Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights. RGNLRESEARCH ORIGINAL Diagnostic Performance of Myocardial Perfusion MR at 3 T Ⅲ in Patients with Coronary IMAGING CARDIAC Disease1

Rolf Gebker, MD Purpose: To prospectively determine the diagnostic performance of Cosima Jahnke, MD myocardial perfusion magnetic resonance (MR) imaging at Ingo Paetsch, MD 3 T for helping depict clinically relevant coronary artery Sebastian Kelle, MD (Ն50% diameter) in patients with suspected or Bernhard Schnackenburg, PhD known (CAD), with coronary an- Eckart Fleck, MD giography as the reference standard. Eike Nagel, MD Materials and The study was approved by the local ethics committee; Methods: written informed consent was obtained. Vasodilator stress perfusion imaging by using a turbo field-echo sequence was obtained in 101 patients (71 men, 30 women; mean age, 62 years Ϯ 7.7 [standard deviation]) scheduled for coro- nary . Myocardial ischemia was defined as stress-inducible perfusion deficit in arterial territories without delayed enhancement (DE) or additional stress- inducible perfusion deficit in territories with nontransmu- ral DE. Images were evaluated in consensus by two blinded readers. Diagnostic performance was determined on per- patient and per–coronary artery territory bases. The num- ber of dark rim artifacts in patients without DE was deter- mined in a second read. Interobserver variability was as- sessed in 40 randomly selected patients.

Results: One hundred one patients underwent MR examinations. Coronary angiography depicted relevant stenosis in 70 (69%) patients. Patient-based sensitivity and specificity were 90% and 71%, respectively. Sensitivity, specificity, and diagnostic accuracy for the detection of coronary ste- nosis in a specific territory were 76%, 89%, and 86%, respectively. In 24% of patients without DE, dark rim artifacts were detected, mostly in the left anterior de- scending artery territory (56%). Among 40 randomly se- lected patients, there was agreement in the determination of myocardial perfusion deficits in 37 (93%, ␬ϭ0.79) patients.

Conclusion: Myocardial perfusion MR imaging by using saturation-re- covery spoiled gradient-echo imaging at3Thasanaccu- racy of 84% for depicting hemodynamically relevant coro- nary artery stenosis in patients with suspected and known 1 From the Department of Internal Medicine/, CAD. German Institute Berlin, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, 13353 Berlin, Germany (R.G., C.J., I.P., S.K., E.F., E.N.); and Clinical Science, ௠ RSNA, 2008 Philips Medical Systems, Hamburg, Germany (B.S.). Re- ceived March 31, 2007; revision requested May 31; revi- sion received July 30; accepted August 20; final version accepted September 24. Address correspondence to R.G. (e-mail: [email protected] ).

஽ RSNA, 2008

Radiology: Volume 247: Number 1—April 2008 57 : Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

iven the pathophysiologic events We hypothesized that by using a nosine-related side effects (n ϭ 2) and during the ischemic cascade, the combined MR imaging approach with (b) poor-quality MR images (n ϭ 1) Gassessment of myocardial perfu- stress-rest first-pass perfusion imaging, (Fig 1). Of the patients with adverse sion appears to be the most promising followed by delayed-enhancement (DE) side effects, one developed severe dys- concept for a noninvasive test to help imaging at 3 T, it is possible to detect pnea and requested termination of the detect hemodynamically relevant coro- relevant coronary stenosis in patients examination; the dyspnea resolved after nary artery disease (CAD) (1). Myocar- with clinically suspected or known 30 seconds without the administration dial perfusion assessment is frequently CAD. Thus, the purpose of our study of theophylline (Bronchoparat; Astellas performed by using stress and rest sin- was to prospectively determine the di- Pharma, Munich, Germany). The other gle photon emission computed tomogra- agnostic performance of myocardial patient developed bradycardia of 30 phy (SPECT) or positron emission to- perfusion MR imaging at 3 T for depict- beats per minute, with atrioventricular mography (PET). However, both mo- ing clinically relevant coronary artery node block being the most likely cause; dalities expose patients to radiation, stenosis (Ն50% diameter stenosis) in 15 seconds after the termination of the SPECT is limited by attenuation arti- patients with suspected or known CAD, adenosine infusion, heart rate returned facts and reduced spatial resolution, with coronary angiography as the refer- to normal. Thus, a total of 101 patients and PET is limited by its reduced avail- ence standard. were evaluated in our study (Table 1). ability. Myocardial perfusion magnetic res- MR Examination onance (MR) imaging has been devel- Materials and Methods MR sequence design.—All patients oped extensively over the past decade. were examined in the supine position by So far, clinical trials have mostly been Patients using a 3-T whole-body imager (Achieva performed with 1.5-T scanners and The study was approved by the Charite´ 3T; Philips, Best, the Netherlands) have shown that MR perfusion yields and Virchow-Klinikum Ethics Commit- equipped with a dual gradient system diagnostic results for the detection of tee. Written informed consent was given (Quasar; Philips) by using gradient CAD comparable to those of clinically by all patients. One hundred four pa- strength of 80 mT/m and slew rate of established nuclear techniques such as tients (72 men, 32 women; mean age, 200 T/m/sec. A six-element cardiac syn- SPECT or PET (2–4). When using coro- 62 years Ϯ 7.7 [standard deviation]) ergy coil was used for signal detection. nary angiography as the reference stan- scheduled for clinically indicated coro- Cardiac synchronization was performed dard (5–9), MR perfusion imaging is as- nary angiography were examined be- by using a four-electrode vector electro- sociated with good diagnostic accuracy tween August 2005 and July 2006. cardiogram, and image acquisitions (sensitivity, 88%–93%; specificity, 75%– Patients with contraindications to were triggered on the R wave (11). 90%). Recently, MR imaging at3Thas either MR imaging (noncompatible bio- For cine imaging, a balanced steady- become available and potentially pro- metallic implants or claustrophobia) or state free precession sequence was used vides a substantial improvement of tis- adenosine-related side effects (atrio- (repetition time msec/echo time msec, sue contrast in T1-weighted imaging ventricular node block grade Ͼ I, acute 3/1.5; flip angle, 40°). A saturation pre- techniques relying on gadolinium-based coronary syndrome, severe hyperten- contrast material enhancement. Yet, sion, asthma) and patients with ar- 3-T systems in general are known to be rhythmia were not considered for study Published online before print prone to imaging artifacts (eg, suscepti- inclusion. All patients were instructed 10.1148/radiol.2471070596 bility artifacts) (10). to refrain from any beverages or foods 2008; 247:57–63 containing caffeine within 24 hours be- fore the MR study. The pertinent medi- Abbreviations: Advances in Knowledge cal history of the patients was recorded CAD ϭ coronary artery disease ϭ Ⅲ Myocardial stress perfusion imag- at the time of the examination by one of DE delayed enhancement LAD ϭ left anterior descending artery ing at 3 T has an accuracy of three authors (R.G., C.J., or I.P.). LCX ϭ left circumflex artery 84%–86% to detect relevant cor- Three of 104 patients could not be RCA ϭ right coronary artery onary artery stenoses. evaluated owing to (a) adverse ade- Ⅲ Dark rim artifacts represented by Author contributions: Guarantors of integrity of entire study, R.G., E.N.; study matched perfusion defects be- Implication for Patient Care concepts/study design or data acquisition or data analy- tween stress and rest imaging oc- sis/interpretation, all authors; manuscript drafting or cur in almost one-fourth of pa- Ⅲ Myocardial perfusion MR imaging manuscript revision for important intellectual content, all tients without evidence of delayed at3Tisanoption for patients authors; approval of final version of submitted manu- enhancement. with suspected or known coro- script, all authors; literature research, R.G., I.P., B.S., Ⅲ There is high interobserver agree- nary artery disease to help depict E.F., E.N.; clinical studies, R.G., C.J., I.P., S.K.; statistical analysis, R.G.; and manuscript editing, all authors ment (␬ϭ0.79) for the detection hemodynamically relevant coro- of perfusion deficits. nary stenosis in patients. Authors stated no financial relationship to disclose.

58 Radiology: Volume 247: Number 1—April 2008 CARDIAC IMAGING: Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

pared (prepulse delay, 95 msec) single- their breath as long as possible during defects were determined qualitatively shot spoiled gradient-echo sequence imaging and to continue breathing shal- by using subjective visualization. Ische- was used for perfusion imaging. The ac- lowly when they could no longer hold mia was defined as (a) territories with- quisition parameters were as follows: their breath. out DE that showed a perfusion deficit 2.8/0.9; flip angle, 18°; field of view, 5. Rest perfusion imaging by using 380 ϫ 380 mm; turbo factor, 59; and geometry identical to the short-axis cine Figure 1 matrix, 128 ϫ 128. For 60 consecutive views and giving a bolus of contrast cardiac cycles, three short-axis sections agent identical to that given during with a spatial resolution of 2.9 ϫ 2.9 ϫ stress imaging after 15 minutes to allow 8 mm recorded every heartbeat up to for myocardial washout. a heart rate of 110 beats per minute. 6. An additional bolus of 0.15 DE imaging was performed by using mmol/kg gadopentetate dimeglumine an inversion-prepared three-dimen- immediately after the rest perfusion sional spoiled gradient-echo sequence scan and standard DE gadolinium-based (1.5 ϫ 1.7 ϫ 5 mm). imaging performed 10 minutes later. Imaging protocol.—All patients re- ceived two 18-gauge intravenous lines MR Image Analysis to allow separate administration of Images were evaluated in consensus by adenosine and the contrast agent to pre- two experienced readers (C.J. and vent high-grade atrioventricular block- R.G., both with 5 years experience in ade during the injection of the bolus. cardiac MR) who were blinded to the The patients underwent a standard MR patientsЈ history and angiographic re- examination that included the following sults. The readers were presented with Figure 1: Flow diagram of patient inclusion. steps: anonymous MR cases, including perfu- AV ϭ atrioventricular. 1. Standardized planning to deter- sion at stress, at rest, and DE. Perfusion mine the actual short axis of the left . 2. Cine imaging of three short-axis Table 1 views and three long-axis views (four-, Patient Demographics two-, and three-chamber views). The Characteristic No. of Patients three short-axis views were distributed to cover the heart at the basal, equato- Age (y) 62 Ϯ 7.6 (44–80)*† rial, and apical positions by adjusting Sex (male/female) 71/30 the gap between the sections. The dis- Body mass index 27 Ϯ 4* tance between the apical section and Risk factors and patient history the apex, as well as the basal section Hypertension 87 (86) and the mitral valve, were identical. Hypercholesterolemia 70 (69) 3. Perfusion test imaging by using Smoking 42 (42) geometry identical to that of the short- Diabetes mellitus 26 (26) axis cine views to carefully exclude any Family history 27 (27) wraparound or trigger artifacts before Suspected CAD 45 (45) starting the actual index test. Known CAD 56 (55) Prior 35 (35) 4. Stress perfusion imaging by using Prior percutaneous coronary intervention 53 (52) the described perfusion sequence. Adeno- Prior coronary artery bypass graft 9 (9) sine (140 ␮g/min/kg) was given for a CAD classification 70 (69) total of 4 minutes. Scanning was started One vessel 25 (36) in the last minute of vasodilator stress Two vessel 32 (46) by using an intravenous bolus of 0.025 Three vessel 13 (19) mmol/kg of gadopentetate dimeglumine None 31 (31) (Magnevist; Schering, Berlin, Germany) Therapy following MR at an injection rate of 4 mL/sec followed Percutaneous coronary intervention 62 (61) by a flush of 20 mL of saline solution at Coronary artery bypass graft 7 (7) the same rate. and heart rate were monitored during the admin- Note.—Numbers in parentheses are percentages, except where noted. Ϯ istration of adenosine and the contrast * Data are the mean standard deviation. † Numbers in parentheses are the range. agent. Patients were instructed to hold

Radiology: Volume 247: Number 1—April 2008 59 CARDIAC IMAGING: Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

of 25% or more of the transmural ex- To assess interobserver variability ease was found in 25% (25 of 101), 32% tent during stress perfusion, but not at for interpretation of perfusion imaging, (32 of 101), and 13% (13 of 101) of rest (stress-inducible deficit), for three two independent observers (I.P. and patients, respectively. The most severe or more consecutive image frames and E.N., with 9 and 11 years experience in stenosis was assigned to the LAD in (b) territories with nontransmural DE cardiac MR, respectively) scored perfu- 41% (29 of 70), to the LCX in 23% (16 demonstrating additional stress-induc- sion imaging qualitatively on the basis of of 70), and to the RCA in 36% (25 of 70) ible perfusion deficits. Myocardial terri- reading criteria mentioned above in a of patients with CAD. tories were assigned to the three major randomly selected sample of 40 studies. coronary according to standard Statistical Analysis definitions (12). The readers were in- Coronary Angiography: Reference Statistical analysis was performed by structed to consider defects as artifacts Standard using computer software (SPSS, release if they had an identical extent during All patients underwent conventional 12.0.1; SPSS, Chicago, Ill). For all con- stress and rest in territories without within tinuous parameters, mean Ϯ standard DE, if defects appeared in a nonphysi- 24–48 hours after the MR examina- deviation is given. Student t test was ologic pattern in the epicardium, if de- tion by using a standard Judkins tech- used to assess significance of continuous fects occurred in less than 25% of the nique. A reduction of the luminal di- variables. ␬ values were calculated to transmural extent of the myocardium ameter 50% or more in a major epi- compare interobserver agreement on a during maximum signal intensity of the cardial coronary artery or the major per-patient basis for myocardial territo- left ventricular blood pool before maxi- branches (Ն2.5 mm) was considered ries. Sensitivity, specificity, diagnostic mum signal intensity in the myocardium to be a relevant stenosis. The angio- accuracy, and negative and positive pre- was reached, or if they appeared only graphic results were classified as one-, dictive values were calculated according briefly (less than three consecutive two-, or three-vessel disease or exclu- to standard definitions. A P value of less frames). The studies were analyzed on a sion of relevant CAD. In patients with than .05 was considered to indicate a per-patient and per-territory basis. CAD, the most severe stenosis was significant difference. As a measure of image quality, we defined by an experienced blinded in- determined the amount of dark rim ar- terventionalist (one of three treating tifacts in the subendocardium associ- physicians with 8–25 years experience Results ated with 3-T perfusion imaging, which in coronary angiography) and as- can pose a substantial challenge to mak- signed to a coronary artery territory Patient Group ing a diagnosis. The blinded perfusion (left anterior descending artery In the 101 patients (303 coronary terri- images of those patients without evi- [LAD], left circumflex artery [LCX], tories) available for comparison be- dence of DE were analyzed again in a or right coronary artery [RCA]) tween coronary angiography and MR second read 4 weeks after the first ses- deemed responsible for the occur- perfusion imaging, hemodynamic re- sion by two readers (C.J. and R.G.) in rence of a perfusion defect. Similarly, sponses were appropriate, such as an consensus to determine cases that in patients with bypass grafts, relevant increase in heart rate of 30% and a mild showed territories with perfusion de- arterial or graft stenoses were decrease in systolic and diastolic blood fects that had similar intensity and ex- considered responsible for ischemia in pressure (Table 2). tent during stress and rest (matched the myocardial territory supplied by perfusion deficit). In the absence of DE, the recipient native coronary vessel. Diagnostic Performance of MR these cases were considered to repre- Relevant coronary artery stenoses Per-patient analysis.—The overall patient- sent artifacts. The territory distribution were present in 69% (70 of 101) of pa- based sensitivity, specificity, and diagnos- of these artifacts was noted. tients; one-, two-, and three-vessel dis- tic accuracy for the detection of coronary artery stenosis (Ն50%) were 90% (63 of Table 2 70), 71% (22 of 31), and 84%, respec- tively (Figs 2 and 3; Table 3). Negative Hemodynamic Data and positive predictive values were 76% Characteristic Rest Stress and 88%, respectively. Heart rate (beats per minute) 66 Ϯ 988Ϯ 11* In patients with one-vessel dis- Blood pressure ease, patient-based sensitivity was Systolic (mm Hg) 139 Ϯ 23 137 Ϯ 23 88% (22 of 25); in two-vessel disease, Diastolic (mm Hg) 76 Ϯ 973Ϯ 11* sensitivity was 88% (28 of 32); and in Pulse pressure product† 9020 Ϯ 2512 11 800 Ϯ 2966* three-vessel disease, sensitivity was 100% (13 of 13). In patients with prior Ϯ Note.—Data are the mean standard deviation. myocardial infarction, sensitivity was * P Ͻ .001. 89% (31 of 35). † Measured as beats per minute times millimeters of mercury. Per-territory analysis.—In a per-

60 Radiology: Volume 247: Number 1—April 2008 CARDIAC IMAGING: Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

Figure 2 territory analysis, respective sensitivity and specificity for the LAD were 83% (24 of 29) and 86% (62 of 72); for the LCX, 75% (12 of 16) and 91% (77 of 85); and for the RCA, 68% (17 of 25) and 91% (69 of 76) (Table 3). Respec- tive negative and positive predictive val- ues for LAD were 93% and 71%; LCX, 95% and 60%; and RCA, 90% and 71%. Sensitivity, specificity, and diagnostic accuracy for the detection of relevant coronary stenosis for all coronary arte- rial territories were 76% (53 of 70), 89% (208 of 233), and 86%, respec- tively. In patients with one-vessel disease, the sensitivity for the detection of an Figure 2: Patient with stress-inducible anterior and anteroseptal perfusion defect (black arrows). Scans ischemic territory was 84% (21 of 25). show A, apical and B, equatorial short-axis views during stress; C and D show perfusion images at rest (2.8/ Of 74 patients suspected of having 0.9; flip angle, 18°). E, Coronary angiography shows 90% stenosis (white arrow) of proximal LAD. relevant CAD with no history of myo- cardial infarction and no electrocardio- graphic changes suggesting myocardial infarction, eight patients had evidence Figure 3 of DE.

Artifacts In 24% (16 of 66) of patients without evidence of DE, territories with a matched perfusion defect during stress and rest were detected, thus representing arti- facts (Fig 4). In a per-territory analy- sis, 56% (nine of 16) of the artifacts were located in the LAD territory, 13% (two of 16) in the LCX territory, and 31% (five of 16) in the RCA terri- tory. In 94% (15 of 16) of patients, no substantial stenosis was detected in the territory of these vessels during coronary angiography.

Interobserver Agreement Of 40 randomly selected patients, there was agreement in the determination of myocardial perfusion deficits in 37 (93%, ␬ϭ0.79) on a per-patient basis. For coro- Figure 3: Short-axis views of patient with stress-inducible perfusion defect show apical septum (A, black nary territories, there was concordance in arrow) and basal inferior wall (B, white arrow); images during rest (C, D) do not (2.8/0.9; flip angle, 18°). Cor- 34 (85%, ␬ϭ0.63) patients for the LAD, onary angiograms shows distal occlusion of LAD (E, black arrow) and two subtotal stenoses (F, black arrows) 34 (85%, ␬ϭ0.67) patients for the LCX, of mid and distal RCA. and 33 (83%, ␬ϭ0.63) patients for the RCA. The principal findings of our study are fects between stress and rest imaging that (a) myocardial stress perfusion im- occur in almost one-fourth of patients Discussion aging at 3 T has an accuracy of 84%– without evidence of DE; and (c) there is In our study, we evaluated myocardial 86% to help detect relevant coronary high interobserver agreement for the perfusion MR imaging at3Tinapopu- artery stenoses; (b) dark rim artifacts detection of perfusion deficits. lation with suspected and known CAD. represented by matched perfusion de- Myocardial perfusion imaging with

Radiology: Volume 247: Number 1—April 2008 61 CARDIAC IMAGING: Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

1.5-T MR has evolved into a clinically We found that adenosine vasodila- with previous interventions, and prior useful technique to help detect myocar- tor MR perfusion imaging is a feasible myocardial infarction, while in previous dial ischemia while still experiencing a and safe stress test in the 3-T environ- studies, these patient groups have often low signal-to-noise ratio. Thus, the 3-T ment. Only two of 104 patients could been excluded, resulting in a higher platform is attractive for its immediate not be evaluated owing to adenosine- specificity (5,8). Second, the number of increase in magnetization by a factor of related side effects, including self-limit- patients with artifacts, represented by two, a higher signal-to-noise ratio, and ing dyspnea and bradycardia, which oc- matching stress-rest defects in patients better tissue contrast (13). Cardiac im- curred probably owing to transient without evidence of DE, was relatively aging at 3 T, however, is substantially atrioventricular node block. Both pa- high. When performing visual assess- different from imaging at 1.5 T because tients recovered within seconds after ment, perfusion defects must be differ- of increased susceptibility artifacts, dif- discontinuation of the adenosine infu- entiated from dark rim artifacts, which ferences in tissue relaxation, and radio- sion. This is in accordance with previ- may appear in the subendocardial layer frequency homogeneity issues. Because ously reported tolerance levels and of the myocardium. findings in initial small studies (14,15) safety profile of adenosine MR perfusion Recently, Klem et al (17) have shown have shown that MR perfusion at 3 T studies at 1.5 T (16). Poor-quality MR that by combining stress and rest perfu- provides improved contrast over a range of images led to the exclusion of only one sion imaging with DE imaging, the diag- gadopentetate dimeglumine doses and im- patient. Although our overall patient- nostic value significantly improves com- proved diagnostic accuracy compared based diagnostic accuracy of 84% is pared with perfusion imaging alone in a with perfusion imaging at 1.5 T, we comparable to previously reported re- population with a low prevalence of CAD. aimed to demonstrate its diagnostic sults from perfusion studies at 1.5 T This improvement was most pronounced performance in a larger patient cohort (8,9), our specificity was lower at 71%. in dark rim artifact spoiled perfusion im- that is typical for a tertiary center (in- There might be several reasons for this ages. The results of our study confirmed cluding an analysis of dark rim artifacts, finding. First, our patient population that the incorporation of a rest perfusion which play an important role in image consisted of a cohort with a high num- scan and DE is helpful in distinguishing analysis). ber of hypertensive patients, patients potential perfusion defects from artifacts. A number of possible causes for the ap- Table 3 pearance of dark rim artifacts have been discussed, including susceptibility effects Diagnostic Performance of Myocardial Perfusion Imaging According to the Extent and of the highly concentrated contrast agent Location of Coronary Stenosis bolus during the first pass (18). We antic- Coronary Stenosis Ն 50% Sensitivity (%) Specificity (%) Accuracy (%) ipated this issue and lowered the contrast agent dose from 0.05 mmol/kg at 1.5 T to Patient-based total 90 (63/70) 71 (22/31) 84 0.025 mmol/kg for our study at 3 T. Nev- Coronary artery 76 (53/70) 89 (208/233) 86 ertheless, the amount of these artifacts in LAD 83 (24/29) 86 (62/72) 85 our study was higher compared with the LCX 75 (12/16) 91 (77/85) 88 results (24% vs 18%) reported by Klem RCA 68 (17/25) 91 (69/76) 85 et al (17). This might reflect the stronger

Note.—Numbers in parentheses are raw data. susceptibility effects caused by larger B0 inhomogeneities at 3 T when compared with 1.5 T, as previously reported. Most of the dark rim artifacts (56%) in our Figure 4 study were found in the LAD territory, which might be related to partial volume effects and susceptibility affecting the sep- tum, as it is located between the right and left ventricular cavities, both of which have high concentrations of gado- pentetate dimeglumine during the first pass of the contrast agent. Our study had limitations. The per- territory analysis does not take into ac- count the correlation owing to territo- ries being clustered within patients. We Figure 4: Short-axis views of patient with suspected CAD show subendocardial perfusion defect (arrow) did not perform quantitative or semi- during A, stress and B, rest (2.8/0.9; flip angle, 18°) in absence of scar (C) at DE, representing typical dark rim quantitative measurements, which have imaging artifact. Patient had no relevant stenosis at coronary angiography. been reported to lead to increased spec-

62 Radiology: Volume 247: Number 1—April 2008 CARDIAC IMAGING: Diagnostic Performance of Myocardial Perfusion MR at 3 T Gebker et al

ificity of myocardial perfusion imaging compared with stress myocardial perfusion mittee of the Council on Clinical Cardiology (8). The drawback of quantitative anal- . AJR Am J Roentgenol 2005; of the American Heart Association. Circula- ysis is that it is time consuming and not 185(1):95–102. tion 2002;105(4):539–542. ideal for day-to-day clinical purposes. 4. Schwitter J, Nanz D, Kneifel S, et al. Assess- 13. Nayak KS, Cunningham CH, Santos JM, Our aim, however, was to establish a ment of myocardial perfusion in coronary Pauly JM. Real-time cardiac MRI at 3 tesla. clinically robust approach for myocar- artery disease by magnetic resonance: a Magn Reson Med 2004;51(4):655–660. comparison with positron emission tomogra- dial perfusion imaging at 3 T. 14. Araoz PA, Glockner JF, McGee KP, et al. 3 phy and coronary angiography. Circulation Tesla MR imaging provides improved con- Myocardial perfusion MR imaging 2001;103(18):2230–2235. by using saturation-recovery spoiled gradi- trast in first-pass myocardial perfusion imag- ing over a range of gadolinium doses. J Car- ent-echo imaging at 3 T has an accuracy 5. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myocardial ischemia from perfusion diovasc Magn Reson 2005;7(3):559–564. of 84%–86% for depicting hemodynam- reserve based on cardiovascular magnetic reso- ically relevant coronary stenosis in pa- 15. Karamitsos TD, Cheng AS, Pegg TJ, Robson nance. Circulation 2000;101(12):1379–1383. MD, Neubauer S, Selvanayagam JB. Myocar- tients with suspected or known CAD. 6. Giang TH, Nanz D, Coulden R, et al. Detection of dial stress perfusion imaging at 3T: compari- Recent studies on the causes of dark rim coronary artery disease by magnetic resonance son of semi-quantitative and qualitative as- artifacts have pointed out that low spa- myocardial perfusion imaging with various con- sessment in patients with suspected coro- tial and temporal resolutions are both trast medium doses: first European multi-centre nary artery disease. J Cardiovasc Magn associated with more pronounced arti- experience. Eur Heart J 2004;25(18):1657– Reson 2007;9(2):369–370. facts (19,20). Future efforts should thus 1665. 16. Miszalski-Jamka K, von der Recke G, be directed at improving both spatial 7. Ishida N, Sakuma H, Motoyasu M, et al. Hardung D, et al. A prospective study in 645 and temporal resolution in an attempt Noninfarcted myocardium: correlation be- consecutive patients to assess safety and fea- sibility of adenosine stress magnetic reso- to optimize image quality at 3-T perfu- tween dynamic first-pass contrast-enhanced nance perfusion for diagnosis of myocardial sion imaging. Multicenter trials incorpo- myocardial MR imaging and quantitative cor- onary angiography. Radiology 2003;229(1): ischaemia. J Cardiovasc Magn Reson 2006; rating a multimodality approach are 209–216. 8(1):43. needed to establish the clinical role of 17. Klem I, Heitner JF, Shah DJ, et al. Improved 8. Nagel E, Klein C, Paetsch I, et al. Magnetic cardiac MR with 3 T for the detection of detection of coronary artery disease by resonance perfusion measurements for the relevant myocardial ischemia. stress perfusion cardiovascular magnetic noninvasive detection of coronary artery dis- resonance with the use of delayed enhance- ease. Circulation 2003;108(4):432–437. References ment infarction imaging. J Am Coll Cardiol 9. Wolff SD, Schwitter J, Coulden R, et al. 2006;47(8):1630–1638. 1. Nesto RW, Kowalchuk GJ. The ischemic Myocardial first-pass perfusion magnetic 18. Arai AE. Magnetic resonance first-pass myo- cascade: temporal sequence of hemody- resonance imaging: a multicenter dose-rang- cardial perfusion imaging. Top Magn Reson namic, electrocardiographic and symptom- ing study. Circulation 2004;110(6):732–737. atic expressions of ischemia. Am J Cardiol Imaging 2000;11(6):383–398. 10. Bernstein MA, Huston J 3rd, Ward HA. Im- 1987;59(7):23C–30C. 19. Di Bella EV, Parker DL, Sinusas AJ. On the aging artifacts at 3.0T. J Magn Reson Imag- 2. McCrohon JA, Lyne JC, Rahman SL, Lorenz dark rim artifact in dynamic contrast-en- ing 2006;24(4):735–746. CH, Underwood SR, Pennell DJ. Adjunctive hanced MRI myocardial perfusion studies. role of cardiovascular magnetic resonance in 11. Schar M, Kozerke S, Fischer SE, Boesiger P. Magn Reson Med 2005;54(5):1295–1299. the assessment of patients with inferior at- Cardiac SSFP imaging at 3 Tesla. Magn Re- 20. Al-Saadi N, Torheim G, Reynolds G, De tenuation on myocardial perfusion SPECT. son Med 2004;51(4):799–806. Marchi D, Burchielli S, Lombardi M. Effects J Cardiovasc Magn Reson 2005;7(2):377– 12. Cerqueira MD, Weissman NJ, Dilsizian V, of spatial and temporal resolution on the 382. et al. Standardized myocardial segmentation dark rim artifact in first pass myocardial per- 3. Sakuma H, Suzawa N, Ichikawa Y, et al. and nomenclature for tomographic imaging fusion. J Cardiovasc Magn Reson 2007;9(2): Diagnostic accuracy of stress first-pass con- of the heart: a statement for healthcare pro- 173–174. trast-enhanced myocardial perfusion MRI fessionals from the Cardiac Imaging Com-

Radiology: Volume 247: Number 1—April 2008 63