofen69 or

naproxen,70 nonselective prostaglandin inhibitors. Recently, it was also demonstrated that effects of prostaglandin inhibition may be compensated by addition of a nitric oxide donor,71 a concept that has lead to new class of pain medication cyclooxygenase-inhibiting nitric oxide donators.72 It was also demonstrated by the use of BOLD MRI that there is a lack of (or reduced) nitric oxide bioavailability in kidneys of hypertensive rats73 and how that may be reversed by suitable pharmacologic interventions, such as superoxide scavengers.74 This was done by demonstrating that the baseline R2* values in the renal medulla of spontaneously hypertensive rat were significantly high (ie, low oxygenation) compared with those in Wistar Kyoto strain and that they do not respond to nitric oxide synthase inhibition by L-NAME. Similarly, it was also shown that renal medullary R2* drops (ie, oxygenation improves) significantly following administration of tempol in spontaneously hypertensive rat but not in Wistar Kyoto. Of significance is that these measurements can be extended to humans.75 Other applications of BOLD MRI measurements include the evaluation of renal artery stenosis76 and diabetes.77 In a carefully performed large animal study, Juillard et al76 found oxygenation in both cortex and medulla significantly reduced

Radiology imaging of renal structure and function by CT, MRI, and US during acute reduction in blood flow. Ries et al77 have used a model for type I diabetes, streptozotocin-induced diabetes in rats, and have shown by using BOLD MRI that the oxygenation in all compartments of the kidney is significantly reduced. They interpret this as being related to hyperfiltration associated increase in oxygen consumption. By comparing the observed changes on BOLD with histologic changes, the authors further conclude, the observed MRI changes are not influenced by anatomical or histologic changes, but by functional changes only. Textor et al78 reported on measurements in human subjects with renal artery stenosis. In normal-sized kidneys downstream of high-grade renal arterial stenoses, R2* was elevated at baseline (suggesting enhanced hypoxia) and decreased after the administration of furosemide. This was true even when the glomerular filtration rate (GFR) was significantly reduced. These results are supported by previous reports of preserved cortical tissue volume in poststenotic kidneys, despite reduced function as measured by isotope renography.79 These in turn may suggest that GFR might be recoverable for such cases and that nonfiltering kidney tissue represents a form of “hibernation” in the kidney with the potential for restoring kidney function after reestablishing blood flow.79 By contrast, atrophic kidneys beyond entirely occluded renal arteries demonstrated low levels of R2* (improved oxygenation) and did not change after the administration of furosemide.78 This finding may suggest a nonfunctioning kidney with limited or no oxygen consumption. Similarly, ureteral obstruction was shown to result in greater oxygenation in the kidneys.80 Renal graft dysfunction is a major concern and early characterization of the underlying cause is important. Delayed treatment can lead to the irreversible loss of nephrons and hasten graft loss over time.81,82 Allograft rejection and acute tubular necrosis are 2 important causes of early kidney allograft dysfunction, and it is difficult to discriminate between them by regular clinical tests. Percutaneous transplant biopsy is the most effective method, but it has risks, such as bleeding, kidney rupture, and rarely, graft loss.83,84 Developing a noninvasive method may be highly desirable. Several groups have evaluated the feasibility of BOLD MRI in patients with renal allografts.85-88

Measurement of Differential Renal Function and GFR with MRI Regular low-molecular-weight gadolinium chelates can be considered as glomerular tracers. However, the relationship between signal intensity (SI) and concentration is highly complex, inducing concomitant reduction of T1 and T2 (or T2*), which is not the case for radioactive agents or iodine compounds. Semiquantitative evaluation of renal function (RF), as split (or differential) RF is sufficient in urological management of most uropathies, mainly obstructive. However, it is usually not useful in daily assessment and follow-up of renal diseases. In the nephrologic field, it can be required

51 when a reduced RF is associated with renal asymmetry, in renovascular diseases, before renal surgery if RF is altered or before renal biopsy. However, the level of GFR is the best index for monitoring chronic kidney diseases and its measurement is difficult to obtain accurately in routine. Applications requiring true GFR measurements in nephrology actually include systematic follow-up of renal transplants, evaluation of living donors, when measurement of creatinine clearance is not reliable (low muscle mass, obese) and during all protocols requiring reliable and reproducible RF estimation. The most common method used clinically for accurate GFR measurement is assessment of the plasma disappearance of a substance that is excreted from the body exclusively by glomerular filtration with no tubular reabsorption nor tubular excretion, for example, ethylenediamine tetraacetic acid, or Iothalamate.89 Most of the time, these standard methods are underused because they are quite time-consuming and require several blood/urinary samples. Therefore, a quantitative method based on an intrarenal tracer kinetics, obtained rapidly, without blood and/or urine sampling, coupled with a morphologic evaluation of the kidneys and the entire excretory system would be extremely useful in clinical management of patients with renal disease.

Measurement of Split Renal Function By using dynamic contrast-enhanced MR imaging, Rohrschneider et al90 obtained calculations of the percentage of the single-kidney “activity” comparable with those derived with gamma camera scintigraphy (Fig. 7). These studies were based on a dynamic radiofrequency-spoiled gradient-echo sequence and one-half of a standard clinical accepted dose of Gd-diethylene triamine pentaacetic acid. An ROI was positioned around the renal parenchyma (omitting the pelvis), and calculation of the relative RF was then determined by the following equation: RF ⫽ AUC(mm2) ⫻ S(mm2) where AUC corresponds to the area under the glomerulotubular segment of the time-intensity curve and S is the ROI area. In both an experimental study of ureteral obstruction91 and in patients,92 a high correlation between MR and renal scintigraphy was found. In addition, conversion from SI to concentration of contrast agent is not necessary as recently demonstrated in rats with acute and chronic ureteral obstruction.93 A large multicentric trial in which the authors compared renal scintigraphy and dynamic MRI in adults and children presenting with unilateral obstruction is ongoing in France (unpublished data). The early results of this study show better concordance of MR estimation with renal scintigraphy when Rutland-Patlak plots rather than the Rorschneider approach are used.

MR Quantification of Global GFR Two methods are proposed for measurement of global GFR using MRI and freely filtered Gd-chelates. The first one is based on the measurement of the clearance of an MR agent

N. Grenier et al

52

between 40 and 55 minutes after injection (with a mean paired difference of ⫺5.9 mL/min/1.73 m2 ⫾ 14.6). All measurement points were within ⫾ 2 SD values but the maximum deviation from the reference GFR was 29%. With such a wide deviation, this technique can hardly be applied to an individual patient. Main drawbacks of this method are the length of MR acquisition, the sensitivity to body movements, and to the selected time intervals when the analysis is performed. More experience is required with this technique.

MR Quantification of Single Kidney GFR (SKGFR)

Figure 7 Evaluation of the split renal function in a young infant with a junction obstruction on the left kidney. MRI shows a left pyelectasis (A). The time-intensity curves (B) show an asymmetry between both kidneys with a lower area under the curve on the left side (blue curve). The relative function was calculated as 37% on the left and 63% on the right.

with the use of blood samplings.94 This method presents few advantages compared with other methods because it is timeconsuming and requires several blood samples. The second method is based on measurement of the slope clearance of a freely filtered Gd-chelate from the extracellular fluid volume (ECFV) using SI changes within abdominal organs.95 GFR is calculated as the product of the ECFV (ECFV ⫽ 0.021 54 · weight 0.6469 · height0.3964) and the time constant of the second exponential phase (␣2). The best concordance between gadobutrol clearance and iopromide clearance was observed within the liver, the exponential fit being performed

Two methods are available for measurement of global GFR with MRI and freely filtered Gd-chelates: (1) monitoring of tracer intrarenal kinetics; and (2) measurement of the extraction fraction of the agent. Monitoring of intrarenal tracer kinetics is the most widely used method. However, a great heterogeneity for parameters of pulse-sequences, dose of injected Gd contrast, methods for conversion of SI into concentration, postprocessing methods, and compartment models is still noted in the literature between the different groups, and no consensus exists.96 This quantification requires an accurate sampling of the vascular phase of the enhancement with a high temporal resolution to measure the arterial input function (AIF), which is characterized by the SI changes within the suprarenal abdominal aorta. Without taking this into account will produce an overestimation of GFR because of recirculation of the agent within the vascular space. All pulse-sequences must have a heavy T1-weighting and be fast enough to characterize the vascular phase of the tracer kinetic which is necessary for assessment of the AIF. Concentration of Gd within the kidney can be very high because of water reasbsorption in the proximal convoluted tubule and within the medulla. Therefore, to avoid T2* contribution to the signal, the injected dose must be lowered and the patient well hydrated.97 The choice between 0.025 and 0.05 mmol/kg depends on the level of signal-to-noise ratio obtained with the sequence and the system used. An oblique-coronal plane, passing through the long axis of the kidneys, has to be preferred to an axial plane because with the later, movement correction is more difficult and the AIF can be severely impaired by inflow effects within the aorta. Concerning renal coverage, some cover the entire parenchyma with several slices via the use of a multislice or a 3D acquisition and calculate GFR by summing the GFR values of each voxel in each slice. Others acquire only one median slice, calculate a mean GFR value, and extrapolate this value to the entire renal volume to obtain a total GFR. Because dynamic MRI of the kidneys is performed when the patient is free breathing, the main problem is correction of respiratory movements. Absence of correction produces artifacts in time-intensity evolution that can lead to incorrect quantification. A movement correction method developed recently98 on the basis of the estimation of a rigid transformation and correction allows one to estimate successfully small motion and provided more than 20% of reduction on GFR uncertainty on transplanted kidneys and up to 60% on

Radiology imaging of renal structure and function by CT, MRI, and US

53

Figure 8 Measurement of GFR by the use of dynamic contrast-enhanced MRI and the Rutland-Patlak method. (A) Time-intensity curve with selection of the filtration phase for analysis. (B) Patlak plot of this time interval. (C) Functional map of the glomerular filtration rate based on a ROI limited to cortex.

native kidneys. Extraction of dynamic data should be ideally limited to the cortex, which requires an accurate segmentation method. In the published clinical studies, it was generally performed manually. A recent review has developed and discussed extensively the principles and the difficulties on segmentation and on renal ROI generation.99 The ideal model reflecting filtration physiology remains to be elucidated. Comparison of these models were discussed in previous reviews.100,101 The Rutland-Patlak plot technique102 has been applied to MRI by several groups (Fig. 8).103-107 Others applied a 2-compartment model confined to the cortex, taking into account the outflow from the tubules during the sampling period and making it possible to draw ROIs

strictly limited to the cortex (Fig. 9).103,104 This method has provided more accurate results than Rutland-Patlak model in rabbits, using 51Cr- ethylenediamine tetraacetic acid as a gold standard.103 A more complex model108 includes 3 cortical compartments (glomerular, capillary, and proximal convoluted tubules), 3 medullary compartments (loops of Henle, distal convoluted tubules, collecting ducts), and the collecting system. Despite being complex, this model has the advantage of assessing some important tubular physiological parameters. Applied to the passage of Gd into the first 2 compartments, the model allows the calculation of GFR. Such studies should be accurate and reproducible when compared with a gold standard technique. However only one

N. Grenier et al

54

clinical and experimental settings,106,107,109,110 evaluation seems still necessary at a larger scale. The alternative method allowing a quantification of a single kidney extraction fraction, on the basis of the measurement of T1 within flowing arterial and venous blood (Look and Locker method) during a continuous Gd infusion111-114 provided concordant results with inulin clearance in animals but was never applied to humans, to our knowledge.

Contribution of X-Ray CT in the Evaluation of Kidney Performance With CT, there is a direct and constant linearity between the tissue absorption of X-rays and the measured density on the image that persists after the administration of iodine contrast agents allowing measuring its concentration in the vascular space or in tissues. This proportionality enables the measurement of functional parameters after modeling time density curves measured on electron-beam computed tomography (EBCT) or multidetector computed tomography (MDCT).

Acquisition Deriving from the standard x-ray CT, EBCT was used in the 1980s in the United States. Located behind the subject, an electron source produces an electron beam deviated electromagnetically on an 8 rows X-ray emitter beneath the patient, producing 8 X-ray beams and 8 transversal cross-sectional images. The temporal resolution was very significantly improved (a few tens of milliseconds) without limitation caused by mechanical rotations. These last 2 characteristics allow extraction of renal perfusion and GFR with the gamma variate modeling.115 MDCT is the latest evolution of CT technology. This type of scanner has multiple rows of the torque emitter-detector (up to 64), allowing a rapid volume acquisition (time resolution of approximately 0.5 seconds), with a high spatial resolution. The MDCT allows the fast acquisition of renal images, with isolation of vascular and tissue phases. Figure 9 Measurement of GFR by the use of dynamic contrast-enhanced MRI and a 2-compartment model confined to the cortex, taking into account the outflow from the tubules during the sampling period and making it possible to draw ROIs strictly limited to the cortex. (A) Time-intensity curve with selection of the entire phase for analysis and an accurate fitting of the curve (green line). (B) Functional map of the glomerular filtration rate.

clinical study undertook the 99mTc-DTPA clearance during the Gd-MRI.108 A recent analysis of the literature emphasized the great heterogeneity of protocols (ie, in acquisition mode, dose of contrast, postprocessing techniques) and regression coefficient values, which are generally considered inadequate for replacing an accepted reference method.96 Although the authors of several published papers conclude that Gd-enhanced MRI provides a reliable estimate of GFR and is suitable for use in both

Modeling The extraction of functional parameters, such as renal perfusion, renal blood flow, and GFR is based on the analysis of tissue and vascular density kinetics. Time density curves are obtained by drawing manually ROI on each cross section. Because of linearity, the density in ROI is proportional to the concentration of contrast agent, allowing for absolute quantitative measurements. To obtain reliable and accurate kinetics, the temporal resolution of the scanner must be high, especially during the vascular transit of the contrast media, to monitor the rapid changes in density. Moreover, to prevent any errors related to respiratory movements, it is essential to acquire simultaneously several kidney levels. Today, 2 types of contrast concentration kinetic modeling are used to measure the RF parameters: the gamma variate (Fig. 10) and the Patlak methods (Fig. 11).116,117

Radiology imaging of renal structure and function by CT, MRI, and US

55

Figure 10 The gamma variate modeling decomposes tissue kinetics into different peaks after intravenous injection of 0.5 mL/kg contrast media at 15 mL/s. (A) The renal cortical time density curve includes 3 peaks: a vascular peak corresponding to the transit of contrast in vascular structures of the cortex (used for the measurement of renal blood flow); a proximal peak, (used for the measurement of glomerular filtration rate), and a distal peak (corresponding to the transit of the contrast media) in the proximal tube and in the distal tube, respectively. (B) The time density curve in the medulla includes 2 peaks, a vascular peak for the transit of contrast in the vascular structures of the medulla used for the measurement of renal blood flow and a tubular peak corresponding to the transit of contrast in the loop of Henlé. Reprinted with permission from Krier et al.117

cortical and overall blood flow was measured, whereas there was a compensatory increase of these parameters on the contralateral kidney. For evaluating the endothelial function, regional perfusion was measured before and after injection of vasodilators in pigs that received hypercholesterolemic diet. Under basal conditions, regional perfusion and tubular function were unchanged. However, after the vasodilatation stimulation, renal perfusion, GFR, and the intratubular concentration index was decreased, suggesting that hypercholesterolemia plays a role in the development of renal impairment, particularly at the level of the tubules and induces damages of the microcirculation.120 The same group also evaluated the impact of blocking oxidative stress during ischemic nephropathy. Chade et al121 observed that renal perfusion and GFR increased significantly after blocking the oxidative stress. They concluded that the administration of antioxidants improved endothelial function, probably by increasing the availability of nitric oxide and reducing free radicals. Feldstein et al122 speculated that these microvascular alterations could play a role in the renal lesions of arteriosclerosis. Subsequently, Krier et al117 examined the effect of vasodilatation drugs on the renal perfusion, glomerular flow rate in pigs with renal artery stenosis and showed that the GFR in stenotic kidney decreased proportionally with the degree of stenosis. Other studies showed that the renal artery stenosis lead to redistribution of intrarenal blood flow.123 Daghini et al116 validated in 2007 the use of MDCT for the measurement of renal perfusion, GFR, and tubular dynamics. They found a close correlation between results performed with MDCT and EBCT. Further work focused on the impact of acute ureteral obstruction on glomerular hemodynamic and tubular function in the obstructed kidney. Pelaez et al124 observed that cortical perfusion decreased with obstruction whereas medullary perfusion remained stable, the GFR was maintained the first 30 minutes and then dropped. There was no hemodynamic change on the contralateral kidney. Intratubular concentration of contrast agent increased in the obstructed kidney while it declined on contralateral kidney. Itano et al125 have considered severe ureteral obstruction in pigs and showed that there was a decrease in cortical and medullary blood flow, while the GFR and proximal tubular function were preserved.

Applications in Renal Disease Experiments

Applications in Human Renal Diseases

Using gamma variate modeling, Lerman and Romero at the Mayo Clinic applied this method in many experimental and clinical studies in hypertension, ischemic nephropathy, and ureteral obstruction. With the EBCT, Bentley et al118 observed in dog a reduction in renal blood flow and cortical perfusion when renal perfusion pressure was decreased. Meanwhile, the medullary perfusion remained unchanged. Lerman et al119 also studied adaptation of the contralateral kidney to when the perfusion of the ipsilateral kidney was changed. After a month of hypertension created by unilateral renal artery stenosis in pigs, a significant reduction of the

Lerman et al studied renal perfusion and cortical and medullary volume in humans according to macrovascular lesions. Renal perfusion was compared in patients with normal RF and high blood pressure due to renal artery stenosis by atherosclerosis or dysplasia, or essential hypertension.126 Global and cortical perfusion was more reduced in patients with renal artery stenosis due to atherosclerosis than due to dysplasia or essential hypertension. By contrast, medullary perfusion was identical in all groups. Tsushima et al127 first developed the Patlak modeling to calculate GFR in clinical studies (Fig. 11). In patients with

N. Grenier et al

56

Figure 11 Upper panel: ROI drawn manually on the entire kidney (left) and the cortex (right) when the density of contrast media is maximum in the cortex. Middle panel: Time density curves measured on the entire kidney (left) and restricted to the kidney. Lower panel: Patlak plot derived from the time density curves measured on the entire kidney (left) and restricted to the cortex (right). The plot is constructed with the abscissa corresponding to 兰 C (T) /C (T), where C (T) is the aortic contrast concentration at the time t and the ordinate corresponding-to-the ratio between the tissue and aortic contrast concentration at the time t. The standardized glomerular filtration rate (mL · min⫺1/g) is calculated as the slope of the linear regression of the plot. The y coordinate at the origin is the renal vascular fraction (quantity of blood in the tissue volume). Reprinted with permission from Daghini et al.116

renal disease (hydronephrosis, renovascular hypertension, or diabetes), they compared the GFR values measured with the Patlak model, with 99mDTPA scintigraphy and with the creatinine clearance using a 24-hour urinary collection. There was a close correlation between the GFR values obtained with CT and creatinine clearance (r ⫽ 0.92, P ⬍ 0.0001), as well as with scintigraphy (r ⫽ 0.97, P ⬍ 0.0001). A good correlation (r ⫽ 0.87; P ⬍ 0.0001) was found in the 24 diabetic patients with or without renal dysfunction when GFR values obtained with CT and with the 24-hour urinary creatinine clearance were compared. However, no true reference method, such as inulin clearance, was used to measure GFR. Subsequently, Hackstein et al128,129 simplified the method using only 2 clusters of points to create the Patlak chart instead of numerous points of the time density curve. These

2 clusters are selected during the arterial peak and the tissue phase. There was a very good correlation between CT-GFR value and the clearance of iopromide (r ⫽ 0.889). The authors stressed the fact that it is possible to use this technique during routine scans being performed to evaluate other diagnostic problems. They also showed in patients with hydronephrosis that there is an overestimation of the GFR with CT when renal parenchyma was enlarged. Daghini et al116 proposed a modified Patlak model where the selected ROI included the cortex only instead of the entire kidney. The delay between the arrival of contrast in the aorta and the kidney was also taken into account.130 This modified version provided a better correlation between the CT measurements of GFR and the reference, ie, inulin clearance, with a slight overestimation. However, a better estimation was provided by the gamma variate method.

Radiology imaging of renal structure and function by CT, MRI, and US

57

Advantages and Limits of the Method EBCT and MDCT are functional imaging modalities providing many functional parameters. They provide standardized and global measurements, after correction by tissue volume. Like other types of functional imaging, CT-based imaging modalities allow study of single-kidney hemodynamics and RF, essential information for the evaluation of asymmetric renal pathologies, such as renal artery stenosis or ureteral obstruction. Direct and constant linearity between concentration of contrast agent and the measured density allows obtaining absolute quantitative measurements and performing inter- and intraindividual comparisons over time. The gamma-variate modeling is validated with reference methods and the reproducibility of its measurements was evaluated.117,123,131 Its noninvasive character is quite relative as a central venous catheter is necessary for injections of contrast agent at high rate (15 mL/s). Work is in progress to validate a compartmental model using a peripheral injection of contrast agent. GFR is easier to calculate with the Patlak method because it is a graphical method that does not need to create estimates by nonlinear regressions. In addition, it is applicable with a peripheral catheter for injection of contrast. Unlike EBCT, MDCT is currently available in all hospitals, so theoretically available for a broad and relatively simple use. A limitation of this modality is the injection of contrast (2 ⫻ 0.5 mL/kg for the gamma variate modeling, 100-150 mL for the Patlak method), restricting the application of these procedures in patients with a significant renal insufficiency. The other limit is the exposure to X-rays.132 This must be taken into account in its clinical applications in its potential functional applications. In summary, radiological techniques are now able to quantify many functional and structural parameters within renal parenchyma. However, if CT seems the most appropriate for quantification of filtration function, nephrotoxicity related to iodine contrast agent may preclude its use in diseased patients. US with microbubbles and MRI with small doses of Gd-chelates are more attractive for perfusion and filtration parameters, respectively, but validation of quantification with reference methods is still worthwhile. BOLD and diffusion MR measurements are unique but their implication in clinics has still to be defined.

References 1. Qin S, Caskey CF, Ferrara KW: Ultrasound contrast microbubbles in imaging and therapy: physical principles and engineering. Phys Med Biol 54:R27-R57, 2009 2. Quaia E: Microbubble ultrasound contrast agents: an update. Eur Radiol 17:1995-2008, 2007 3. Sboros V, Tang MX: The assessment of microvascular flow and tissue perfusion using ultrasound imaging. Proc Inst Mech Eng H 224:273290, 2010 4. Cosgrove D, Eckersley R, Blomley M, et al: Quantification of blood flow. Eur Radiol 11:1338-1344, 2001 5. Cosgrove D, Harvey C: Clinical uses of microbubbles in diagnosis and treatment. Med Biol Eng Comput 47:813-826, 2009 6. Quaia E, Blomley MJ, Patel S, et al: Initial observations on the effect of irradiation on the liver-specific uptake of levovist. Eur J Radiol 41: 192-199, 2002 7. Yanagisawa K, Moriyasu F, Miyahara T, et al: Phagocytosis of ultra-

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18. 19.

20.

21.

22. 23.

24.

25.

26.

27.

28.

29.

sound contrast agent microbubbles by Kupffer cells. Ultrasound Med Biol 33:318-325, 2007 Barnett SB, Duck F, Ziskin M: WFUMB Symposium on Safety of Ultrasound in Medicine: Conclusions and Recommendations on Biological Effects and Safety of Ultrasound Contrast Agents, 2006. Ultrasound Med Biol 33:233-234, 2007 Piscaglia F, Bolondi L: The safety of Sonovue in abdominal applications: retrospective analysis of 23188 investigations. Ultrasound Med Biol 32:1369-1375, 2006 Dijkmans PA, Visser CA, Kamp O: Adverse reactions to ultrasound contrast agents: is the risk worth the benefit? Eur J Echocardiogr 6:363-366, 2005 Kusnetzky LL, Khalid A, Khumri TM, et al: Acute mortality in hospitalized patients undergoing echocardiography with and without an ultrasound contrast agent: results in 18,671 consecutive studies. J Am Coll Cardiol 51:1704-1706, 2008 Mulvagh SL, Rakowski H, Vannan MA, et al: American Society of Echocardiography consensus statement on the clinical applications of ultrasonic contrast agents in echocardiography. J Am Soc Echocardiogr 21:1179-1201, 2008; quiz: 1281 Senior R, Becher H, Monaghan M, et al: Contrast echocardiography: evidence-based recommendations by European Association of Echocardiography. Eur J Echocardiogr 10:194-212, 2009 Correas JM, Burns PN, Lai X, et al: Infusion versus bolus of an ultrasound contrast agent: in vivo dose–response measurements of BR1. Invest Radiol 35:72-79, 2000 Mari J, Hibbs K, Stride E, et al: An approximate nonlinear model for time gain compensation of amplitude modulated images of ultrasound contrast agent perfusion. IEEE Trans Ultrason Ferroelectr Freq Control 57:818-829, 2010 Mule S, De Cesare A, Lucidarme O, et al: Regularized estimation of contrast agent attenuation to improve the imaging of microbubbles in small animal studies. Ultrasound Med Biol 34:938-948, 2008 Phillips P, Gardner E: Contrast-agent detection and quantification. Eur Radiol 14:4-10, 2004 (suppl 8) Gu X, Zhong H, Wan M, et al: Parametric perfusion imaging based on low-cost ultrasound platform. Ultrasound Med Biol 36:130-144, 2010 Wiesmann M, Meyer K, Albers T, et al: Parametric perfusion imaging with contrast-enhanced ultrasound in acute ischemic stroke. Stroke; J Cereb Circ 35:508-513, 2004 Wei K, Jayaweera AR, Firoozan S, et al: Quantification of myocardial blood flow with ultrasound-induced destruction of microbubbles administered as a constant venous infusion. Circulation 97:473-483, 1998 Potdevin TC, Fowlkes JB, Moskalik AP, et al: Analysis of refill curve shape in ultrasound contrast agent studies. Med Phys 31:623-632, 2004 Potdevin TC, Fowlkes JB, Moskalik AP, et al: Refill model of rabbit kidney vasculature. Ultrasound Med Biol 32:1331-1338, 2006 Lucidarme O, Kono Y, Corbeil J, et al: Validation of ultrasound contrast destruction imaging for flow quantification. Ultrasound Med Biol 29:1697-1704, 2003 Arditi M, Frinking PJ, Zhou X, et al: A new formalism for the quantification of tissue perfusion by the destruction-replenishment method in contrast ultrasound imaging. IEEE Trans Ultrason Ferroelectr Freq Control 53:1118-1129, 2006 Porter TR, Xie F, Silver M, et al: Real-time perfusion imaging with low mechanical index pulse inversion Doppler imaging. J Am Coll Cardiol 37:748-753, 2001 Meier P, Zierler KL: On the theory of the indicator-dilution method for measurement of blood flow and volume. J Appl Physiol 6:731-744, 1954 Krix M, Kiessling F, Farhan N, et al: A multivessel model describing replenishment kinetics of ultrasound contrast agent for quantification of tissue perfusion. Ultrasound Med Biol 29:1421-1430, 2003 Wei K, Le E, Bin JP, et al: Quantification of renal blood flow with contrast-enhanced ultrasound. J Am Coll Cardiol 37:1135-1140, 2001 Wei K, Ragosta M, Thorpe J, et al: Noninvasive quantification of

N. Grenier et al

58

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43. 44. 45. 46.

47.

48.

49.

50.

coronary blood flow reserve in humans using myocardial contrast echocardiography. Circulation 103:2560-2565, 2001 Meyer-Wiethe K, Cangur H, Seidel GU: Comparison of different mathematical models to analyze diminution kinetics of ultrasound contrast enhancement in a flow phantom. Ultrasound Med Biol 31: 93-98, 2005 Lucidarme O, Franchi-Abella S, Correas JM, et al: Blood flow quantification with contrast-enhanced US: “entrance in the section” phenomenon—phantom and rabbit study. Radiology 228:473-479, 2003 Quaia E, Nocentini A, Torelli L: Assessment of a new mathematical model for the computation of numerical parameters related to renal cortical blood flow and fractional blood volume by contrast-enhanced ultrasound. Ultrasound Med Biol 35:616-627, 2009 Kishimoto N, Mori Y, Nishiue T, et al: Renal blood flow measurement with contrast-enhanced harmonic ultrasonography: evaluation of dopamine-induced changes in renal cortical perfusion in humans. Clin Nephrol 59:423-428, 2003 Kihm LP, Hinkel UP, Michael K, et al: Contrast enhanced sonography shows superior microvascular renal allograft perfusion in patients switched from cyclosporine A to everolimus. Transplantation 88:261265, 2009 Kim JH, Eun HW, Lee HJ, et al: Clinical use of renal perfusion imaging by means of harmonic sonography with a microbubble contrast agent in patients after renal transplantation: preliminary study. J Ultrasound Med 24:755-762, 2005 Schwenger V, Korosoglou G, Hinkel UP, et al: Real-time contrastenhanced sonography of renal transplant recipients predicts chronic allograft nephropathy. Am J Transplant 6:609-615, 2006 Claudon M, Barnewolt CE, Taylor GA, et al: Renal blood flow in pigs: changes depicted with contrast-enhanced harmonic US imaging during acute urinary obstruction. Radiology 212:725-731, 1999 Quaia E, Siracusano S, Palumbo A, et al: Detection of focal renal perfusion defects in rabbits after sulphur hexafluoride-filled microbubble injection at low transmission power ultrasound insonation. Eur Radiol 16:166-172, 2006 Taylor GA, Barnewolt CE, Adler BH, et al: Renal cortical ischemia in rabbits revealed by contrast-enhanced power Doppler sonography. AJR Am J Roentgenol 170:417-422, 1998 Taylor GA, Barnewolt CE, Claudon M, et al: Depiction of renal perfusion defects with contrast-enhanced harmonic sonography in a porcine model. AJR Am J Roentgenol 173:757-760, 1999 Taylor GA, Ecklund K, Dunning PS: Renal cortical perfusion in rabbits: visualization with color amplitude imaging and an experimental microbubble-based US contrast agent. Radiology 201:125-129, 1996 Bertolotto M, Martegani A, Aiani L, et al: Value of contrast-enhanced ultrasonography for detecting renal infarcts proven by contrast enhanced CT. A feasibility study. Eur Radiol 18:376-383, 2008 Correas JM, Claudon M, Tranquart F, et al: [Contrast-enhanced ultrasonography: renal applications]. J Radiol 84:2041-2054, 2003 Hazlewood CF, Rorschach HE, Lin C: Diffusion of water in tissues and MRI. Magn Reson Med 19:214-216, 1991 Rorschach HE, Lin C, Hazlewood CF: Diffusion of water in biological tissues. Scan Microsc Suppl 5:S1-S9, 1991; discussion: S9-10 Muller MF, Prasad P, Siewert B, et al: Abdominal diffusion mapping with use of a whole-body echo-planar system. Radiology 190:475478, 1994 Muller MF, Prasad PV, Bimmler D, et al: Functional imaging of the kidney by means of measurement of the apparent diffusion coefficient. Radiology 193:711-715, 1994 Kim S, Naik M, Sigmund E, et al :Diffusion-weighted MR imaging of the kidneys and the urinary tract. Magn Reson Imaging Clin N Am 16:585-596:vii-viii, 2008 Thoeny HC, De Keyzer F, Oyen RH, et al: Diffusion-weighted MR imaging of kidneys in healthy volunteers and patients with parenchymal diseases: initial experience. Radiology 235:911-917, 2005 Basser PJ, Mattiello J, LeBihan D: Estimation of the effective selfdiffusion tensor from the NMR spin echo. J Magn Reson B 103:247254, 1994

51. Mori S, van Zijl PC: Fiber tracking: principles and strategies—a technical review. NMR Biomed 15:468-480, 2002 52. Engelter ST, Wetzel SG, Bonati LH, et al: The clinical significance of diffusion-weighted MR imaging in stroke and TIA patients. Swiss Med Wkly 138:729-740, 2008 53. Carbone SF, Gaggioli E, Ricci V, et al: Diffusion-weighted magnetic resonance imaging in the evaluation of renal function: a preliminary study. Radiol Med 112:1201-1210, 2007 54. Afdhal NH, Nunes D: Evaluation of liver fibrosis: a concise review. Am J Gastroenterol 99:1160-1174, 2004 55. Lu L, Lee G, Gulani V, et al: Diffusion tensor imaging as a biomarker of diabetic nephropathy. International Society of Magnetic Resonance in Medicine, Stockholm, Sweden, 2010, p 97 56. Brezis M, Rosen S: Hypoxia of the renal medulla—its implications for disease. N Engl J Med 332:647-655, 1995 57. Eckardt KU, Bernhardt WM, Weidemann A, et al: Role of hypoxia in the pathogenesis of renal disease. Kidney Int Suppl 99:S46-S51, 2005 58. Norman JT, Fine LG: Intrarenal oxygenation in chronic renal failure. Clin Exp Pharmacol Physiol 33:989-996, 2006 59. Epstein FH, Agmon Y, Brezis M: Physiology of renal hypoxia. Ann NY Acad Sci 718:72-81, 1994; discussion: 81-72 60. Matthews PM, Jezzard P: Functional magnetic resonance imaging. J Neurol Neurosurg Psychiatry 75:6-12, 2004 61. Rajagopalan P, Krishnan KR, Passe TJ, et al: Magnetic resonance imaging using deoxyhemoglobin contrast versus positron emission tomography in the assessment of brain function. Prog Neuropsychopharmacol Biol Psychiatry 19:351-366, 1995 62. Ugurbil K, Hu X, Chen W, et al: Functional mapping in the human brain using high magnetic fields. Philos Trans R Soc Lond B Biol Sci 354:1195-1213, 1999 63. Dunn JF, Swartz HM: Blood oxygenation. Heterogeneity of hypoxic tissues monitored using bold MR imaging. Adv Exp Med Biol 428: 645-650, 1997 64. Prasad PV, Chen Q, Goldfarb JW, et al: Breath-hold R2* mapping with a multiple gradient-recalled echo sequence: application to the evaluation of intrarenal oxygenation. J Magn Reson Imaging 7:11631165, 1997 65. Prasad PV, Edelman RR, Epstein FH: Noninvasive evaluation of intrarenal oxygenation with BOLD MRI. Circulation 94:3271-3275, 1996 66. Priatna A, Epstein FH, Spokes K, et al: Evaluation of changes in intrarenal oxygenation in rats using multiple gradient-recalled echo (mGRE) sequence. J Magn Reson Imaging 9:842-846, 1999 67. Brezis M, Agmon Y, Epstein FH: Determinants of intrarenal oxygenation. I. Effects of diuretics. Am J Physiol 267:F1059-F1062, 1994 68. Epstein FH, Veves A, Prasad PV: Effect of diabetes on renal medullary oxygenation during water diuresis. Diabetes Care 25:575-578, 2002 69. Prasad PV, Epstein FH: Changes in renal medullary pO2 during water diuresis as evaluated by blood oxygenation level-dependent magnetic resonance imaging: effects of aging and cyclooxygenase inhibition. Kidney Int 55:294-298, 1999 70. Tumkur SM, Vu AT, Li LP, et al: Evaluation of intra-renal oxygenation during water diuresis: a time-resolved study using BOLD MRI. Kidney Int 70:139-143, 2006 71. Lin JL, Li L, Schnitzer T, et al: Intra-renal oxygenation in Rat kidneys during water-loading: effects of COX inhibition and NO donation. J Magn Reson Imaging in press 72. Muscara MN, Wallace JL: COX-inhibiting nitric oxide donors (CINODs): potential benefits on cardiovascular and renal function. Cardiovasc Hematol Agents Med Chem 4:155-164, 2006 73. Li L, Storey P, Kim D, et al: Kidneys in hypertensive rats show reduced response to nitric oxide synthase inhibition as evaluated by BOLD MRI. J Magn Reson Imaging 17:671-675, 2003 74. Li LP, Li BS, Storey P, et al: Effect of free radical scavenger (tempol) on intrarenal oxygenation in hypertensive rats as evaluated by BOLD MRI. J Magn Reson Imaging 21:245-248, 2005 75. Li LP, Ji L, Santos EA, et al: Effect of nitric oxide synthase inhibition on intrarenal oxygenation as evaluated by blood oxygenation level-dependent magnetic resonance imaging. Invest Radiol 44:67-73, 2009 76. Juillard L, Lerman LO, Kruger DG, et al: Blood oxygen level-depen-

Radiology imaging of renal structure and function by CT, MRI, and US

77.

78.

79.

80.

81. 82.

83.

84.

85.

86.

87.

88.

89. 90.

91.

92.

93.

94.

95.

96.

97.

dent measurement of acute intra-renal ischemia. Kidney Int 65:944-950, 2004 Ries M, Basseau F, Tyndal B, et al: Renal diffusion and BOLD MRI in experimental diabetic nephropathy. Blood oxygen level-dependent. J Magn Reson Imaging 17:104-113, 2003 Textor SC, Glockner JF, Lerman LO, et al: The use of magnetic resonance to evaluate tissue oxygenation in renal artery stenosis. J Am Soc Nephrol 19:780-788, 2008 Cheung CM, Shurrab AE, Buckley DL, et al: MR-derived renal morphology and renal function in patients with atherosclerotic renovascular disease. Kidney Int 69:715-722, 2006 Thoeny HC, Kessler TM, Simon-Zoula S, et al: Renal oxygenation changes during acute unilateral ureteral obstruction: assessment with blood oxygen level-dependent MR imaging—initial experience. Radiology 247:754-761, 2008 Breza J, Navratil P: Renal transplantation in adults. BJU Int 84:216223, 1999 Ojo AO, Wolfe RA, Held PJ, et al: Delayed graft function: risk factors and implications for renal allograft survival. Transplantation 63:968974, 1997 Gainza FJ, Minguela I, Lopez-Vidaur I, et al: Evaluation of complications due to percutaneous renal biopsy in allografts and native kidneys with color-coded Doppler sonography. Clin Nephrol 43:303-308, 1995 Preda A, Van Dijk LC, Van Oostaijen JA, et al: Complication rate and diagnostic yield of 515 consecutive ultrasound-guided biopsies of renal allografts and native kidneys using a 14-gauge Biopty gun. Eur Radiol 13:527-530, 2003 Djamali A, Sadowski EA, Muehrer RJ, et al: BOLD-MRI assessment of intrarenal oxygenation and oxidative stress in patients with chronic kidney allograft dysfunction. Am J Physiol Ren Physiol 292:F513F522, 2007 Han F, Xiao W, Xu Y, et al: The significance of BOLD MRI in differentiation between renal transplant rejection and acute tubular necrosis. Nephrol Dial Transplant 23:2666-2672, 2008 Malvezzi P, Bricault I, Terrier N, et al: Evaluation of intrarenal oxygenation by blood oxygen level-dependent magnetic resonance imaging in living kidney donors and their recipients: preliminary results. Transplant Proc 41:641-644, 2009 Thoeny HC, Zumstein D, Simon-Zoula S, et al: Functional evaluation of transplanted kidneys with diffusion-weighted and BOLD MR imaging: initial experience. Radiology 241:812-821, 2006 Prigent A: Monitoring renal function and limitations of renal function tests. Semin Nucl Med 38:32-46, 2008 Rohrschneider WK, Hoffend J, Becker K, et al: Combined static-dynamic MR urography for the simultaneous evaluation of morphology and function in urinary tract obstruction. I. Evaluation of the normal status in an animal model. Pediatr Radiol 30:511-522, 2000 Rohrschneider WK, Becker K, Hoffend J, et al: Combined static-dynamic MR urography for the simultaneous evaluation of morphology and function in urinary tract obstruction. II. Findings in experimentally induced ureteric stenosis. Pediatr Radiol 30:523-532, 2000 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 224:683-694, 2002 Pedersen M, Shi Y, Anderson P, et al: Quantitation of differential renal blood flow and renal function using dynamic contrast-enhanced MRI in rats. Magn Reson Med 51:510-517, 2004 Choyke PL, Austin HA, Frank JA, et al: Hydrated clearance of gadolinium-DTPA as a measurement of glomerular filtration rate. Kidney Int 41:1595-1598, 1992 Boss A, Martirosian P, Gehrmann M, et al: Quantitative assessment of glomerular filtration rate with MR gadolinium slope clearance measurements: a phase I trial. Radiology 242:783-790, 2007 Mendichovszky I, Pedersen M, Frokiaer J, et al: How accurate is dynamic contrast-enhanced MRI in the assessment of renal glomerular filtration rate? A critical appraisal. J Magn Reson Imaging 27:925-931, 2008 Rusinek H, Lee VS, Johnson G: Optimal dose of Gd-DTPA in dynamic MR studies. Magn Reson Med 46:312-316, 2001

59 98. de Senneville BD, Mendichovszky IA, Roujol S, et al: Improvement of MRI-functional measurement with automatic movement correction in native and transplanted kidneys. J Magn Reson Imaging 28:970-978, 2008 99. Michoux N, Vallee JP, Pechere-Bertschi A, et al: Analysis of contrastenhanced MR images to assess renal function. Magma 19:167-179, 2006 100. Bokacheva L, Rusinek H, Zhang JL, et al :Assessment of renal function with dynamic contrast-enhanced MR imaging. Magn Reson Imaging Clin N Am 16:597-611:viii, 2008 101. Grenier N, Mendichovszky I, de Senneville BD, et al: Measurement of glomerular filtration rate with magnetic resonance imaging: principles, limitations, and expectations. Semin Nucl Med 38:47-55, 2008 102. Patlak CS, Blasberg RG, Fenstermacher JD: Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 3:1-7, 1983 103. Annet L, Hermoye L, Peeters F, et al: Glomerular filtration rate: assessment with dynamic contrast-enhanced MRI and a cortical-compartment model in the rabbit kidney. J Magn Reson Imaging 20:843849, 2004 104. Hermoye L, Annet L, Lemmerling P, et al: Calculation of the renal perfusion and glomerular filtration rate from the renal impulse response obtained with MRI. Magn Reson Med 51:1017-1025, 2004 105. Pedersen M, Dissing T, Deding D, et al: MR renography based on contrast-enhanced T1-mapping. In: Medicine ISfMRi (ed) International Society for Magnetic Resonance in Medicine, Miami, 2005, p 526 106. Hackstein N, Cengiz H, Rau WS: Contrast media clearance in a single kidney measured on multiphasic helical CT: results in 50 patients without acute renal disorder. AJR Am J Roentgenol 178:111-118, 2002 107. 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 18:714-725, 2003 108. Lee VS, Rusinek H, Bokacheva L, et al: Renal function measurements from MR renography and a simplified multicompartmental model. Am J Physiol 292:F1548-F1559, 2007 109. Laurent D, Poirier K, Wasvary J, et al: Effect of essential hypertension on kidney function as measured in rat by dynamic MRI. Magn Reson Med 47:127-134, 2002 110. Lee VS, Rusinek H, Noz ME, et al: Dynamic three-dimensional MR renography for the measurement of single kidney function: initial experience. Radiology 227:289-294, 2003 111. Dumoulin CL, Buonocore MH, Opsahl LR, et al: Noninvasive measurement of renal hemodynamic functions using gadolinium enhanced magnetic resonance imaging. Magn Reson Med 32:370-378, 1994 112. Niendorf ER, Santyr GE, Brazy PC, et al: Measurement of Gd-DTPA dialysis clearance rates by using a look-locker imaging technique. Magn Reson Med 36:571-578, 1996 113. Niendorf ER, Grist TM, Frayne R, et al: Rapid measurement of GdDTPA extraction fraction in a dialysis system using echo-planar imaging. Med Phys 24:1907-1913, 1997 114. Niendorf ER, Grist TM, Lee FT Jr, et al: Rapid in vivo measurement of single-kidney extraction fraction and glomerular filtration rate with MR imaging. Radiology 206:791-798, 1998 115. Lerman LO, Bentley MD, Bell MR, et al: The effect of a low-osmolar radiographic contrast medium on in vivo and postmortem renal size. Invest Radiol 26:992-997, 1991 116. Daghini E, Juillard L, Haas JA, et al: Comparison of mathematic models for assessment of glomerular filtration rate with electron-beam CT in pigs. Radiology 242:417-424, 2007 117. Krier JD, Ritman EL, Bajzer Z, et al: Noninvasive measurement of concurrent single-kidney perfusion, glomerular filtration, and tubular function. Am J Physiol Ren Physiol 281:F630-F638, 2001 118. Bentley MD, Lerman LO, Hoffman EA, et al: Measurement of renal perfusion and blood flow with fast computed tomography. Circ Res 74:945-951, 1994

60 119. Lerman LO, Schwartz RS, Grande JP, et al: Noninvasive evaluation of a novel swine model of renal artery stenosis. J Am Soc Nephrol 10: 1455-1465, 1999 120. Chade AR, Rodriguez-Porcel M, Grande JP, et al: Distinct renal injury in early atherosclerosis and renovascular disease. Circulation 106: 1165-1171, 2002 121. Chade AR, Krier JD, Rodriguez-Porcel M, et al: Comparison of acute and chronic antioxidant interventions in experimental renovascular disease. Am J Physiol Ren Physiol 286:F1079-F1086, 2004 122. Feldstein A, Krier JD, Sarafov MH, et al: In vivo renal vascular and tubular function in experimental hypercholesterolemia. Hypertension 34:859-864, 1999 123. Romero JC, Lerman LO: Novel noninvasive techniques for studying renal function in man. Semin Nephrol 20:456-462, 2000 124. Pelaez LI, Juncos LA, Stulak JM, et al: Non-invasive evaluation of bilateral renal regional blood flow and tubular dynamics during acute unilateral ureteral obstruction. Nephrol Dial Transplant 20:83-88, 2005 125. Itano NB, Sherrill LE, Lerman LO, et al: Electron beam computerized tomography assessment of in vivo single kidney glomerular filtration rate and tubular dynamics during chronic partial unilateral ureteral obstruction in the pig. J Urol 166:2530-2535, 2001

N. Grenier et al 126. Lerman LO, Taler SJ, Textor SC, et al: Computed tomography-derived intrarenal blood flow in renovascular and essential hypertension. Kidney Int 49:846-854, 1996 127. Tsushima Y, Blomley MJ, Okabe K, et al: Determination of glomerular filtration rate per unit renal volume using computerized tomography: correlation with conventional measures of total and divided renal function. J Urol 165:382-385, 2001 128. Hackstein N, Bauer J, Hauck EW, et al: Measuring single-kidney glomerular filtration rate on single-detector helical CT using a two-point Patlak plot technique in patients with increased interstitial space. AJR Am J Roentgenol 181:147-156, 2003 129. Hackstein N, Wiegand C, Rau WS, et al: Glomerular filtration rate measured by using triphasic helical CT with a two-point Patlak plot technique. Radiology 230:221-226, 2004 130. Tsushima Y: Functional CT of the kidney. Eur J Radiol 30:191-197, 1999 131. Lerman LO, Flickinger AL, Sheedy PF 2nd, et al: Reproducibility of human kidney perfusion and volume determinations with electron beam computed tomography. Invest Radiol 31:204-210, 1996 132. Einstein AJ, Henzlova MJ, Rajagopalan S: Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA 298:317-323, 2007