Jeffrey H. Maki

Accuracy of Computed Tomographic Angiography and Magnetic Resonance Angiography for Diagnosing Renal Artery Stenosis

Context Physicians sometimes use computed tomographic angiography (CTA) or magnetic resonance angiography (MRA) to diagnose renal artery stenosis. Contribution This prospective multicenter study compared CTA and MRA with digital subtraction angiography (the reference standard) in 402 hypertensive patients with suspected renal artery stenosis. Multiple experienced physicians sometimes disagreed about whether the CTA and MRA tests showed renal artery stenosis. The sensitivity estimates of CTA and MRA for detecting renal artery stenosis were 64% and 62%. Implications In this study, even trained physicians had difficulty interpreting some CTA and MRA tests, and neither test was sensitive enough to rule out renal artery stenosis. Renal artery stenosis may cause renovascular hypertension and renal impairment. Accurate detection and treatment of clinically relevant stenoses may cure or improve hypertension and preserve renal function. Current treatment options include surgery, percutaneous transluminal renal angioplasty with or without stent placement, and medical therapy. Despite the availability of several other diagnostic tests, intra-arterial digital subtraction angiography (DSA) remains the reference standard for anatomic diagnosis of renal artery stenosis. This test, however, is an invasive procedure that carries a risk for serious complications and is burdensome for patients (1, 2). For this reason, less invasive diagnostic alternatives, such as computed tomographic angiography (CTA) and 3-dimensional contrast-enhanced magnetic resonance angiography (MRA), are widely used for diagnostic work-up in patients with suspected renal artery stenosis. A recent meta-analysis (3) found that CTA and MRA were significantly better than non-contrast-enhanced magnetic resonance angiographic techniques, ultrasonography, captopril renal scintigraphy, and the captopril test at identifying renal artery stenosis when DSA was used as the reference standard. To date, however, only a limited number of small, well-designed studies have been published on the diagnostic accuracy of either CTA or MRA for detection of renal artery stenosis in patients with suspected renovascular hypertension (4-14). Because CTA and MRA seemed to be promising techniques with the potential to reduce the number of patients requiring conventional angiography, we set up a large-scale multicenter study to investigate the diagnostic performance of these tests, using DSA as reference standard, in hypertensive patients clinically deemed at risk for renal artery stenosis. The purpose of our study was to determine the interobserver agreement and diagnostic accuracy of CTA and MRA in comparison with DSA and to examine whether CTA or MRA can be used as an initial test for detection of renal artery stenosis. Methods We performed a prospective comparative study among CTA, MRA, and the reference standard, DSA, for the detection of renal artery stenosis. Each included patient underwent all 3 diagnostic tests. Participants Over a 3-year period, patients were prospectively recruited from the internal medicine outpatient clinics of 3 large teaching hospitals and 3 university hospitals in the Netherlands. The ethical review board of each hospital approved the study, and written informed consent was obtained from all participants. At the 2 hospitals that recruited most of the participating patients, enrollment was consecutive; the other participating hospitals included patients by using nonsystematic convenience samples. At the 6 participating centers, all hypertensive patients between 18 and 75 years of age with a diastolic blood pressure greater than 95 mm Hg were routinely screened for predefined clinical clues indicating renal artery stenosis, as described by the Working Group on Renovascular Hypertension (15) and others (16, 17). Patients were eligible for participation in the study if they exhibited at least 1 clinical clue. Exclusion criteria were known allergy to iodinated contrast agents; pregnancy; contraindications to MRA, CTA, or DSA (18, 19); contraindications to intervention; or previous participation in the study. All included patients were scheduled to have CTA, MRA, and DSA within a 3-month window. At the coordinating center (Maastricht University Hospital), included patients were scheduled to undergo CTA and MRA on the same day, followed by DSA the next day. At the other centers, the tests were performed on the basis of availability. No treatments that could affect the test results were allowed before all tests were completed. The case record forms for all patients were collected at the coordinating center, and the information was entered into a database. Imaging Techniques Each participating hospital was equipped with state-of-the art magnetic resonance scanners (1.0 or 1.5 Tesla), helical computed tomography scanners (single- or multi-detector row systems), and DSA equipment. In addition, hospitals were allowed to optimize scan protocols during the study when new insights emerged or when equipment was upgraded, an approach that conforms to usual clinical practice. Changes in scan protocols occurred twice (Appendix Table 1). To ensure state-of-the art magnetic resonance imaging, all scan protocols had to meet minimal quality standards in terms of spatial resolution and scan duration. The quality standards were defined by the coordinating center and were based on the protocols that were published at the start of the study. During the entire study, the coordinating center continuously monitored the quality of all images. Information about manufacturers, scan protocols, and contrast agents is shown in Appendix Table 1. All imaging was performed or supervised by experienced radiologists and radiologic technologists. Renal CTA, MRA, and DSA had already been part of clinical routine before the start of the study. Image Evaluation At the conclusion of study enrollment, 2 panels of 3 observers evaluated the CTA and MRA image data at the coordinating center. All observers had more than 3 years of experience evaluating such data on a regular basis, and for each method 1 observer had more than 6 years of experience. Each observer independently performed the evaluations and was blinded to all other results, including clinical information and DSA results. Digital image data for all CTA and MRA examinations were evaluated by using a work station equipped with all commonly used image-processing tools (EasyVision, release 4.2.1, Philips Medical Systems, Best, the Netherlands). Source images had to be examined in all cases before a final diagnosis could be made. The DSA images were evaluated by 4 vascular radiologists, all with more than 10 years of experience in this particular field. The first observer was the radiologist who actually performed the test; the evaluation took place during the DSA procedure. The second and third observers who judged each DSA examination knew the first observers judgment. If discrepancies existed among the first 3 observers with respect to the number of renal arteries involved or the nature, location, or severity of disease (differences of >10% in the degree of stenosis), a fourth radiologist, who had access to the diagnoses of the other observers, made the final diagnosis. This consensus approach has been used in several other CTA and MRA studies (6, 7, 12-14). All DSA observers were blinded to the results of CTA and MRA. To determine the degree of stenosis, the diameter of the most severely affected part of a renal artery was measured and related to the reference diameter, which was defined as the diameter of a representative nonaffected portion of the artery, preferably immediately distal to the stenosis (that is, beyond the site of poststenotic dilatation, if present). Fibromuscular dysplasia was diagnosed when multiple aneurysms separated by focal narrowing (string-of-beads sign) were observed. For CTA, MRA, and DSA, luminal narrowing of at least 50%, as well as all cases of fibromuscular dysplasia, was defined as clinically relevant renal artery stenosis (3). For each patient, the observers first recorded the number of renal arteries. Subsequently, these arteries were judged with respect to the presence or absence of stenosis (expressed as percentage of luminal narrowing), the nature of the stenosis (atherosclerotic or fibromuscular dysplasia), the location of the stenosis (ostial or truncal), and the level of confidence in the diagnosis (high, moderate, or poor) (6). Inconclusive examination results were noted on the standardized form used to collect all relevant data. Statistical Analysis The severity of the stenoses as seen on CTA and MRA was categorized on a 5-point scale (grade 1, 0% to 19%; grade 2, 20% to 49%; grade 3, 50% to 74% or fibromuscular dysplasia; grade 4, 75% to 99%; and grade 5, total occlusion [100% stenosis]). The Cohen weighted analysis was used to test for agreement beyond that of chance among the 3 observers of MRA and among the 3 observers of CTA (20). Unless stated otherwise, all analyses on the diagnostic accuracy of CTA and MRA (sensitivity, specificity, and receiver-operating curve [ROC] analysis) compared with DSA are based on patients as the unit of analysis. In the by-patient analysis, a patient was classified as having positive results if 1 or more renal arteries were found to be stenotic ( 50%) on DSA. The most severe stenosis per patient was used for analysis. Inconclusive CTA and MRA results were considered as positive test results because further diagnostic work-up would be required in clinical practice and these patients would be referred for DSA. Exact 2-sided 95% CIs for proportions were calculated by using a binomial distribution. Overall estimates of sensitivity, specificity, positive predictive value, and negative predictive value for all observers per method, including 95% CIs, were calculated by using the cluster option of Stata, version 8.2 (Stata Corp., College Station, Texas) (21). This

Motion of the distal renal artery during three-dimensional contrast-enhanced breath-hold MRA.

To study the potential detrimental effects of renal motion on breath‐hold three‐dimensional contrast‐enhanced (CE) magnetic resonance angiography (MRA).

Parallel imaging in MR angiography.

The recently developed techniques of parallel imaging with phased array coils are rapidly becoming accepted for magnetic resonance angiography (MRA) applications. This article reviews the various current parallel imaging techniques and their application to MRA. The increased scan efficiency provided by parallel imaging allows increased temporal or spatial resolution, and reduction of artifacts in contrast-enhanced MRA (CE-MRA). Increased temporal resolution in CE-MRA can be used to reduce the need for bolus timing and to provide hemodynamic information helpful for diagnosis. In addition, increased spatial resolution (or volume coverage) can be acquired in a breathhold (eg, in renal CE-MRA), or in otherwise limited clinically acceptable scan durations. The increased scan efficiency provided by parallel imaging has been successfully applied to CE-MRA as well as other MRA techniques such as inflow and phase contrast imaging. The large signal-to-noise ratio available in many MRA techniques lends these acquisitions to increased scan efficiency through parallel imaging.

Use of a three‐station phased array coil to improve peripheral contrast‐enhanced magnetic resonance angiography

To explore the imaging capabilities of a new commercially available, three‐station, 129‐cm long, 12‐element phased array coil for contrast‐enhanced magnetic resonance angiography (CE‐MRA) in patients with symptomatic peripheral arterial occlusive disease.

SNR improvement in NMR microscopy using DEFT

Abstract This paper examines the use of a driven equilibrium Fourier transform (DEFT) pulse sequence for improving the signal per unit time and hence image resolution in NMR microscopy. DEFT vs partial saturation (PS) is modeled and it is shown that DEFT is most useful in physiologic materials provided short TE values (TE ⪡ T 2 ) and short TR values (TR T 1 ) are used. Under these conditions, DEFT can yield up to a fourfold signal increase compared to PS. It is shown that DEFT can provide spin density and T 1 / T 2 -ratio-weighted images. DEFT is also shown to have SNR advantages as T 1 increases—an important consideration at higher magnetic fields. Experimental data that verify the theoretical predictions and the functioning of a DEFT pulse sequence to produce high-quality 2D spin-warp images of a phantom are presented. Studies performed on small animals demonstrate the utility of the DEFT sequence in MR microscopy by providing increased SNR and new contrast mechanisms over limited fields of view.

Initial Clinical Experience with Dual-Agent Relaxation Contrast for Isolated Lymphatic Channel Mapping

Purpose To evaluate the clinical performance of dual-agent relaxation contrast (DARC) magnetic resonance (MR) lymphangiography compared with that of conventional MR lymphangiography in the creation of isolated lymphatic maps in patients with secondary lymphedema. Materials and Methods This retrospective study was approved by the institutional review board. The diagnostic quality of 42 DARC MR lymphangiographic studies was compared with that of 42 conventional MR lymphangiographic studies. Two independent readers rated venous contamination as absent, mild, or moderate to severe. Interreader agreement on venous contamination grades was assessed by using the linearly weighted Cohen κ statistic. The Mann-Whitney U test was used to compare the distribution of grades at each station between conventional MR lymphangiography and DARC MR lymphangiography for each reader separately. Results DARC MR lymphangiography had significantly less venous contamination than did conventional MR lymphangiography (P < .001). The two radiologists rated venous contamination as moderate to severe in 64% (27 of 42) and 69% (29 of 42) of distal limbs, 23% (10 of 42) of midlimbs, and 2% (one of 42) and 9% (four of 42) of proximal limbs at conventional MR lymphangiography compared with 0% (0 of 42) of distal limbs, 2% (one of 42) of midlimbs, and 0% (0 of 42) of proximal limbs at DARC MR lymphangiography. Lymphatic signal was partially attenuated (median 45% decrease) when longer echo times were used for venous suppression, but it did not subjectively degrade diagnostic quality. Conclusion DARC MR lymphangiography yields isolated lymphatic maps through nulling of venous contamination, thereby simplifying diagnostic interpretation and communication with surgical colleagues.

Effect of injection rate on contrast-enhanced MR angiography image quality

Contrast‐enhanced (CE)‐MRA optimization involves interactions of sequence duration, bolus timing, contrast recirculation, and both R1 relaxivity and R2* ‐related reduction of signal. Prior data suggest superior image quality with slower gadolinium injection rates than typically used.

Human whole-blood1H2O longitudinal relaxation with normal and high-relaxivity contrast reagents: Influence of trans-cell-membrane water exchange: CR Longitudinal Relaxivities in Whole Blood

Accurate characterization of contrast reagent (CR) longitudinal relaxivity in whole blood is required to predict arterial signal intensity in contrast‐enhanced MR angiography (CE‐MRA). This study measured the longitudinal relaxation rate constants (R1) over a concentration range for non‐protein‐binding and protein‐binding CRs in ex vivo whole blood and plasma at 1.5 and 3.0 Tesla (T) under physiologic arterial conditions.

Human whole-blood 1 H 2 O longitudinal relaxation with normal and high-relaxivity contrast reagents

Accurate characterization of contrast reagent (CR) longitudinal relaxivity in whole blood is required to predict arterial signal intensity in contrast‐enhanced MR angiography (CE‐MRA). This study measured the longitudinal relaxation rate constants (R1) over a concentration range for non‐protein‐binding and protein‐binding CRs in ex vivo whole blood and plasma at 1.5 and 3.0 Tesla (T) under physiologic arterial conditions.

Measuring aortic diameter with different MR techniques

To compare nongated three‐dimensional (3D) contrast‐enhanced magnetic resonance angiography (CE‐MRA) with 3D‐navigated cardiac‐gated steady‐state free‐precession bright blood (3D‐nav SSFP) and noncontrast 2D techniques for ascending aorta dimension measurements.