| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, New York University Medical Center, 530 First Ave., HCC Basement-MRI, New York, NY 10016.
Received June 4, 1999;
accepted after revision July 14, 1999.
Address correspondence to V. S. Lee.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Thirty-five patients referred for MR angiography of renal arteries underwent non-breath-hold oblique sagittal velocity-encoded phase-contrast MR imaging through both renal hila (TR/TE, 24/5; flip angle, 30°; signal averages, two; encoding velocity, 75 cm/sec) before and after IV administrastion of an ACE inhibitor (enalaprilat). We analyzed velocity waveforms using established Doppler sonographic criteria. A timing examination with a test bolus of gadolinium contrast material was performed to ensure optimal arterial enhancement during breath-hold gadolinium-enhanced three-dimensional gradient-echo MR angiography.
RESULTS. MR phase-contrast waveform pattern analysis was 50% (9/18) sensitive and 78% (40/51) specific for the detection of renal artery stenosis equal to or greater than 60% as shown on MR angiography. Sensitivity (67%, 12/18) and specificity (84%, 42/50) increased slightly, but not significantly, after IV administration of an ACE inhibitor. Also, the accuracy of quantitative criteria such as acceleration time and acceleration index did not improve after the administration of ACE inhibitor.
CONCLUSION. Renal hilar velocity waveforms, measured using non-breath-hold MR phase-contrast techniques with or without an ACE inhibitor, are insufficiently accurate to use in predicting renal artery stenosis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Conventional angiography is the standard in the diagnosis of renovascular disease. However, the technique has inherent diagnostic limitations and disadvantages such as invasiveness, the use of ionizing radiation, and the risks of reactions to contrast material. The location and distribution of renal artery ostia vary. In approximately 10% of patients, errors in projectional angles for angiography can obscure the proximal renal arteries [3]. Also, frontal view angiography may cause an underestimation of en face atheromatous plaques. Furthermore, the presence of a stenosis alone does not indicate sufficient hypoperfusion to cause renal ischemia and renovascular disease [1]. Patients with essential hypertension are prone to accelerated atherosclerosis and subsequent renal artery stenosis. Nearly half of normotensive patients more than 60 years old have atherosclerotic lesions in their renal arteries [4].
Physiologic tests that assess the functional significance of a narrowed renal artery may better predict a patient's response to revascularization than to anatomic tests. However, tests such as renal scintigraphy and Doppler sonography cannot provide the anatomic information needed for interventional planning and for assessing a lesion's amenability to angioplasty. Until recently, no imaging technique has had the capacity to provide both an anatomic and a physiologic evaluation of renovascular disease. Recent advances in MR imaging suggest that it may have the potential to provide both.
With improvements in software and hardware, three-dimensional (3D) MR angiography can be performed during a single breath-hold [5, 6, 7, 8]. MR angiography has distinct advantages over conventional contrast angiography: it is less invasive, it does not require ionizing radiation, and the contrast agent has minimal to no nephrotoxicity [9, 10]. The near-isotropic 3D MR angiograms can be reconstructed to facilitate the evaluation of tortuous vessels and eccentric plaques. An additional advantage of MR imaging is its versatility; pulse sequences can be designed to measure physiologic data. Specifically, velocity waveforms in the renal artery, akin to Doppler sonographic tracings, can be measured noninvasively using phase-contrast velocity-encoded MR imaging. Phase-contrast MR imaging of renal artery velocity waveforms has shown correlation with measurements obtained from implanted Doppler sonographic probes [11]. Moreover, the location and angle of velocity measurements using MR imaging can be proscribed easily regardless of body habitus and other factors that limit routine sonographic methods.
Renal artery stenosis is believed to cause dampening of normal velocity waveforms measured in the distal renal artery, resulting in a decreased and delayed systolic upstroke (pulsus parvus et tardus). Several groups [12, 13, 14, 15] have reported that Doppler sonographic measurements of flow-velocity waveforms in renal hilar or intrarenal vessels can be used to diagnose renovascular disease with high accuracy. Recently, one group [16] advocated the use of an angiotensin-converting enzyme (ACE) inhibitor to improve the accuracy of Doppler waveform analysis. In their study of 71 hypertensive patients, Doppler sonography of segmental arteries was performed before and after the administration of oral captopril. The authors reported a sensitivity and specificity for 50% or greater renal artery stenosis that increased from 81% (39/48) and 98% (85/87), respectively, to 100% (26/26 and 70/70, respectively) for both. The authors hypothesized that ACE inhibitors, by blocking angiotensin II-mediated vasoconstriction of afferent and efferent intrarenal arterioles, cause an exaggerated pulsus parvus et tardus appearance to arterial waveforms in patients with renal artery stenosis.
We investigated the usefulness of phase-contrast MR imaging in measuring renal artery velocity waveforms as an adjunct to MR angiography for the assessment of renal artery stenosis. Also, in an attempt to validate recent results in the sonography literature [16], we examined whether the administration of an ACE inhibitor improves the diagnostic accuracy of waveform analysis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our patients' average systolic and diastolic blood pressures before the examination were 166 ± 28 mm Hg and 90 ± 23 mm Hg, respectively. The following diagnostic criteria were used to indicate an increased risk of renovascular hypertension: hypertension refractory to medication (n = 24); disease onset at less than 20 years or greater than 50 years old (n = 15); peripheral, coronary, or carotid atherosclerotic disease (n = 15); sudden onset or worsening of hypertension (n = 14); abdominal or flank bruit (n = 4); and hypertensive retinopathy (n = 3) [1]. Most patients (n = 21) met more than one diagnostic criteria. For 14 patients (40%), an ACE inhibitor was part of the routine medicine regimen at the time of examination and continued for this study.
MR Imaging Protocol
All patients were instructed to drink two cups of water (to minimize risks
of hypotension after ACE inhibitor administration) and to void just before the
study. A 22-gauge IV catheter was placed in an arm vein and attached to a
power injector (Spectris; Medrad, Pittsburgh, PA). An MR-compatible blood
pressure device (Multigas Monitor 9500; MR Equipment, Bay Shore, NY) was used
to record blood pressure at baseline and after ACE inhibitor injection with
the cuff placed on the upper part of the arm contralateral to the IV catheter.
MR-compatible ECG leads were placed on the patient's back before the
examination.
Patients were examined using a 1.5-T system (Vision; Siemens, Erlangen, Germany) with high-performance gradients (maximum gradient strength, 25 mT/m; rise time, 600 µsec). A torso phased array coil was used for all studies, and the patient's arms were propped up at the sides using small cushions to minimize wraparound artifacts during imaging.
The imaging protocol and sequence details were as follows. After routine axial breath-hold T1-weighted gradient-echo imaging through the kidneys and adrenal glands, oblique sagittal ECG-gated phase-contrast acquisitions at the level of the renal hila bilaterally were obtained both before and at least 15 min after slow IV injection of enalaprilat (Vasotec; Merck, West Point, PA) at a dose of 0.04 mg/kg (up to a maximum of 2.5 mg, injected over 3-4 min). Blood pressure and pulse were recorded before injection and at 5- to 10-min intervals thereafter for at least 40 min or until blood pressure returned to within 10% of baseline. Patients were questioned for symptoms during the course of the examination. A 3D spoiled gradient-echo acquisition was acquired without contrast material to serve as a mask for subtraction from contrast-enhanced MR angiograms. A timing examination using a test dose of 1 ml of gadopentetate dimeglumine (Magnevist; Berlex, Wayne, NJ) was then performed according to the method described by Earls et al. [17]. Briefly, the test dose was used to measure the transit time (patient circulation time) from arm vein to abdominal aorta. This measurement was then used to time the arterial phase 3D acquisition to ensure maximum arterial enhancement after the injection of 19 ml of contrast material (average dose was 0.13 mmol/kg ± 0.03) at a rate of 2 ml/sec using the power injector.
Sequence parameters are detailed as follows. We performed cine phase-contrast acquisitions (TR/TE, 24/5; flip angle, 30°; signal averages, two; matrix, 128 x 256; slice thickness, 8 mm; field of view, 350 mm, using a rectangular field of view when possible; encoding velocity, 75 cm/sec) with a temporal resolution of 24 msec. The number of phases per cardiac cycle varied from 24 to 32 depending on each patient's heart rate. No temporal data interpolation or view-sharing was used. Acquisition times ranged from 2.5 to 4.5 min and were performed during free breathing. MR angiographic acquisitions were performed in the oblique coronal plane using a spoiled gradient echo acquisition (TR/TE, 3.8/1.3; flip angle, 25°; field of view, 350-400 mm, using a rectangular field of view when possible; matrix, 256 x 160; slab thickness, 70-120 mm). We achieved 1- to 2.5-mm resolution in the slice-select direction using sinc interpolation (zero-filling). For MR angiography, all acquisition times were less than 25 sec, and all studies were acquired during breath-hold at end-expiration.
Image Analysis
One radiologist analyzed MR angiograms at a Vision satellite console
workstation (Siemens) using standard multiplanar and
maximum-intensity-projection reconstructions. The radiologist was unaware of
clinical information and phase-contrast results. The degree of renal artery
stenosis was classified as equal to or greater than 60% or less than 60%
diameter reduction on the basis of contrast-enhanced 3D source images.
MR phase contrast data were analyzed on a Sparc 5 workstation (Sun Microsystems, Palo Alto, CA) using modified Interactive Data Language software (IDL; Research Systems, Boulder, CO) developed in house. Velocity images were produced using the standard algorithm and were analyzed by one investigator unaware of patient information and MR angiography findings. On phase images, renal artery mean and maximum velocity values were plotted against time. The following parameters, defined in the sonography literature [15], were derived from velocity waveforms: peak systolic velocity (m/sec), acceleration time (sec, time to early systolic peak), and acceleration index (m/sec2, slope of the early systolic rise). Additionally, we performed qualitative evaluation of the waveform shape. The presence or absence of an early systolic peak was recorded. The waveforms were classified according to a modification of the scheme described by Oliva et al. [16]. Waveform classification was performed both independently and by consensus by two investigators who were unaware of MR angiography results. Results were analyzed for classification into one of 10 patterns (Fig. 1) and then into one of two categories, normal (patterns I-VI) and abnormal (patterns VII-X), according to published criteria [16].
|
After data analysis, waveforms were retrospectively compared with angiography findings to evaluate potential sources of error.
Conventional Angiography
After MR imaging, eight patients (23%) were referred for conventional
angiography within 3 months. All studies were performed using a femoral artery
approach with a 5-French pigtail catheter. For most patients, selective
injections of both renal arteries were performed. Angiograms were interpreted
by a radiologist who was unaware of the results of our phase-contrast MR
study.
Statistical Analysis
We used the Student's t test to compare the characteristics of
patients with and without renal artery stenosis. The t test was also
used to compare quantitative parameters of phase-contrast waveforms before and
after the administration of an ACE inhibitor and to compare vessels with and
without significant (
60%) stenosis. We used the McNemar test to compare
the accuracy of waveform analysis in predicting renal artery stenosis before
and after the administration of an ACE inhibitor. A threshold of p
< 0.05 was used to define statistical significance.
We used the kappa coefficient to evaluate interobserver agreement; 
= 0 indicates agreement no greater than chance, and = 1 indicates
perfect agreement. Agreement was calculated for the classification of
waveforms into one of 10 waveform patterns
(Fig. 1) and then into one of
two categories, normal (patterns I-VI) and abnormal (patterns VII-X).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Overall, 27% of renal arteries (19/70) showed equal to or greater than 60% stenosis on MR angiography, corresponding to 40% of patients (14/35). Bilateral renal artery stenosis was found in 14% of patients (5/35). In 11% of patients (4/35), single accessory renal arteries were identified on MR angiography; however, none of them showed significant stenosis.
When compared with patients without renal artery stenosis, those with
significant disease (60% narrowing) were older (71 years versus 57 years;
p = 0.003), had higher systolic blood pressure (178 mm Hg versus 159
mm Hg; p = 0.04), and had longer circulation times (24 sec versus 20
sec; p = 0.02). There was no significant difference in diastolic
blood pressure or degree of blood pressure change after the administration of
an ACE inhibitor for the two groups (p > 0.2).
At the baseline, MR phase-contrast acquisitions were successfully obtained for all but one renal artery that appeared occluded on renal MR angiography. After the administration of ACE inhibitor, flow waveforms were obtained from all but two renal arteries; one corresponded to the same occluded vessel and a second was considered uninterpretable because of motion artifacts during scanning. The quantitative parameters for MR phase-contrast waveforms before and after the administration of ACE inhibitor are shown in Table 1 (Figs. 2A, 2B, 2C and 3A, 3B, 3C, 3D). The difference in acceleration time for renal arteries with and without significant stenosis was statistically significant at the baseline and after the administration of ACE inhibitor. Differences in acceleration index between the two groups approached statistical significance after the administration of ACE inhibitor; however, when each parameter was compared before and after the administration of ACE inhibitor, no significant difference was found (p = 0.14-0.96).
|
|
|
|
|
|
|
|
The results were further analyzed using the following previously published
Doppler sonographic parameters for hemodynamically significant stenosis
(60% narrowing): acceleration time equal to or greater than 0.07 sec,
acceleration index less than 3.0 m/sec2, and absence of early
systolic peak [12,
15]. Corresponding
sensitivities and specificities for the three criteria are shown in
Table 2 for the baseline and
ACE inhibitor-enhanced data. Again, there was no difference in results after
the administration of ACE inhibitor.
|
Qualitative assessments of phase-contrast waveforms obtained by consensus are shown in Table 3. According to the criteria established in the sonography literature [16], patterns I-VI (Fig. 1) were considered normal, whereas patterns VII-X were abnormal. Using these criteria, the sensitivity of waveform pattern analysis was 50% (9/18) before the administration of ACE inhibitor and 67% (12/18) after. The specificity improved slightly from 78% (40/51) to 84% (42/50); however, neither change was statistically significant (p > 0.2). Additionally, we examined the subset of 14 patients who were using ACE inhibitors as part of their routine medical regimen at the time of examination and found similar sensitivity (33%, 1/3) and specificity (82%, 9/11) that was unchanged after the administration of ACE inhibitor.
|
For classification to one of 10 waveform patterns, observer concordance was
34%, with a poor interobserver agreement value (=0.27). For
classification to normal (patterns I-VI) versus abnormal (VII-X) waveforms,
concordance was 83% with a good interobserver agreement (=0.60).
Of the eight patients who underwent subsequent contrast angiography, six
patients also underwent concurrent angioplasty, and three had concurrent renal
stent placement (one bilateral). When comparing MR angiography findings with
those of conventional angiography, we found a sensitivity of 100% (8/8), and a
specificity of 75% (6/8) for detecting significant (60%) stenosis. Of the
two patients with false-positive results on MR angiographic studies, one
patient was found at contrast angiography to have heavy calcifications at the
renal ostium (Fig. 4A,
4B,
4C), and in the second
patient, the vessel appeared to have a tortuous proximal course (Fig.
5A,
5B,
5C).
|
|
|
|
|
|
We retrospectively examined the MR phase-contrast waveforms corresponding to the two false-positive MR angiography studies. In the first false-positive study in which the heavy calcifications at the ostium presumably created susceptibility artifacts that mimicked stenosis on MR angiography, MR phase-contrast waveforms were interpreted as normal by both observers (Fig. 4A, 4B, 4C). In the patient with a tortuous renal artery and false-positive results on MR angiography, phase-contrast MR imaging measurements of velocity waveforms were interpreted as abnormal by both observers (Fig. 5A, 5B, 5C). Thus, in this patient, both MR phase-contrast and angiography data favored a false diagnosis of renal artery stenosis.
For one patient in whom normal-appearing phase-contrast waveforms were discrepant with the MR angiography findings suggesting right renal artery stenosis, we performed an additional phase-contrast acquisition at the level of the stenosis [18] (Fig. 6A, 6B, 6C, 6D). Systolic velocity greater than 2.5 m/sec was measured, consistent with significant stenosis based on sonographic criteria [19]. The stenosis was confirmed on conventional angiography, and the patient was successfully treated with angioplasty.
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Experimental models using flow phantoms may provide valuable insight into this controversy [23, 24]. Using the analogy of a resistor-capacitor electronic circuit, Bude et al. [23] hypothesized that the combination of elevated resistance at the stenosis or increased compliance of the downstream vessel serves effectively as a high-frequency filter, causing absence of the early systolic peak. On the basis of these models, patients with heavily calcified (noncompliant) vessels may not show the expected waveform dampening distal to significant renal artery stenosis. Conversely, abnormally dampened waveforms can be seen in the absence of renal artery stenosis in younger patients with highly compliant vessels [23]. These confounding effects may in part explain the poor sensitivity and specificity of waveform analysis observed by others using sonography and in this study using MR imaging.
Our overall baseline sensitivity (50%) and specificity (78%) for MR phase-contrast waveform pattern analysis in predicting at least 60% renal artery stenosis are in close agreement with Doppler sonographic findings reported by Kliewer et al. [20]. Of the quantitative measures analyzed, we found only acceleration time to be significantly different between normal and diseased vessels. We suspect that interobserver variability also contributes to the inaccuracy we found in waveform pattern analysis. The most common sources of error were related to questionable definition of the early systolic peak (e.g., differentiating patterns III and IV or II and V). These errors may be reduced by improving temporal resolution and decreasing artifacts arising from respiratory motion.
We found no significant change in qualitative or quantitative parameters of flow waveforms after the administration of ACE inhibitor. In initial studies one group [16] described an improvement in the accuracy of Doppler waveform analysis after the administration of ACE inhibitor; however, the underlying mechanism for this observation is unclear. Although Oliva et al. [16] argue that ACE inhibitors should exaggerate the pulsus parvus et tardus pattern, Bude et al. [25] state that an ACE inhibitor should increase waveform pulsatility because of decreased resistance. Clearly, further physiology and modeling experiments are needed.
One limitation of our study is our inclusion of 14 patients who were receiving ACE-inhibitor therapy as a part of their routine anti-hypertensive regimen. Renal scintigraphic studies have shown that captopril renography had a decreased sensitivity (75% [12/16] versus 98% [39/40]) for detecting renal artery stenosis in patients taking ACE inhibitors compared with those not taking these drugs [26]. However, we found no substantial difference in sensitivity and specificity of waveform pattern classification for predicted renal artery stenosis in the subset of 14 patients taking ACE inhibitors.
MR phase-contrast techniques have been validated in vitro and in vivo models as accurate noninvasive means of measuring blood flow [27, 28]. Using an implantable sonographic probe in animals and in humans intraoperatively, Schoenberg et al. [11] validated cine phase-contrast MR techniques for the measurement of flow velocity curves and mean renal artery flow. In a study of 23 patients with renal artery stenosis, they measured velocity waveforms 1-2 cm downstream from stenosis and observed that for stenoses greater than 60%, the early systolic peak decreased below the compliance peak (similar to patterns II and V) (Fig. 1). Using criteria of decreased early systolic peak and delayed peak velocity (corresponding to the compliance peak rather than early systolic peak), these researchers reported a sensitivity of 100% (19/19) and a specificity of 97% (28/29) for predicting renal artery stenosis greater than 50%. Given that the location of MR measurements in their study was 1-2 cm downstream from the stenoses—that is, in the midportion of the main renal artery—a direct comparison of these results with those in the sonography literature (measured either at the stenosis or in hilar or intrarenal vessels) or with our results (measured in the renal hila) is difficult.
MR phase-contrast estimates of flow provide several advantages over sonographic approaches. The method is not limited by patient body habitus or anatomic factors such as bowel gas. The angle of insonation in Doppler measurements is vital to the accuracy of measurements of flow, and likely contributes to the high degree of variability in waveform patterns observed in healthy subjects [29]. MR phasecontrast techniques require slice positioning orthogonal to flow that can be easily achieved regardless of patient anatomy. Disadvantages of the MR approach include the lower temporal resolution compared with Doppler sonography, reliance on consistent cardiac rhythms for ECG triggering, and, until recently, relatively long examinations that prohibit real-time sampling at multiple locations.
Although we found relatively low sensitivity and specificity for MR phase-contrast waveform analysis for predicting significant renal artery stenosis, limitations in our protocol and study design should be considered. Phase-contrast images that have at least 30 msec temporal resolution (24 msec in our study) typically require lengthy imaging times (1-3 min) with commercially available sequences, precluding breath-hold acquisitions. Breath-hold segmented K-space phase-contrast methods with view-sharing are available on selected commercial systems and can be modified using higher receiver bandwidths to obtain temporal resolution of less than 30 msec, depending on the patient's ability to breath-hold [30, 31]. This high temporal resolution is necessary to ensure accurate definition of the period of early systolic acceleration [24]. The breath-hold approach may help to eliminate potential artifacts arising from motion [32] and allow repeated samplings along the length of the renal artery. Early efforts in real-time MR fluoroscopic measures of flow are also promising (Heid O, presented at the International Workshop on Magnetic Resonance Angiography, September 1998; Nayak KS et al., presented at the Society for Cardiovascular Magnetic Resonance, January 1999).
Present spatial resolution limitations preclude the evaluation of intrarenal vessels with MR phase-contrast techniques. Conclusions based on measurements of hilar segmental branches may therefore not be comparable with sonographic results for intrarenal segmental branch velocities. With Doppler sonography, the accuracy of measurements in the main renal artery exceeds that of intrarenal artery measurements [22]; however, the approach has been significantly limited by the inability to see the main renal arteries in as many as 45% of patients [19, 32]. MR imaging may provide an impetus and opportunity for reevaluating velocity measurements at the level of stenosis. With faster acquisition times, phase-contrast velocity measurements can be obtained in the same setting as routine MR angiography, with minor additional costs of time and ECG lead pads.
In our study, most patients did not undergo conventional angiography after MR angiography, and thus one further limitation is the lack of conventional angiographic validation of MR results. However, reliance on MR angiography results is an emerging trend based on consistently excellent correlation with conventional angiography [33, 34, 35].
Tortuous vessels can cause nonlaminar flow and lead to the dissipation of phase coherence. This dissipation manifests as signal loss on all MR techniques [36] and necessitates alternative approaches. This limitation may explain the false-positive waveforms observed in one of our patients (Fig. 5A, 5B, 5C). MR perfusion and renographic methods [37, 38] may be able to assess for significant renovascular disease in this setting.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Miyazaki and V. S. Lee Nonenhanced MR Angiography Radiology, July 1, 2008; 248(1): 20 - 43. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
