| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Departments of Cardiovascular Radiology (S.W., M.F., C.L., J.B., M.F., V.G., J.P.B.), Nephrology (O.M.), and Cardiology (C.V.), CHRU de Lille, 59037 Lille Cedex, France. Received January 29, 2005; revision requested March 31; revision received September 28; accepted October 17; final version accepted March 1, 2006. Supported by the EA2693. Address correspondence to S.W. (e-mail: s-willoteauxs{at}chru-lille.fr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Materials and Methods: Institutional review board approval was obtained, with waiver of informed consent. The results of renal contrast-enhanced MR angiography were retrospectively analyzed in 25 patients with hypertension (24 women, one man; mean age, 48 years ± 19 [standard deviation]; age range, 18–72 years) who had FMD diagnosed on the basis of clinical and angiographic features. All examinations were performed at 1.5 T. Results were analyzed by two readers, and a third reader established a consensus in case of discrepancy. Sensitivity, specificity, and 95% confidence intervals (CIs) were calculated for FMD and for each possible type of FMD lesion ("string of pearls" appearance, stenosis, and aneurysm). A linear-weighted 
statistic was calculated to determine agreement between digital subtraction angiography and contrast-enhanced MR angiography for the diagnosis of FMD and to determine inter- and intraobserver agreement regarding FMD diagnosis.
Results: Fifty main renal arteries were analyzed, 35 of which demonstrated abnormal arteriographic features of FMD (stenosis, 22 arteries; string of pearls, 21 arteries; and aneurysm, four arteries). The sensitivity and specificity of contrast-enhanced MR angiography for the diagnosis of FMD were 97% (95% CI: 83%, 100%) and 93% (95% CI: 66%, 100%), respectively. Sensitivity was 68% (95% CI: 83%, 100%), 95% (95% CI: 74%, 100%), and 100% (95% CI: 40%, 100%) for the diagnosis of stenosis, string of pearls, and aneurysm, respectively. Linear-weighted statistics for inter- and intraobserver agreement regarding FMD diagnosis were 0.63 and 0.92, respectively.
Conclusion: In patients with renal FMD, contrast-enhanced MR angiography can reliably facilitate diagnosis by demonstrating characteristic lesions.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In a previous study (2), renal artery FMD was described as a cause of hypertension in a young predominantly female population. Conventional arteriography is currently used to determine the location and the extent of renal artery involvement. Focal stenoses, a "string-of-pearls" appearance (indicating multiple stenoses), and aneurysm are classic arteriographic signs of FMD (3). In elderly patients, these classic signs may coexist with atherosclerotic disease (4).
Renal FMD represents a group of disorders characterized by fibrous or muscular hypersplasia in one or more layers (intima, media, and adventitia) of the renal artery wall (5). The most frequent form of renal FMD is medial fibromuscular disease, which occurs in approximately 65%–70% of cases and has a classic string-of-pearls appearance at conventional angiography, with or without aneurysm formation. Less common is the perimedial (subadventitial) form, which occurs in approximately 15%–20% of cases and is characterized angiographically by aneurysm formation and focal or long stenoses. Medial hyperplasia, which occurs in 8%–10% of cases, has no pathognomonic radiologic appearances. Isolated intimal and adventitial involvement is extremely rare (1%–2% of cases) (4).
Several different imaging modalities are currently available for the detection of renal artery stenosis (6): isotope scanning before and after the administration of captopril, duplex ultrasonography (US) (7), computed tomography (CT) (8–10), and contrast material–enhanced magnetic resonance (MR) angiography (11). The reported sensitivities and specificities of these modalities range from 75% to 100%. In most published studies, however, researchers have included patients with atherosclerotic renal artery stenosis. The reference standard for the diagnosis of renal artery FMD is the depiction of classic renal artery abnormalities on digital subtraction angiograms or in pathologic specimens.
Thus, the purpose of our study was to retrospectively evaluate the sensitivity and specificity of contrast-enhanced MR angiography by using digital subtraction angiography as the reference standard in patients with hypertension and renal artery FMD.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Inclusion criteria were a clinical diagnosis of hypertension, contrast-enhanced MR angiography performed as one of the noninvasive diagnostic tests for the detection of potential renovascular hypertension, and conventional digital subtraction angiography performed after contrast-enhanced MR angiography. A total of 25 patients (24 women, one man; mean age, 48 years ± 19 [standard deviation]; age range, 18–72 years) were included. All patients had hypertension that was resistant to multiple medications or had one or more criteria for secondary hypertension. In each case, contrast-enhanced MR angiography was performed as one of the noninvasive diagnostic tests for the detection of potential renovascular hypertension.
In all patients, conventional angiography was performed 1–60 days after contrast-enhanced MR angiography (mean, 30 days). Contrast-enhanced MR angiograms, which were obtained by using different methods of reformation, were reviewed and compared with subsequent arteriograms. The institutional review board at our institution approved the study and waived the requirement for individual informed consent.
Arteriography
Digital subtraction angiography of the renal arteries was performed by five different angiographers (among them J.P.B. and V.G.) who had 2–9 years of experience in catheterization. These examinations were performed with a 5-F pigtail catheter by using a transfemoral or transbrachial approach. After the pigtail catheter was positioned in the abdominal aorta, contrast material (sodium meglumine ioxaglate, Hexabrix; Guerbet, Aulnay-Sous-Bois, France) that contained 320 mg of iodine per milliliter was injected at a rate of 18–20 mL/sec by using a power injector. One arteriographic examination was performed by using an Integris V5000 system (Philips, Eindohoven, the Netherlands) with a matrix of 1024 x 1024. Seven arteriographic examinations were performed by using an Integris V3000 system (Philips) with a matrix of 512 x 512. Seventeen arteriography examinations were performed by using an Infinix system (Toshiba Medical Systems, Zoetermeer, the Netherlands) with a matrix of 1024 x 1024. A 38-cm field of view was used for global acquisitions in a 5° left anterior oblique projection.
When visualization of the renal vasculature was considered inadequate (either because of superimposition of the superior mesenteric artery or because of inadequate visualization of the ostia), we obtained right anterior oblique or frontal views. Selective renal arteriography with manual contrast material injection (10 mL) was systematically performed by using a preshaped visceral catheter and a 25-cm field of view. When indicated, arterial blood pressures were also measured in both the renal artery and the aorta to assess the significance of stenosis, particularly when a string-of-pearls appearance was observed. The total volume of contrast material that was injected varied between 100 and 150 mL. All examinations were performed by using a computer-assisted digital subtraction technique.
Contrast-enhanced MR Angiography
Three-dimensional (3D) contrast-enhanced MR angiography was performed with four different imagers, all of which were 1.5-T systems equipped with high performance gradients. Seven examinations were performed by using a Magneton Vision system (Siemens Medical Systems, Erlangen, Germany), with imaging parameters of 3.2/1.2 [repetition time msec/echo time msec], flip angle of 40°, and field of view of 450 x 450. Three examinations with different sequences were performed by using a Magneton Symphony system (Siemens Medical Systems), with imaging parameters of 3.4–4.6/1.2–1.8, flip angle of 30°–40°, and field of view of 350–400 x 263–325. One examination was performed by using a Genesis Signa system (GE Medical Systems, Milwaukee, Wis), with imaging parameters of 4/1, flip angle of 30°, and field of view of 340 x 340. Fourteen examinations were performed by using a Gyroscan Intera system (Philips), with imaging parameters of 5.2/1.5, flip angle of 40°, and field of view of 400 x 400.
All examinations were performed by using a phased-array surface coil. Parallel imaging and elliptical centric phase encoding techniques were used for 15 examinations (14 with Gyroscan Intera and one with Genesis Signa). For the other examinations, phase encoding was performed sequentially. The 3D slab was oriented in the coronal plane. Prior to contrast-enhanced MR angiography, a 20–22-gauge intravenous catheter was placed in an antecubital vein and was connected to a power injector that was loaded with gadolinium (gadoterate meglumine, Dotarem; Guerbet); injection rates ranged from 1.2 to 2.0 mL/sec. Each injection of contrast material was followed by a 20-mL saline flush. The volume of contrast material varied between 0.1 and 0.2 mmol per kilogram body weight. In each case, contrast-enhanced MR angiography comprised a single acquisition at the arterial phase of intravenous injection of contrast material.
A total of eight examinations were performed with a bolus timing technique to determine the bolus transit time; bolus timing was performed with a single-section gradient-recalled-echo sequence, which was used to collect images of the abdominal aorta after injection of a 2-mL test bolus of contrast material every 1 second for 30 seconds. Real-time monitoring of the arrival of the contrast material bolus (fluoroscopic MR angiography) was used for 17 examinations. When enhancement of the abdominal aorta was achieved, contrast-enhanced MR angiography was initiated by the radiographer under the supervision of the radiologist.
The following rendering algorithms were applied to the coronal source images: transverse multiplanar reformations, maximum intensity projections, and volume-rendered reconstructions. All of these algorithms were performed at the time of examination in 16 cases and were retrospectively performed during the study period in three cases for which the digital data were still available.
Image Analysis
All renal arteriograms were reviewed by one radiologist (J.P.B., with 12 years of experience in the interpretation of digital subtraction angiograms) by using film images, and the number of renal arteries was determined. The proximal, middle, and intrarenal parts of these arteries were analyzed, and the presence or absence of the characteristic signs of FMD (string-of-pearls appearance, focal stenoses, or aneurysms) was noted. A stenosis was defined as a narrowing of 50% or more in diameter of the renal artery; an aneurysm was defined as an abnormal widening of the artery, with loss of parallelism of the vascular wall of the artery; and a string-of-pearls appearance was defined as a sequence of outpouchings along the course of the renal artery, simulating beads threaded onto a string (3). The evaluation was visual, and digital subtraction angiograms of the renal arteries were used as the reference standard.
Two other radiologists (readers 1 and 2), with 4 and 6 years of experience in contrast-enhanced MR angiography and vascular imaging, reviewed each series of image reformations without knowledge of the corresponding arteriographic results. These two readers knew that all patients had FMD. Eleven readings were performed by using film images, and 14 were performed by using a workstation. The two readers evaluated the same images and did not perform the image reformations. Coronal source images were available in 23 patients (46 main renal arteries), transverse multiplanar reformations were available in 22 patients (44 main renal arteries), maximum intensity projections were available in all 25 patients (50 main renal arteries), and volume-rendered images were available in 19 patients (38 main renal arteries). Thus, each reader had to evaluate 89 reformatted images for the right and left renal arteries (178 readings).
Reader 1 analyzed each set of images once, and reader 2 analyzed each set of images twice in a different order, with at least 1 week between analyses in order to limit learning bias. Analyses included an assessment of the main renal arteries and the accessory arteries. The proximal and middle parts of the renal arteries were studied, as well as their hilar and intrarenal parts (when possible). The number of arteries with evidence of FMD was recorded, and each specific sign (stenosis, string-of-pearls appearance, or aneurysm) was sought on each type of reformation.
For each contrast-enhanced MR angiographic examination and during each interpretation, all of the different reconstruction modalities were analyzed together during the same session. For each artery, each of the different reformatted images was assessed for the presence of one or more signs of FMD. In addition, the number of stenoses and/or aneurysms was noted.
At the end of each interpretation session, a diagnosis of FMD in an artery was confirmed if at least one of the reconstruction modalities depicted evidence of at least one lesion that demonstrated characteristics of FMD. The results of the analyses by reader 1 and those of the first analysis by reader 2 were subsequently compared. If a difference in interpretation was found between the two readers for the presence or absence of a lesion, a third reader (S.W., with 7 years of experience in reading contrast-enhanced MR angiograms) reviewed the image reformation in question and established a final diagnosis.
Statistical Analysis
With conventional angiography as the reference standard, sensitivities and specificities of contrast-enhanced MR angiography were calculated with 95% confidence intervals (CIs) for the final diagnosis. These parameters were calculated for (a) the diagnosis of FMD, stenosis, string-of-pearls appearance, and aneurysm in the 50 main renal arteries and (b) the diagnosis of each specific sign by using each reconstruction modality. In the particular case of stenosis, calculations were made according to the number of stenoses and not according to the number of stenosed arteries because an artery could have more than one stenosis.
A linear-weighted statistic was calculated to determine the level of agreement between digital subtraction angiography and contrast-enhanced MR angiography in the diagnosis of FMD and each lesion type. A linear-weighted statistic was also calculated to determine inter- and intraobserver agreement with respect to FMD diagnosis (12). These calculations were performed by using spreadsheet software (Excel, 2004; Microsoft, Redmond, Wash) and StatView for Windows (release 5; SAS Institute, Cary, NC). A value of 0.20 or less was considered slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
= 0.91). One main renal artery was classified as having a string-of-pearls appearance at contrast-enhanced MR angiography but was normal at digital subtraction angiography. In one patient, one of the main renal arteries was seen as normal at contrast-enhanced MR angiography but was found to have a stenosis at digital subtraction angiography; in this patient, the contralateral artery was identified as having a string-of-pearls appearance at both contrast-enhanced MR angiography and digital subtraction angiography. In two of the eight renal arteries with two lesions, only one of the lesions was identified. In the renal artery with both a string-of-pearls appearance and a stenosis, the stenosis was missed at MR angiography; in the renal artery with both a string-of-pearls appearance and an aneurysm, the string-of-pearls appearance was missed.
A definite diagnosis of FMD was established in all patients at 3D contrast-enhanced MR angiography because at least one diagnostic feature was identified at 3D contrast-enhanced MR angiography in every patient (Figs 1–3). Sensitivities and specificities for the diagnosis of FMD are reported in Table 2.
|
|
|
|
|
|
|
|
|
|
= 0.65), 20 (95%) of 21 lesions with a string-of-pearls appearance ( = 0.92), and four (100%) of four aneurysms. The sensitivities and specificities of contrast-enhanced MR angiography for each type of lesion are reported in Table 2. Sensitivity and specificity values for source coronal images and for each reconstruction modality are reported in Table 3 for stenoses, string-of-pearls appearance, and aneurysm.
|
= 0.63). A third reader was necessary to establish a final diagnosis for 19 of 178 analyses. Agreement between the first and second readings by reader 2 was almost perfect ( = 0.92) (Table 4).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
To our knowledge, the first study to evaluate the abdominal aorta and its branches by using breath-hold 3D contrast-enhanced MR angiography was published in 1995 by Prince et al (13). Many investigators have reported sensitivities of 96%–100% and specificities of 71%–95% for the detection of renal artery stenosis (14,15). However, most studies have focused on only atherosclerotic stenosis at the renal artery ostia. FMD is far less frequent than stenosis of atheromatous origin. However, the diagnosis of FMD is still important because angioplasty has shown good results in the treatment of hypertension secondary to FMD (16). Renal artery FMD depicted at contrast-enhanced MR angiography has already been described, but the sensitivity and specificity of this technique have not been previously established (11,17).
The main limitation of contrast-enhanced MR angiography for the diagnosis of renal FMD has been its lack of spatial resolution and its inability to demonstrate the distal portion and small intrarenal branches of the main renal artery (18). For the study of the renal artery, particularly for the diagnosis of FMD, the challenge with contrast-enhanced MR angiography has been to maintain a high spatial resolution and signal-to-noise ratio while minimizing the imaging time such that artifacts due to inherent patient motion are minimized.
In the present study, which was conducted between 1998 and 2004, several different sequences were used with different MR imagers. Important technical improvements occurred during the 6 years of the study period. The first major improvement was the use of the bolus tracking technique (used in 16 examinations in our study) instead of the test bolus technique (used in nine examinations in our study) to determine the synchronization of injection and acquisition. The disadvantages of the test bolus technique were the persistence of enhancement in the renal veins and the enhancement of the urinary tract, which occurred during main acquisition and interfered with assessment of arterial renal structures.
The use of real-time monitoring to determine the arrival time of the contrast material bolus and to trigger contrast-enhanced MR angiography (19) allowed for the reliably precise timing of the beginning of contrast-enhanced MR angiography to coincide with the arrival of the contrast material bolus (20). However, the present study was not designed to compare the value of the test bolus technique with that of the bolus tracking technique.
Another technical improvement was the acquisition of images with elliptical centric view ordering, which minimized interference from venous signal intensity and respiratory motion artifacts and permitted longer acquisition times. Combining the bolus tracking technique with elliptical centric view ordering has been shown to provide reliably high-quality MR angiograms of the renal arteries, with high arterial contrast and excellent suppression of venous signal intensity (19,20).
Another major and recent improvement in MR imaging is the use of parallel imaging, such as sensitivity encoding, that increases the spatial and/or temporal resolution. Sensitivity encoding has enabled a substantial decrease in the overall time constraints for imaging. Preliminary results have compared favorably with digital subtraction angiography, with diagnostic accuracy exceeding 90% for the diagnosis of renal artery stenosis by using contrast-enhanced MR angiography (21,22).
Seven stenoses were missed during contrast-enhanced MR angiography in our study. Consequently, the level of agreement between contrast-enhanced MR angiography and digital subtraction angiography for the diagnosis of stenosis was only substantial ( = 0.65). Two arteries had more than one stenosis at digital subtraction angiography (one artery with two stenoses and one artery with three stenoses), but only the more proximal stenosis was seen at contrast-enhanced MR angiography in each artery. Although the number of stenoses was underestimated in these arteries, the presence of stenosis was correctly identified. Correlation with digital subtraction angiography for the diagnosis of a string-of-pearls appearance was almost perfect ( = 0.92).
The identification of 10 (77%) of the 13 accessory renal arteries at contrast-enhanced MR angiography, with one false-positive finding, is concordant with published results (14). We separated main and accessory renal arteries because it has been shown that the ability of a noninvasive modality to demonstrate hemodynamically significant stenoses in accessory arteries provides additional information but should not substantially improve the detection rate of renovascular hypertension with that modality (23). Despite the fact that the detection of FMD in accessory renal arteries was not the main objective of the study, two of the three pathologic accessory renal arteries were correctly diagnosed at contrast-enhanced MR angiography.
The level of agreement between the two readers for the diagnosis of FMD on contrast-enhanced MR angiograms was substantial ( = 0.63). This could be interpreted in one of two ways: Either reader 1 relied on diagnostic criteria favoring a high test sensitivity (97%; 95% CI: 83%, 100%) at the expense of specificity (80%; 95% CI: 51%, 95%), or reader 2 employed criteria favoring a high test specificity (100%; 95% CI: 75%, 100%) at the expense of sensitivity (86%; 95% CI: 69%, 95%). Intraobserver agreement was almost perfect ( = 0.92).
Algorithms for the screening of renal artery stenosis have been proposed and encompass a wide range of available noninvasive techniques, including duplex US, multisection CT, and contrast-enhanced MR angiography (1,24). It has been shown that CT angiography gives good results in the depiction of FMD (4) but exposes the patient to ionizing radiation and to potentially nephrotoxic iodinated contrast material. Contrast-enhanced MR angiography, with injection of gadolinium-based contrast material, avoids these two hazards (25,26).
Our study has limitations. The radiologists reviewing the contrast-enhanced MR angiograms knew that each patient had FMD and that there was no control group. Because our results were based on findings in a selected group of patients with confirmed FMD, the reported diagnostic performance of contrast-enhanced MR angiography cannot be extrapolated to a more general population with hypertension of mixed causes.
Another limitation could be the use of renal artery stenosis as one of the three signs of FMD. The discovery of a renal artery stenosis in a patient with hypertension does not first evoke the diagnosis of FMD, but our series included patients with FMD that was diagnosed on the basis of angiographic observation, as well as on clinical arguments that ruled out other causes. The study included 25 examinations performed by using four different imagers over a 6-year period; this inhomogeneity could constitute a limitation, but our study examined contrast-enhanced MR angiography rather than a specific technique of acquisition. The retrospective nature of the study constituted another limitation but permitted us to gather data on a large population for a disease with a low prevalence.
In conclusion, for patients in whom renal artery FMD was diagnosed on the basis of clinical and angiographic features, contrast-enhanced MR angiography can facilitate diagnosis by demonstrating characteristic lesions. Recent improvements in MR imaging and contrast-enhanced MR angiography increase spatial and temporal resolution without diminishing the signal-to-noise ratio. The exact role of contrast-enhanced MR angiography should be defined in further prospective studies by using the most up-to-date developments for the diagnosis of FMD in a more general population that is suspected of having renovascular hypertension and by employing the most up-to-date developments in MR imaging.
|
FOOTNOTES |
|---|
Abbreviations: CI = confidence interval • FMD = fibromuscular dysplasia • 3D = three-dimensional
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, S.W., J.P.B.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, S.W., M.F., C.L., V.G.; clinical studies, S.W., M.F., O.M., C.L., V.G., J.P.B.; statistical analysis, S.W., O.M.; and manuscript editing, S.W., J.B.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  U. Kramer, J. Wiskirchen, M. C. Fenchel, A. Seeger, G. Laub, G. Tepe, J. P. Finn, C. D. Claussen, and S. Miller Isotropic High-Spatial-Resolution Contrast-enhanced 3.0-T MR Angiography in Patients Suspected of Having Renal Artery Stenosis Radiology, April 1, 2008; 247(1): 228 - 240. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
