Eur. Radiol. (2001) 11: 355±372 Ó Springer-Verlag 2001
Claus C. A. Nolte-Ernsting Gerhard B. Adam Rolf W. Günther
Received: 7 July 2000 Revised: 21 August 2000 Accepted: 21 August 2000
)
C. C. A. Nolte-Ernsting ( ) ´ G. B. Adam ´ R. W. Günther Department of Diagnostic Radiology, University of Technology, Pauwelsstrasse 30, 52057 Aachen, Germany e-mail:
[email protected] Tel.: + 49-2 41-8 08 83 32 Fax: + 49-2 41-8 88 84 11
U R O G E N I TAL
MR urography: examination techniques and clinical applications
Abstract Modern MR urography is performed on the basis of two different imaging strategies, which can be used complementarily to cover almost all aspects in the diagnosis of upper urinary tract diseases. The first technique utilizes unenhanced, heavily T2-weighted pulse sequences to obtain static-fluid images of the urinary tract. T2-weighted MR urograms have proved to be excellent in the visualization of the markedly dilated urinary tract, even if the renal excretory function is quiescent. Static-fluid MR urography is less suitable for imaging of disorders that occur in the nondilated collecting system. The second MR urography technique is analogous to the methodology of conventional intravenous pyelography and is, therefore, designated as excretory MR urography. For this purpose, a nonnephrotoxic gadolinium chelate is intravenously administered and after its renal excretion, the gadolinium-enhanced urine is visualized using fast T1-weighted gradient-echo sequences. The combination of gadolinium and low-dose furosemide (5±10 mg) is the key for achieving a uniform distribution of the contrast
Introduction Especially during the past 5 years, MR urography has advanced from the state of basic research to use in clinical practice. Clinical MR urography, however, is still in
material inside the entire urinary tract and, secondly, to avoid high endoluminal gadolinium concentrations, which cause signal loss of the urine due to T2* effects. Gadolinium excretory MR urography allows to obtain high-quality images of both nondilated and obstructed urinary tracts in patients with normal or moderately impaired renal function. This article reviews the principles of T2- and T1-weighted MR urography in detail and informs how to use these techniques safely in potential clinical applications such as chronic urolithiasis, intrinsic and extrinsic tumor diseases, and congenital anomalies. Magnetic resonance urography performed in combination with standard MR imaging offers a potential to reduce the need for invasive retrograde pyelography. Although the economic aspect is still problematic, it is obvious that MR urography will continue to increase its role in clinical uroradiology. Key words Kidney ´ Urography ´ MR imaging ´ Gadolinium ´ Furosemide
its infancy and seeks widespread dissemination. Most of us have already seen impressive MR urograms, but the various underlying examination techniques, hardware requirements, MR pulse sequences, and potential indications are not commonly known. The goal of this re-
356
view is to offer a thorough approach to this new way of imaging of the upper urinary tract and to provide general guidelines for the performance of MR urography, which is easy to put into operation. Unenhanced T2weighted and gadolinium-enhanced T1-weighted MR urography techniques are described in detail, which forms the basis for a safe application of MR urography in the clinical routine. With regard to the major urinary tract disorders, potential indications for MR urography are presented and illustrated by numerous clinical examples.
Hardware requirements Magnetic resonance urography is most practically carried out at 1.5 T. Nevertheless, several investigators have successfully performed MR urography at low- and mid-field strength between 0.23 and 0.5 T [1, 2, 3, 4, 5]. High field strength magnets can be equipped either with a conventional gradient system [6] or, preferably, with ultrafast high-performance gradients [7]. Different surface coils were employed in several studies dealing with MR urography [3, 6, 8, 9, 10, 11]. The body coil alone is also able to do a good job in most of the MR urographic examinations performed in adult patients [7, 12]. In the near future, modern surface coil techniques will certainly help to improve the image quality in high-resolution MR urography. The use of external ureteric compression has been recommended by a few authors [10, 13, 14]; however, modern MR urography techniques actually obviate the need for an inconvenient compression device [6, 15, 16].
T2-weighted MR urography T2-weighted MR urograms simply generate static water images (static-fluid MR urography). The water that we want to see is, of course, the urine itself, which can be regarded as an ªintrinsic contrast medium.º Static-fluid MR urography offers a diagnostic tool, which is independent of the renal excretory function. Thus, the important advantage is that T2-weighted MR urography can be performed even in patients with non-excreting kidneys [3, 6, 8, 12, 13, 14, 17]. On the other hand, the methodology of static-fluid MR urography does not yield functional information about the urinary tract [3, 8, 12, 16, 17]. In order to obtain water images of the urinary tract, heavily T2-weighted pulse sequences are necessary, which display the urine with a hyperintense signal intensity and with a high contrast toward surrounding extraurinary tissues. Turbo spin-echo sequences have proved to be optimal for obtaining heavily T2-weighted images within short data acquisition times. The principle of turbo spin-echo imaging was pursued in the mid-
1980 s by Hennig and co-workers who called this technique rapid acquisition with relaxation enhancement (RARE) [18]. Indeed, it was the RARE sequence which was first used to obtain urogram-like MR images [1]. In RARE imaging, more generally referred to as turbo spin-echo (TSE) imaging, the reduction in scan time is achieved by the characteristic feature that each 90 excitation pulse is followed by multiple 180 pulses generating a train of consecutive echoes, each of which experiences a different phase-encoding step. The number of echoes following a 90 pulse is called echo train length (ETL) or TSE factor. The number of necessary 90 pulses per image equals the number of repetition times (TR), which importantly determines the sequence duration. Compared with a standard spin-echo pulse sequence, the number of 90 pulses applied in a TSE sequence is markedly reduced, namely by the division of the TSE factor. In a multi-shot TSE sequence, still several 90 pulses have to be applied per acquisition of a single MR image. With increasing ETL, the total number of necessary 90 pulses decreases and, accordingly, the TSE sequence becomes faster. A single-shot TSE sequence is even more rapid, because only one 90 pulse is applied and the echo train must be long enough to complete a whole single image. Since we only apply a single 90 pulse per slice, there actually does not exist a real TR. The TR that is usually given by the sequence protocol of our magnets (Table 1) equals the time between two single shots, each of which is applied for an individual slice. Another important parameter in TSE imaging is the so-called effective echo time (TE). In a single-shot TSE sequence, the effective TE indicates the echo profile, which is placed in the center line of the k-space and which is, therefore, the most pronounced echo of the train because it experiences the weakest phase-encoding gradient. Since we want to obtain a heavily T2-weighted image, the effective TE should be long (Table 1). Finally, the combination of a single-shot TSE sequence and a half-Fourier data acquisition provides the fastest variant of the RARE technique which is called HASTE (half-Fourier acquisition single-shot turbo spin echo). The HASTE sequence allows us to acquire a single T2-weighted MR image within a breath-hold of less than 5 s and is, to date, the most commonly recommended technique for use in static-fluid MR urography [9, 12, 15, 17]. Typical parameter values of the HASTE sequence are listed in Table 1. Additional spectral fat suppression can help to further improve the contrast between urinary tract and retroperitoneum [5, 9, 12, 13, 14, 17]. Basically, there are two possibilities of how to use the RARE or HASTE sequences in T2-weighted MR urography: On one hand, the whole urinary tract can be visualized by acquisition of only a single-slice projection image (Table 1). The slice thickness must be very thick,
357
Table 1 Typical parameters of T2- and T1-weighted MR urography pulse sequences. a flip angle; FOV field of view; RFOV rectangular FOV; NSA no. of signals averaged; GRE gradient-echo; EPI echo-planar imaging Sequence
Type of MR urogram
Respiratory TR/TE/a compensation
1) T2 HASTE
Static-fluid
Breath-hold
Projection imagea 2) T2 HASTE
Static-fluid
Multislice
Breath-hold
Excretory
Gating
Excretory
Breath-hold
5000c/200 msb 130/± 15/3 ms 4±6/2±3 ms
Excretory
Breath-hold
6) T1 3D GRE-EPI Excretory
Breath-hold
Survey MRU
-/-
Excretory
Breath-hold
360±390 mm
-/-
212 256
360±390 mm
179±196 256
85±100 %
4±6/2±3 ms
-/-
300±330 mm
30±40
-/-
85±95 %
14/6 ms
-/5
32/4 ms
256 256
80±100 %
30±40
Projection imagea
360±390 mm 85 %
30±40
Detail MRU
7) T1 2D GRE
256 256
360±390 mm 85 %
30±40
Survey MRU 5) T1 3D GRE
130/-
or Triggering 90
Survey MRU 4) T1 3D GRE
Matrix
90
a
3) T1 3D GRE
-/200 msb
TSE factor/ FOV/ EPI factor RFOV
360±390 mm
196±212 256 179 256
85±100 % -/-
70
256±270 mm
256 256
85 %
No. of slices/ Scan duration/ slice thickness NSA 1
3±5 s
70±100 mm
1
20±30
1.5±2.5 min
3±5 mm
2
50±60
3 min
1.5 mm
1
50±60
17±25 s
1.5 mm
1
35±45
18±25 s
1.3 mm
1
50±60
14±18 s
1.4 mm
1
1
13 s
35±50 mm
2
a
With spectral fat suppression Effective echo time c Time between two single shots b
e. g., 6±8 cm, in order to include the entire pelvicaliceal system and the whole three-dimensional course of the ureters (Fig. 1). The anisotropic voxel geometry of a projection MR image consists of multiple columns with edges of, for example, 1 1 80 mm3. The advantage of the projection technique is that we are able to obtain a complete survey image of the urinary tract in 3±5 s without motion artifacts. Furthermore, projection MR urograms are immediately visible on the screen without requiring any postprocessing. The acquisition of a T2weighted projection MR urogram is often good enough for the gross diagnosis of a ureterohydronephrosis and the location of an obstruction (Fig. 1) [8, 17]. T2weighted projection MR urograms, however, often fail to visualize the cause of obstruction [8]. Projection MR urograms are usually obtained in the coronal plane and, additionally, each side of the urinary tract can also be imaged in the sagittal plane (Fig. 1) [3, 8]. The second way of imaging with heavily T2-weighted sequences utilizes the multislice technique. Either 2D or 3D TSE sequences with or without half-Fourier data acquisition can be employed for this purpose. Multiple overlapping slices are obtained with a section thickness of less than 5 mm (Table 1) and, thereafter, the individual slices are mathematically postprocessed to generate maximum intensity projection (MIP) images parallel to the long axis of the body using radial increments of, for example, 15±30 . The multislice method is more timeconsuming, but the acquisition of thin single slices reduces partial-volume averaging and offers a better op-
portunity to detect small pathologic details, which might be missed if only a thick projection image is available [13, 16, 17]. One general disadvantage of T2-weighted MR urography is that superimposed abdominal fluid collections can occasionally degrade the visibility of the urinary tract on both MIP images and projection images [9, 17, 19]. The right pelvicaliceal system can be superimposed by the duodenum, the common bile duct, and the gall bladder. Moreover, a frequent finding is that fluid-filled bowel loops obscure segments of the ureters. Again, the thin single sections of a multislice sequence are preferable because they are not affected by superimposed intestinal structures.
T1-weighted MR urography The second MR urography technique imitates the methodology of conventional X-ray urography and is, therefore, referred to as excretory MR urography [6]. With this technique, we exclusively visualize the contrast-material-enhanced urine with fast T1-weighted pulse sequences obtained after renal excretion of an intravenously injected low-molecular weight gadolinium chelate [6, 7]. The practicability of gadolinium-enhanced T1-weighted MR urography is dependent on the renal excretory function. Consequently, T1-weighted MR urography provides both morphologic and at least gross functional information about the urinary tract [6].
358
a
b Fig. 1 a, b A 71-year-old male patient with a giant left-sided hydronephrosis demonstrated by heavily T2-weighted half-Fourier acquisition single-shot turbo spin echo (HASTE) projection images using a slice thickness of 70 mm in a the coronal plane and b the sagittal plane. Only the sagittal section uncovers a prevesical obstruction (arrowhead), which cannot be further characterized by the projection urograms. This filling defect was caused by a calculus that was finally identified with a standard spin-echo sequence; however, ureterolithiasis is unlikely to be the exclusive origin for such a severe hydronephrosis
The basic mechanisms in excretory MR urography are complex and are outlined in detail herein. Pharmacological aspects in excretory MR urography Firstly, the value of gadolinium contrast agents for the use in excretory MR urography must be addressed. In this review, gadolinium-DTPA (Magnevist, Schering,
Berlin, Germany) is regarded as a representative of all commercially available low-molecular gadolinium compounds that are characterized by renal excretion. After intravenous administration, gadolinium-DTPA is eliminated almost exclusively via glomerular filtration in the kidneys [20, 21]. Renal excretion remains the primary route of the elimination of gadolinium even in patients with impaired kidney function [22]. In clinical doses, gadolinium chelates are characterized by a very low nephrotoxicity [23, 24, 25, 26] and also a good dialysability [27, 28]. The kinetic properties and the good safety profile recommend gadolinium-DTPA or similar compounds as potential contrast agents for excretory MR urography. However, there exists a general gadolinium-related problem which results from the fact that paramagnetic contrast agents lead to a shortening of both the T1- and T2-relaxation times of fluids and tissues. The T1-shortening effect of gadolinium enhances the signal intensity of the urine on T1-weighted images and is therefore desired. The T1-enhancing effect predominates at low gadolinium concentrations in the urine. Conversely, the T2-effect (or T2*-effect in gradient-echo sequences) of gadolinium is undesirable because it destroys the positive contrast enhancement of the urine and results in a signal void. The T2/T2* effect predominates with increasing gadolinium concentrations. Since the normal kidneys are able to concentrate the excreted gadolinium by a factor of 50±100 [21], the T2/T2* effect is a relevant disturbance source, which has to be minimized in order to achieve optimal contrast enhancement for excretory MR urography. The most effective key to control the problem of endoluminal T2* effects is to combine gadolinium with furosemide [6]. Furosemide is a very powerful diuretic agent that belongs to the subclass of loop diuretics [29]. These agents begin to act immediately after the first pass in the kidneys, usually 10±20 s after intravenous injection. The diuretic effect is explained by the blocking of the Na + /K + /2Cl± symporter in the ascending limb of the loop of Henle which leads to a rapid retention of water inside the tubulus [29]. The glomerular filtration rate remains unaffected by furosemide, but urine volume and flow get markedly increased [29]. In diuretic-enhanced excretory MR urography, the positive interaction of furosemide and gadolinium for achieving optimal contrast enhancement in the urine is manifold [6]: At first, the furosemide-induced increase in urine volume causes a mild distension of the urinary tract. Potentially more important is that the endoluminal retention of fluid especially achieves a dilution of the excreted gadolinium concentration. Moreover, furosemide increases the urine flow markedly, which leads to a rapid and uniform distribution of gadolinium within the entire urinary tract including the caliceal fornices. In fact, ªdilutionº and ªdistributionº of the excreted amount of
359
gadolinium are the two components that best explain the benefits of furosemide for avoiding T2* effects in the gadolinium-enhanced urine. In other words, furosemide optimizes the T1-enhancing properties of gadolinium agents inside the urinary tract [6]. In excretory MR urography, a regular intravenous gadolinium dose of 0.1 mmol per kilogram of body weight (bw) has proved to be well suited for the combination with furosemide [6, 7]. The positive interaction of furosemide and gadolinium fortunately already occurs at low intravenous diuretic doses of 5±10 mg. In order to further optimize the interplay of both drugs, it has been recommended to inject furosemide shortly before the contrast agent [6]. A delay of 1±5 min in between the administration of both drugs has proved to be effective for achieving a rapid and complete enhancement of the urinary tract. It has been shown in experimental and clinical studies that the sole administration of a gadolinium agent without furosemide cannot provide sufficient image quality in excretory MR urography [30, 31]. Theoretically, the infusion of physiologic saline solution may serve as a potential alternative to a diuretic injection. From our own experience, the hydration of the collecting system using exclusively physiologic saline is less controllable and less effective than achievable with furosemide. Furthermore, one may also suppose that the injection of gadolinium doses lower than 0.1 mmol/kg bw can be effective enough to avoid T2* effects without the need for furosemide. Experimentally, the injection of only 0.06 mmol gadolinium per kilogram body weight could not generate a sufficient urographic effect and the supplementation of a diuretic still remained necessary [30]. Nevertheless, lowering the intravenous gadolinium dose makes sense especially to reduce the costs for excretory MR urography. Therefore, dose-optimizing studies are necessary to determine low gadolinium dosages, which allow us to obtain a safe urographic effect for all clinical applications. Feasibility of excretory MR urography with regard to the serum creatinine level The practicability of T1-weighted excretory MR urography is substantially determined by the ability of the kidneys to excrete the intravenously administered gadolinium agent. Using a standard dose of gadoliniumDTPA in combination with low-dose furosemide, a good urographic effect is regularly obtained up to a serum creatinine of 2 mg/dl [7]. This range, fortunately, already includes a large number of patients. The exact value and feasibility of excretory MR urography in patients with serum creatinine levels greater than 2 mg/dl is not yet clarified definitively. The production of a sufficient urine volume seems to determine the technical
success of excretory MR urography in patients with impaired renal function [7]. Here again, the administration of furosemide, which increases the urine volume, becomes an important part of the examination. It is obviously useless to perform excretory MR urography in patients with shrunken kidneys and oliguria. Pulse sequences for excretory MR urography In the clinical practice of excretory MR urography, T1weighted, spoiled, 3D gradient-echo (GRE) sequences with low repetition and echo times have proved to be well suited for imaging of the gadolinium-enhanced urinary tract (Table 1) [6, 7]. Typical brand names of such GRE sequences are T1 fast-field echo (T1-FFE) or fast low-angle shot (FLASH). The sequences are usually performed in the coronal or paracoronal orientation. We recommend varying the field of view between two sizes (Table 1) in order to obtain MR urograms with different anatomic coverage and spatial resolution [6, 7]. A large field of view is preferable for acquisition of survey MR urograms, which display the entire urinary tract with moderate spatial resolution. On the other hand, a smaller field of view should be selected to perform high-resolution detail MR urograms that either focus on the pelvicaliceal systems or on the distal ureters. Maximum intensity projection images are postprocessed from the source images of each 3D sequence data set to obtain an anatomic overview of the urinary tract. In gadolinium excretory MR urography, extraurinary fluid collections cannot degrade the quality of MIP images because the signal of unenhanced intestinal water is always hypointense on T1-weighted images. A spectral fat suppression is not really necessary but allows very good demarcation of the contours of the renal parenchyma [6]. For the compensation of respiratory motion artifacts, which potentially can occur in the pelvicaliceal system [6], the patient should suspend breathing during the acquisition of the MR urographic sequence data [7]. Breath-hold 3D GRE imaging, however, requires the implementation of ultrafast high-performance gradient systems, which allow minimization of the repetition time below 10 ms. With ultrafast gradients, the entire 3D GRE data set, including 50 partitions, can be obtained during a breath-hold of 15±25 s (Table 1). In contrast, on magnets equipped with a conventional gradient system, the acquisition time for a corresponding 3D GRE sequence usually lasts several minutes (Table 1), which requires the use of respiratory gating instead of a breath-hold technique. The effectiveness of respiratory gating, however, depends on a regular and shallow respiration, which presupposes good patient cooperation during the acquisition of the pulse sequences [6, 7]. Furthermore, the gating parameters, such
360
a
b Fig. 2 a, b A 34-year-old female patient with recurrent urinary tract infections. a The maximum intensity projection (MIP) of the T1weighted breath-hold 3D gradient-echo (GRE) sequence demonstrates subtle caliceal deformities and diverticula in the nondilated collecting system on the left side (arrowheads). In the right kidney, a mild dilatation of the upper caliceal group is caused by a vascular compression of the infundibulum (arrow). The crossing vascular segment is visible only because of inflow effects. No extra gadolinium was administered. b The comparison with radiography underscores the good morphologic accuracy achieved with excretory MR urography
as respiratory cycle length, gate delay, and gate width, must be adjusted exactly to the patient's individual respiration curve. Recently, it has been demonstrated for excretory MR urography that breath-hold imaging provides superior compensation of motion artifacts than achievable with respiratory gating [7]. Breath-hold MR imaging, in contrast to respiratory gating, more constantly yields good depiction of subtle morphologic de-
Fig. 3 This is not a radiograph but a T1-weighted projection MR urogram obtained with a 2D GRE sequence, which demonstrates a postoperative follow-up after left-sided pyeloplasty in a 28-yearold male patient. Similar to radiography, the contrast material-enhanced urine displays a certain transparency on GRE projection images
tails of the pelvicaliceal system (Fig. 2) [7]. Whenever ultrafast gradients are available, the breath-hold technique should be preferred. On the other hand, the use of respiratory gating offers the possibility of obtaining excretory MR urograms with good anatomic resolution in small children and even in newborns [11]. In breath-hold excretory MR urography, the 3D GRE technique can be combined with echo-planar imaging (GRE-EPI), which allows further reduction of the breath-hold period down to 15±20 s or even lower [7]. Ultrafast multishot 3D GRE-EPI sequences (Table 1) are especially suggested for use in elderly or critically ill patients who are not able to suspend breathing for more than 20 s. The increase in imaging speed with use of EPI, however, is achieved at the expense of a greater sensitivity of the MR urograms toward magnetic susceptibility effects and, secondly, also at the expense of a decrease in signal-to-noise ratio [7, 32]. Both effects can result in a coarse image texture, which especially becomes apparent on high-resolution detail MR urograms [7]. Comparable to static fluid T2-weighted MR urography, 2D projection imaging techniques are also available for T1-weighted excretory MR urography. A special spin-echo projection technique has been designed for the assessment of the renal excretory function by means of MR imaging in a way comparable to scintigraphy [33, 34]. Recently, another projection imaging technique has been introduced using conventional T1weighted 2D GRE sequences (Table 1) in order to obtain ªquick-shotº excretory MR urograms that achieve
361
Fig. 4 Algorithm for a combination of static-fluid MR urography and excretory MR urography in accordance with the grade of dilatation visible on the T2-weighted baseline image. (*A serum creatinine level of 3 mg/dl is only a practical estimate and not a rigid guideline.)
a good morphologic resolution (Fig. 3) [7]. Due to the short acquisition time below 15 s, 2D GRE projection MR urograms can also be recommended for reducing the breath-hold period in patients unable to cooperate [7]. In addition, projection images also suggest a future potential for the MR guidance of antegrade catheter interventions inside the urinary tract [35]. As is the same with T2-weighted projection images, there is limited value also for T1-weighted projection urograms to visualize small filling defects using a slice thickness of several centimeters [7]. Whenever possible, 3D GRE sequences should be used because subtle pathologic details can be assessed more accurately on the very thin source images of the original 3D data set [7].
Standard MR urographic examination Patient preparation Each patient should be carefully informed about the entire examination procedure. It should also be explained that good cooperation during the acquisition of the breath-hold sequences is of crucial importance for the best image quality achievable. The latest serum creatinine level should be available. Because the injection of low-dose furosemide is intended, the patient should be asked to void before undergoing the examination. Further preparations, such as fasting, hydration, or the application of laxatives, are not required. The patient is placed with elevated arms, head first, in supine position inside the bore of the magnet. A surface coil can be used but is not really necessary in adult patients. Monitoring of the patient's respiration curve throughout the entire examination is obligatory for achieving an effective compensation of respiratory motion artifacts.
Imaging strategies As always in MR imaging, T1- and T2-weighted pulse sequences usually complement one another. The same general guideline refers to the standard MR urographic examination. In our department, we combine static-fluid MR urography with excretory MR urography according to an algorithm which has proved useful for the clinical practice because it primarily orientates according to the grade of dilatation and grossly also to the kidney function (Fig. 4): At the beginning of each MR urographic examination, a fast T2-weighted HASTE projection image is acquired within a few seconds in order to obtain a quick baseline look at the current distension of the urinary tract before injection of furosemide [6]. The degree of distension can be estimated according to the following common grading system [36]: Grade 1: Slight blunting of the fornices and infundibula. The papillary impression is well demarcated in all calices. Grade 2: Pyelocaliectasis with obvious blunting of the fornices. The papillary impression is flattened. Grade 3: Pyelocaliectasis with distinct rounding of the calices and obliteration of the papillary impression. Parenchymal atrophy is often associated but not a prerequisite. Grade 4: Extreme caliceal ballooning, usually associated with marked parenchymal atrophy and severe malfunction or nonfunction. With the baseline information on the latest grade of dilatation together with the knowledge of the serum creatinine level, there are four options to carry out a comprehensive MR urographic examination (Fig. 4). In addition to the baseline image, we also perform a standard axial T2-weighted TSE sequence (TR/TE:
362
MR urograms (Fig. 5) [7]. Additional T1-weighted scans can be added, if necessary. It is also possible to perform delayed MR urograms up to several hours after contrast material injection. Two practical recommendations for obtaining highquality excretory MR urograms are:
Fig. 5 Examination schedule for a combined T2- and T1-weighted MR urography. The time scale refers to patients with a normal serum creatinine level. (*Alternatively, furosemide can be injected 30±60 s before contrast material injection.)
1800/100 ms; TSE factor : 16) of the kidneys. The additional sequence requires only 3±4 min of extra time but allows easily to combine imaging of the upper urinary tract with the assessment of the parenchymal morphology of the kidneys. The examination schedule for a combined T2- and T1-weighted MR urography (option B in Fig. 4) lasts approximately 30 min (Fig. 5). Furosemide is preferably administered using a low-dose of 5±10 mg [6, 7]. In patients with a serum creatinine level greater than 1.5 mg/dl, the dose may be increased to 10±20 mg of furosmide [7]. The diuretic agent can be injected either a few minutes before the multislice HASTE sequence is obtained (Fig. 5) or immediately thereafter. For a good practicability, however, gadolinium-DTPA should be administered not later than 5 min after the furosemide injection. Another 5 min later, a series of at least three T1-weighted breath-hold 3D GRE sequences is obtained including survey and detail
1. Good patient cooperation, which is mandatory especially during breath-hold data acquisition. Before starting the first breath-hold MR sequence, it is advisable that the patient lying inside the magnet be exercised once or twice to suspend breathing for 20±30 s in order to get familiar with this special requirement of the examination. The patient should also be reminded to avoid swallowing during the breath-hold. The quality of the patient's breath-hold can be easily checked with the respiration signal simultaneously displayed on the monitor. 2. With increasing degree of obstruction, the urine flow declines accordingly and the chances for obtaining a uniform gadolinium distribution with the help of furosemide diminish. Inside the pelvicaliceal system, a reduced distribution of contrast material may cause undesired layers of gadolinium and unenhanced urine. Furthermore, the undistributed amount of gadolinium may now induce signal voids due to T2* effects. A simple trick for achieving an additional mechanical mixing of the excreted gadolinium in the urine is to rotate the patient, while lying on the magnet table, several times to the left and right, back and forth. Alternatively, it is also possible to change the patient position from supine to prone. Such easy but effective maneuvers can be done without extra time in between two MR urography sequences (Fig. 6). Aspects of image analysis For interpretation of MR urograms, it is very important to consider that filling defects are often not visible on MIP images of both T2-weighted and T1-weighted sequences. Because all kinds of filling defects display a hypointense signal intensity, they often become suppressed by the reconstruction algorithm of the MIP [6]. This problem especially arises when filling defects are located in the renal pelvis where they are almost completely surrounded by fluid or contrast material [6]. Even large filling defects may be obscured on MIP images, although they are clearly visible on the original source images (Fig. 7). Because of the small diameter of the ureters, filling defects in this location often remain visible even on MIPs [6]. As a general guideline, it has been emphasized by several authors not to rely exclusively on MIP images [6, 13, 16, 17]. It is always necessary to include the original source images in the analysis
363
a
a
b b Fig. 6 a, b A 61-year-old female patient with bilateral grade-II dilatation of the collecting system caused by a large retroperitoneal hematoma. a Ten minutes after contrast material injection, the source images of the 3D GRE sequence show insufficient gadolinium distribution inside the renal pelvis of the right kidney (arrowheads). b After rotating the patient a few times to the right and left, a uniform endoluminal gadolinium enhancement is rapidly achieved even in the markedly dilated renal pelvis with reduced urine flow
of both multislice T2-weighted MR sequences and T1weighted 3D GRE MR urograms.
Applications of MR urography Non-dilated urinary tract and ureterohydronephrosis On unenhanced T2-weighted MR urograms, the nondilated urinary tract is either invisible or incompletely visualized [6, 9, 12, 17, 19, 37], whereas excellent depiction is constantly achieved using gadolinium excretory MR urography [6, 7, 37]. The vast majority of studies investi-
Fig. 7 a, b A 71-year-old male patient with left-sided nephrolithiasis. a The single slice of the T1-weighted 3D GRE sequence clearly shows a 17-mm large calculus inside the renal pelvis (arrow). b The hypointense signal of the filling defect is completely suppressed on the MIP image. This case underscores the importance of the source images, which always have to be included in the analysis of MR urograms
gating the value of T2-weighted static-fluid MR urography refrain from any diuretic injection [1, 2, 3, 4, 5, 8, 9, 12, 15, 16, 17]. However, the administration of furosemide helps to improve the visibility of the nondilated urinary tract also in T2-weighted static-fluid MR urography [13, 37]. Interestingly, a low-dose of 10 mg of furosemide seems to be less effective for T2-weighted MR urography than attainable in combination with gadolinium in T1weighted MR urography [37]. Even an intravenous dose of 20 mg of furosemide cannot safely achieve a sufficient visualization of the nondilated collecting system on T2weighted MR urograms [19]. It is certainly a mistake to assume that a nondilated urinary tract must always be nonpathologic. Especially early-stage urinary tract disorders often do not yet cause an obstruction and also filling defects do not have to be associated with a dilatation when they are located inside the pelvicaliceal system (Fig. 7).
364
a
b
c
d Fig. 8 a±d A 54-year-old woman with an extrinsic ureteral stenosis caused by a metastasis of an ovarian carcinoma. a The anteroposterior MIP from excretory MR urography demonstrates the extension of the stenosis. The obstruction is incomplete because gadolinium-enhanced urine can pass the stenosis (arrow). b The near lateral MIP image shows a displacement of the ureter (arrow) and even a very thin stream of contrast-enhanced urine is visible within the stenosis (arrowheads). c Static fluid MR urography is less accurate in the visualization of the morphology of the stenosis. d The axial standard T2-weighted TSE sequence directly visualizes the metastasis (asterisk) but cannot demonstrate the ureteral stenosis itself
It has been confirmed with many studies that MR urography easily allows diagnosis of a dilatation of the urinary tract and determination of the level of obstruction reaching a sensitivity and specificity of up to 100 % [3, 6, 8, 12, 13, 16, 17]. Static-fluid MR urography is especially superior to excretory MR urography in patients with a chronic obstruction and marked impairment of the renal excretory function [6, 37]. With use of T2-
weighted sequences, it even becomes possible to obtain MR urograms of a non-excreting grade-4 hydronephrosis (Fig. 1) [3, 6, 8, 12, 14, 17, 37]. Preliminary results indicate that in incompletely obstructed urinary tracts with mild or moderate dilatation (grades 1 and 2), unenhanced T2-weighted MR urograms tend to show only the site of the obstruction simulating a complete discontinuation in the ureteral course [37]. The ureter appears ªamputatedº on T2-weighted images, although the obstruction is incomplete (Fig. 8) [37]. In gadolinium-enhanced T1-weighted MR urography, incompletely obstructing ureteral filling defects allow the contrast material to pass the stenosis. Consequently, there is gadolinium-enhanced urine before and behind the obstruction, and it may also be possible to see a fine stream of contrast material within the stenosis (Fig. 8 b) [6]. In moderately dilated urinary tracts, the morphology and extension of ureteral filling defects or stenoses is accurately demonstrated with high contrast and spatial resolution using preferably excretory MR urography (Fig. 8) [6, 37]. With increasing obstruction (grades 3
365
Fig. 9 a, b A 64-year-old woman with agenesis of the right kidney and a large parapelvic cyst in the left renal sinus causing a diverticulum-like ballooning of the upper calix. a The MIP from T2-weighted HASTE sequence shows a severe dilatation of the left-sided upper calices without revealing the cause. b Two MIP images from excretory MR urography disclose the true pathology. The left part of b displays anteroposterior MIP with gadoliniumenhancing dilated calix (arrowhead) and non-enhancing cyst (arrow). The right part of b more clearly demonstrates the severe compression of the caliceal infundibulum (arrow)
9a and 4) and diminishing excretory function, T2-weighted static-fluid MR urography becomes the preferred technique (Fig. 1) [6, 37]. We should be aware, however, that gadolinium excretory MR urography can provide urograms of high morphologic detail even in patients with markedly dilated urinary tracts, provided that the renal excretory function allows a sufficient gadolinium distribution. Current MR urography techniques do not yet allow a safe and direct differentiation between obstructive ureterohydronephrosis and dilatation due to vesicoureteral reflux [2, 5, 16]. Gadolinium-enhanced voiding cystourethrography carried out under MR imaging control is, at least, technically feasible but appears impractical for the clinical routine because of the lengthy in-bore examination times of the patients [38]. Anomalies of the urinary tract Virtually all types of urinary tract anomalies can be diagnosed with MR urography. The early detection of anomalies is especially important in pediatric radiology [2, 11, 39, 40]. Moreover, anomalies of the urinary tract anatomy are also common findings in MR urography of adult patients [6, 13, 14, 15, 17]. T1- and T2-weighted sequences are again complementary imaging techniques: T2-weighted static-fluid MR urography has proved useful for the visualization of anomalies associated with an increased amount of fluid such as large
9b ureteroceles [2, 15], bladder diverticula [14], markedly obstructed duplex systems [2, 13, 14, 17], megaureters [2, 39], stenoses of the ureteropelvic junction (UPJ) [14, 17, 39, 40], and cystic kidney diseases [2, 13, 14, 39, 40]. In T2-weighted MR urography, however, the presence of parapelvic cysts can simulate a hydronephrosis or partial pyelocaliectasis (Fig. 9) [7, 8, 15]. In those cases, T1-weighted excretory MR urography is very well suited to demonstrate the true anatomy of the pelvicaliceal system with typically displaced and partly compressed calices (Fig. 9 b) [7, 37]. From our experience, diureticenhanced gadolinium excretory MR urography suggests a great potential for the imaging of small anatomic variations and anomalies that occur in the nondilated urinary tract such as pelvicaliceal diverticula (Fig. 2), ectopic ureters [11], nondilated duplex systems [11], and also tubular ectasias in a medullary sponge kidney (Fig. 10). T2- and T1-weighted MR urography techniques are also able to demonstrate the urinary tract of dystopic kidneys and horseshoe kidneys [2, 6, 11]. Urolithiasis In the era of helical CT, the acute stone colic should not be a primary indication for MR urography. However, because urolithiasis is the most common differential diagnosis in all kinds of urinary tract disorders, it is definitely important to be aware of the findings of stones in MR urography. Although several studies could demon-
366
Fig. 10 a, b A 30-year-old female patient with a medullary sponge kidney. a The T1weighted detail MR urogram demonstrates the classic sign of a ªbunch of flowersº (arrow) caused by multiple dilated collecting ducts filled with contrast material. Furosemide avoids T2*-effects within the ectatic ducts. Excretory MR urography cannot achieve the same high image resolution as obtained with radiography in b. The nephrostomy tube visible in b (arrow) had been removed at the time the MR urogram was performed
10 a
Fig. 11 T1-weighted detail MR urogram of a 26-year-old woman with left-sided nephrolithiasis. Lithotripsy had failed to disintegrate the calculus. The MIP image shows a calculus inside the nondilated inferior caliceal group (arrowhead). Magnetic resonance urography especially provides the information that the calculus can potentially obstruct the narrow bifid renal pelvis (long arrow). The tip of a double-J stent is seen in the upper infundibulum (short arrow). The gadolinium-filled catheter shaft is invisible on the MIP
10 b strate that calculi can be detected with use of static-fluid MR urography [3, 5, 12, 16, 17] and excretory MR urography [6, 7], the strength of the current numerical data is weak because of the limited number of cases investigated thus far. Since it is not possible to identify calcifications on MR images in the way we are used to when reading radiographs or CTs, the MR urographic diagnosis is exclusively based on the detection of typical filling defects. In both static fluid and excretory MR urography, most of the stones cause round or branched endoluminal signal voids (Figs. 7, 11) [3, 6, 12, 16, 17]. As mentioned previously, those filling defects must be found by analysis of the original thin source images of either a T2-weighted multislice sequence or T1-weighted 3D GRE sequence. Nevertheless, the search for filling defects is somewhat unspecific and the differentiation between a calculus and a blood clot, a surgical clip, debris, or a small tumor may occasionally remain difficult because of their similar MR urographic appearance [3, 6, 12, 16]. Together with the presence of the clinical data, however, it should be possible to make the correct MR urographic diagnosis of urolithiasis in the majority of patients [6, 7]. It has also been recommended to include a single plain-film radiograph in the analysis of MR urograms for better detection of calcifications [6, 16]. Apart from its limited value in acute urolithiasis, MR urography may play a potential role in patients with chronic urolithiasis, in whom neither ultrasonography nor CT can sufficiently explain the complicated state of the chronically affected urinary tract. The MR urography not only helps to avoid radiation exposure due to
367
repeated CT examinations. Especially in chronic nephrolithiasis resistant to treatment, MR urography can be used to obtain valuable information about the true location of calculi and the anatomy of the stone-bearing calix [7]. Those data are clinically relevant to determine whether a stone has a potential chance to pass the infundibulum and renal pelvis after lithotripsy or during endo-urologic stone removal (Fig. 11) [7]. Intrinsic tumor disease in the upper urinary tract The vast majority of intrinsic tumors located in the pelvicaliceal system or ureter are, histologically, transitional cell carcinomas of the papillary type [41]. The particular propensity to develop multicentric tumor lesions in the upper urinary tract in up to 40 % of patients [41, 42] emphasizes the importance of imaging techniques that enable a comprehensive bilateral assessment of the kidneys and collecting systems. Computed tomography has already proved to be an excellent diagnostic modality for the imaging of uroepithelial tumors located in the pelvicaliceal system [41, 42, 43, 44]. Since retrograde pyelography is commonly regarded as the mainstay in the diagnostic range of ureteral transitional cell carcinomas [42], probably one of the greatest challenges for MR urography is to provide a noninvasive imaging tool for the recognition especially of small nonobstructing ureteral tumors. Indeed, initial results suggest a potential for MR urography in reducing the total need for invasive retrograde pyelography in the assessment of uroepithelial tumor diseases [6]. T2-weighted static-fluid MR urography has already proved its usefulness for imaging the nonfunctioning urinary tract which is severely obstructed by an advanced-stage intrinsic mass [6, 8, 13]. T1-weighted excretory MR urography allows to obtain high-resolution images of the tumor morphology and extension provided that the excretory function is not quiescent [6]. Although several studies have documented that advanced intrinsic tumor growth can be diagnosed with static-fluid and excretory MR urography [5, 6, 13, 14, 17], there currently exist no data describing the sensitivity and specificity of MR urography for the detection of early tumor disease. The signs of endoluminal tumor growth described for standard intravenous urography and retrograde pyelography [41, 42] also apply to the morphologic features found in MR urography: Transitional cell carcinomas often present as a sessile filling defect with either smooth polypoid, or more commonly, irregular exophytic configuration [41, 42] and hypointense signal intensity on MR urography sequences [6, 8, 13]. Characteristic urographic findings are [42]: ballooning of a tumor-filled calix (ªoncocalixº), obstruction of a caliceal infundibulum with preserved excretion of the affected
a
b Fig. 12 a, b Magnetic resonance urography and standard MR pulse sequence in a 72-year-old man with a transitional cell carcinoma in the renal pelvis. a The MIP from excretory MR urography shows a large tumor in the right renal pelvis. The calices appear ªamputatedº (arrows). b A soft tissue mass with heterogenous signal intensity (black arrow) is seen on the axial standard T2-weighted TSE image
calix (ªamputationº; Fig. 12) or ceased function (ªphantom calixº), the ªsipple signº (Fig. 13), and finally, the ªgoblet signº in the ureter (Fig. 13). The latter finding refers to a characteristic accumulation of contrast material in a prestenotic cup-shaped ureteral dilatation, which is seen right inferior to a tumor obstruction in retrograde pyelography or just proximal to the tumor in conventional urography [42, 45] and also MR urography (Fig. 13) [6]. In patients suspected of having a transitional cell carcinoma, there is a good opportunity to combine MR urography with standard MR imaging in the axial plane [5, 6, 13], thus offering a potential alternative to CT. On unenhanced T1-weighted MR images of standard spinecho sequences, transitional cell carcinomas are mainly isointense to the normal renal parenchyma, whereas they display a more heterogeneous signal morphology including hyperintense and cystic-like areas on T2-
368
Extrinsic obstruction of the urinary tract
a
b Fig. 13 a, b Transitional cell carcinoma of the right ureter in a 62year-old male patient. a Maximum intensity projection from T2weighted HASTE sequence demonstrates the classic ªgoblet signº (arrowhead) and ªsipple signº (arrow). b Corresponding retrograde pyelography is unable to visualize the entire tumor extension
weighted images (Fig. 12). According to the limited contrast-material enhancement of transitional cell carcinomas in CT [43], we also expect a moderate gadolinium enhancement of such tumors on axial T1-weighted spin-echo images. Standard spin-echo pulse sequences can help to distinguish a central soft tissue mass from blood clots or calculi, which display a signal void on T1and T2-weighted spin-echo images and do not enhance with gadolinium [5, 6]. Conventional pulse sequences may also prove useful for the assessment of a parenchymal tumor infiltration or, vice versa, for the differentiation of an intrinsic tumor from a renal cell carcinoma with secondary invasion of the collecting system; however, statistical data concerning the value of MR urography in combination with conventional MR imaging in the diagnosis of transitional cell carcinomas are not yet available.
There are numerous pathologic entities that potentially can involve the urinary tract extrinsically, such as multiple types of retroperitoneal malignancies, aortic aneurysms and y-grafts, retroperitoneal fibrosis, inflammatory diseases, postoperative or traumatic disorders, and anomalies. Typical MR urographic findings that are highly suggestive of extrinsic affection of the urinary tract are a partial or complete displacement and a compression of the tract mainly associated with different degrees of ureterohydronephrosis (Fig. 8) [6]. The morphology of a ureteral stenosis seen with extrinsic compression is different from that observed in transitional cell carcinomas. A filling defect is missing. A goblet sign is not seen. The extrinsic compression of the lumen usually appears as a concentric stricture either with gradual tapering of the ureteral wall [3, 17, 46] or abrupt reduction of the ureteric caliber (Fig. 8). Static-fluid MR urography, again, is useful for the detection of the level and degree of obstruction; however, differentiation between the intrinsic or extrinsic cause of a urereteric stenosis is, at times, difficult if T2weighted MR urography is obtained solely [3, 8, 12, 17]. In patients with a mild or moderate extrinsic obstruction, T2-weighted MR urograms may display a stop of the ureteral fluid column at the site of obstruction simulating a complete occlusion (Fig. 8 c) [37]. In these patients, gadolinium excretory MR urography provides a more accurate assessment of the morphologic features of such extrinsic stenoses including the depiction of the exact extension and the additional functional information that the occlusion is frequently incomplete (Fig. 8 a,b) [6, 37]. In analogy to intrinsic tumor diseases, the combination of MR urography and conventional axial MR pulse sequences offers a ªone-stop shoppingº diagnostic procedure, which is well suited for the assessment of extrinsic urinary tract disorders [6, 13, 17]. Miscellaneous Both static-fluid MR urography and gadolinium excretory MR urography have already proved to be excellent tools for the assessment of the urinary tract of transplant kidneys [34, 47]. Furthermore, especially gadolinium excretory MR urography is a very accurate imaging technique for the detection of urine leaks and fistulas (Fig. 14), and also for the assessment of the urinary tract after orthotopic neobladder reconstruction (Fig. 15) [7, 37]. Occasionally, it is also practical to combine MR urography with MR angiography, e. g. in the preoperative assessment of potential renal transplant donors [10, 31], or in patients with arterial compression of the UPJ (Fig. 16). In dynamic functional MR imaging of the kidneys, designated as MR renography or MR nephrogra-
369
14 a
14 b
14 c
Fig. 14 a±c A 48-year-old woman who underwent kidney re-transplantation and presented postoperatively with increasing flank pain. a Static-fluid MR urography demonstrates an unspecific fluid collection in the lower abdomen (arrows). b Maximum intensity projection from excretory MR urography reveals leakage of gadolinium-enhanced urine. c The exact site of the ureteral perforation is identified on the thin T1-weighted source image (arrow). Parenchymal infarction is also visible (arrowheads)
15 phy, the exclusive glomerular filtration of gadolinium agents can also be used to derive bilateral time±intensity curves of the kidneys comparable to renal scintigraphy [20, 33, 34, 39, 40, 47, 48, 49, 50].
Current status and future perspectives With respect to the current discussion on cost restraints in medical care, it should be clear that MR urography is regarded mainly as a diagnostic test of secondary preference behind ultrasonography, IV pyelography, and CT. However, the diagnostic utility of MR urography is certainly much better than only being a procedure of the second or third choice. Theoretically, MR urography has the potential to become an economic, time-effective, and radiation-saving imaging modality if perform-
Fig. 15 Maximum intensity projection images from excretory MR urography in a 66-year-old woman who underwent cystectomy with urinary diversion via an ileal conduit. The left part of the image shows an anteroposterior view with dilatation of the right-sided urinary tract due to a stenosis of the ureteral anastomosis to the conduit (arrow). The right part demonstrates the anatomy of the ileal conduit from the lateral view (arrowhead)
ed in conjunction with standard pulse sequences, MR nephrography, or MR angiography. This integrative aspect of MR imaging in uroradiology can help to avoid multiple separate diagnostic procedures, which in the sum are obviously costly, time-consuming, and sometimes even invasive. The use of MR urography should not be confined to patients with known intolerance of iodinated contrast agents. Review of the current studies substantiates that T2-weighted static-fluid MR urography and T1-weighted excretory MR urography are complementary imaging tools that can be easily combined to obtain a comprehensive and noninvasive diagnostic test applicable in almost all urinary tract disorders. Especially in the imaging of intrinsic and extrinsic neoplastic ureteral diseases, MR urography suggests the potential to reduce the total number of necessary retrograde pyelographies
370
[6]. In suspected ureteral diseases, MR urography should be obtained before retrograde pyelography or ureterorenoscopy are considered. Although MR urography appears to be less relevant in acute urolithiasis, it can provide valuable clinical information in patients with chronic stone disease [7]. Both static-fluid and excretory MR urography will certainly become very important imaging modalities in pediatric uroradiology [2, 11, 39, 40]. In our department, MR urography has already replaced conventional urography in children. The use of exclusively unenhanced T2-weighted pulse sequences makes particular sense in nonfunctioning hydronephrotic kidneys [3, 8, 12, 14,17] and in obstructive uropathy during pregnancy [4, 12, 17,51]. In the not too distant future, MR urography may also form the basis for such interesting applications as virtual ureterorenoscopy [46], MR-guided percutaneous nephrostomy [35], and MR micturating cystourethrography [38].
a
b Fig. 16 a, b Combination of MR angiography and MR urography in a 40-year-old woman with obstruction of the ureteropelvic junction caused by a vascular compression. a Magnetic resonance angiography obtained with a conventional 3D GRE sequence during the first pass of only a single-dose of gadolinium-DTPA reveals an aberrant renal artery crossing the ureteropelvic junction (arrow). b During the subsequent excretory phase, MR urography depicts the exact length of the ureteral stenosis (arrow) and provides the functional information that the obstruction is incomplete
References 1. Friedburg HG, Hennig J, Frankenschmidt A (1987) RARE-MR urography: imaging of the urinary tract with a new fast nontomographic MR technique. Radiologe 27: 45±47 2. Sigmund G, Stoever B, Zimmerhackl LB, Frankenschmidt A, Nitzsche E, Leititis JU, Struwe FE, Hennig J (1991) RARE-MR-urography in the diagnosis of upper urinary tract abnormalities in children. Pediatr Radiol 21: 416±420
3. Roy C, Saussine C, Jahn C, Vinee P, Beaujeux R, Campos M, Gounot D, Chambron J (1994) Evaluation of RARE-MR urography in the assessment of ureterohydronephrosis. J Comput Assist Tomogr 18: 601±608 4. Roy C, Saussine C, Le Bras Y, Delepaul B, Jahn C, Steichen G, Jacqmin D, Chambron J (1996) Assessment of painful ureterohydronephrosis during pregnancy by MR urography. Eur Radiol 6: 334±338
5. Catalano C, Pavone P, Laghi A, Scipioni A, Panebianco V, Brillo R, Fraioli F, Passariello R (1999) MR pyelography and conventional MR imaging in urinary tract obstruction. Acta Radiol 40: 198±202 6. Nolte-Ernsting CCA, Bücker A, Adam GB, Neuerburg JM, Jung P, Hunter DW, Jakse G, Günther RW (1998) Gadolinium-enhanced excretory MR urography after low-dose diuretic injection: Comparison with conventional excretory urography. Radiology 209: 147±157
371
7. Nolte-Ernsting CCA, Tacke J, Adam GB, Haage P, Jung P, Jakse G, Günther RW (2001) Diuretic-enhanced gadolinium excretory MR urography: comparison of conventional gradient-echo sequences and echo-planar imaging. Eur Radiol 11: 18±27 8. Reuther G, Kiefer B, Wandl E (1997) Visualization of urinary tract dilatation: value of single-shot MR urography. Eur Radiol 7: 1276±1281 9. Balci NC, Mueller-Lisse UG, Holzknecht N, Gauger J, Waidlich R, Reiser M (1998) Breathhold MR urography: comparison between HASTE and RARE in healthy volunteers. Eur Radiol 8: 925±932 10. Low RN, Martinez AG, Steinberg SM, Alzate GD, Kortman KE, Bower BB, Dwyer WJ, Prince SK (1998) Potential renal transplant donors: evaluation with gadolinium-enhanced MR angiography and MR urography. Radiology 207: 165±172 11. Staatz G, Nolte-Ernsting CCA, Adam GB, Hübner D, Rohrmann D, Stollbrink C, Günther RW (2000) Feasibility and utility of respiratory gated gadolinium-enhanced T1-weighted magnetic resonance urography in children. Invest Radiol 35: 504±512 12. Regan F, Bohlman ME, Khazan R, Rodriguez R, Schultze-Haakh H (1996) MR urography using HASTE imaging in the assessment of ureteric obstruction. AJR 167: 1115±1120 13. Rothpearl A, Frager D, Subramanian A, Bashist B, Baer J, Kay C, Cooke K, Raia C (1995) MR urography: technique and application. Radiology 194: 125±130 14. Klein LT, Frager D, Subramanium A, Lowe FC (1998) Use of magnetic resonance urography. Urology 52: 602±608 15. Aerts P, Van Hoe L, Bosmans H, Oyen R, Marchal G, Baert AL (1996) Breathhold MR urography using the HASTE technique. AJR 166: 543±545 16. O'Malley ME, Soto JA, Yucel EK, Hussain S (1997) MR urography: evaluation of a three-dimensional fast spinecho technique in patients with hydronephrosis. AJR 168: 387±392 17. Tang Y, Yamashita Y, Namimoto T, Abe Y, Nishiharu T, Sumi S, Takahashi M (1996) The value of MR urography that uses HASTE sequences to reveal urinary tract disorders. AJR 167: 1497±1502 18. Hennig J, Nauerth A, Friedburg H (1986) RARE imaging: a fast imaging method for clinical MR. Magn Reson Med 3: 823±833
19. Hattery RR, King BF (1995) Technique and application of MR urography. Radiology 194: 25±27 20. Knesplova L, Krestin GP (1998) Magnetic resonance in the assessment of renal function. Eur Radiol 8: 201±211 21. Krestin GP (1999) Genitourinary MR: kidneys and adrenal glands. Eur Radiol 9: 1705±1714 22. Schuhmann-Giampieri G, Krestin G (1991) Pharmacokinetics of Gd-DTPA in patients with chronic renal failure. Invest Radiol 26: 975±979 23. Niendorf HP, Haustein J, Cornelius I, Alhassan A, Clauss W (1991) Safety of gadolinium-DTPA: extended clinical experience. Magn Reson Med 22: 222±228 24. Haustein J, Niendorf HP, Krestin G, Louton T, Schuhmann-Giampieri G, Clauss W, Junge W (1992) Renal tolerance of gadolinium-DTPA/dimeglumine in patients with chronic renal failure. Invest Radiol 27: 153±156 25. Arsenault TM, King BF, Marsh JW Jr, Goodman JA, Weaver AL, Wood CP, Ehman RL (1996) Systemic gadolinium toxicity in patients with renal insufficiency and renal failure: retrospective analysis of an initial experience. Mayo Clin Proc 71: 1150±1154 26. Swan SK, Lambrecht LJ, Townsend R, Davies BE, McCloud S, Parker JR, Bensel K, LaFrance ND (1999) Safety and pharmacokinetic profile of gadobenate dimeglumine in subjects with renal impairment. Invest Radiol 34: 443±448 27. Haustein J, Schuhmann-Giampieri G (1990) Elimination of Gd-DTPA by means of hemodialysis. Eur J Radiol 11: 227±229 28. Ueda J, Furukawa T, Yamamoto T, Sakaguchi K, Araki Y (1998) Permeability of gadolinium-DTPA through two types of hemodialysis membrane. Invest Radiol 10: 734±737 29. Jackson EK (1996) Diuretics. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Goodman Gilman A (eds) Goodman & Gilman's The pharmacological basis of therapeutics, 9th edn. McGraw-Hill, New York, pp 691, 692, 697±701 30. Nolte-Ernsting C, Adam G, Bücker A, Berges S, Bjùrnerud A, Günther RW (1997) Contrast-enhanced magnetic resonance urography: first experimental results with a polymeric gadolinium bloodpool agent. Invest Radiol 32: 418±423
31. Steinborn M, Wintersperger BJ, Heck A, Theodorakis J, Waggershauser T, Hillebrand GF, Reiser M (1999) Contrast-enhanced MR angiography in the preoperative evaluation of living kidney donors. Fortschr Röntgenstr 171: 313±318 32. Edelman RR, Wielopolski P, Schmitt F (1994) Echo-planar MR imaging. Radiology 192: 600±612 33. Schad LR, Semmler W, Knopp MV, Deimling M, Weinmann H-J, Lorenz WJ (1993) Preliminary evaluation: magnetic resonance of urography using a saturation inversion projection spinecho sequence. Magn Reson Imaging 11: 319±327 34. Knopp MV, Dörsam J, Oesingmann N, Piesche S, Hawighorst H, Wiesel M, Schad LR, van Kaick G (1997) Functional MR urography in patients with renal transplants. Radiologe 37: 233±238 35. Nolte-Ernsting CCA, Bücker A, Neuerburg JM, Glowinski A, Adam G, Günther RW (1999) MR imaging-guided percutaneous nephrostomy and use of MR compatible catheters in the nondilated porcine urinary tract. J Vasc Intervent Radiol 10: 1305±1314 36. Talner LB (1990) Obstructive uropathy. In: Pollack HM (ed) Clinical urography. Saunders, Philadelphia, pp 1535±1536 37. Nolte-Ernsting CCA, Tacke J, Staatz G, Bücker A, Haage G, Adam GB, Jung P, Jackse G, Günther RW (2000) Comparison of gadolinium-enhanced T1weighted excretory MR urography and static fluid T2-weighted MR urography. Twelfth European Congress of Radiology ECR 2000. Eur Radiol 10 (Suppl 1):212 38. Nolte-Ernsting CCA, Glowinski A, Katterbach FJ, Adam G, Rasche V, Günther RW (1998) Micturating cystourethrography by MR imaging using radial k-space sampling. Fortschr Röntgenstr 168: 385±389 39. Borthne A, Nordshus T, Reiseter T, Geitung JT, Gjesdal KI, Babovic A, Bjerre A, Loe B (1999) MR urography: the future gold standard in paediatric urogenital imaging? Pediatr Radiol 29: 694±701 40. Borthne A, Pierre-Jerome C, Nordshus T, Reiseter T (2000) MR urography in children: current status and future development. Eur Radiol 10: 503±511 41. Leder RA, Reed Dunnick N (1990) Transitional cell carcinoma of the pelvicalices and ureter. AJR 155: 713±722
372
42. Goldman SM, Gatewood OMB (1990) Neoplasms of the renal collecting system, pelvis, and ureters. In: Pollack HM (ed) Clinical urography. Saunders, Philadelphia, pp 1292±1352 43. Urban BA, Buckley J, Soyer P, Scherrer A, Fishman EK (1997) CT appearance of transitional cell carcinoma of the renal pelvis. Part 1. Early-stage disease. AJR 169: 157±161 44. Urban BA, Buckley J, Soyer P, Scherrer A, Fishman EK (1997) CT appearance of transitional cell carcinoma of the renal pelvis. Part 2. Advanced-stage disease. AJR 169: 163±168 45. Daniels RE III (1999) The goblet sign. Radiology 210: 737±738
46. Nolte-Ernsting CCA, Krombach G, Staatz G, Kilbinger M, Adam GB, Günther RW (1999) Virtual endoscopy of the upper urinary tract based on contrast material-enhanced MR urography data sets. Fortschr Röntgenstr 170: 550±556 47. Dupas B, Bach-Gansmo T, Blancho G, Franconi JM, Havet T, LØautØ F (1999) Gadolinium-enhanced MR imaging of normal renal transplants. Acta Radiol 40: 250±254 48. Laissy JP, Faraggi M, Lebtahi R, Soyer P, Brillet G, MØry J-P, Menu Y, Le Guludec D (1994) Functional evaluation of norml and ischemic kidney by means of gadolinium-DOTA enhanced turboFLASH MR imaging: a preliminary comparison with 99 mTc-MAG3 dynamic scintigraphy. Magn Reson Imaging 12: 413±419
49. Grenier N, Trillaud H, Combe C, Degrze P, Jeandot R, Gosse P, Douws C, Palussire J (1996) Diagnosis of renovascular hypertension: feasibility of captopril-sensitized dynamic MR imaging and comparison with captopril scintigraphy. AJR 166: 835±843 50. Taylor J, Summers PE, Keevil SF, Saks AM, Diskin J, Hilton PJ, Ayers AB (1997) Magnetic resonance renography: optimization of pulse sequence parameters and Gd-DTPA dose, and comparison with radionuclide renography. Magn Reson Imaging 15: 637±649 51. Friedburg HG, Westenfelder M, Roeren T, Hennig J (1987) MR of the urinary tract in pregnancy. Fortschr Röntgenstr 147: 430±432
