Introduction 1 2 3 5 6 8 6 10 0 1 The aim of this study was to test whether CMR scanning at 3.0 T is safe and feasible in patients with coronary stents in the acute and chronic phase after MI, and to prospectively compare image quality at 3.0 T with the current clinical standard at 1.5 T. Furthermore, the presence and significance of different image artifacts are considered. Methods Patient population 11 12 CMR parameters Eighteen CMR scans were performed with a 3.0 T MR system (Intera, Philips, Best, The Netherlands), with a gradient performance of 30 mT/m and slew rate of 150 T/m/s, using a six element cardiac phased array surface coil. Another 18 CMR scans were acquired with a 1.5 T MR system (Magnetom Sonata, Siemens, Erlangen, Germany), with a gradient performance of 40 mT/m and slew rate of 200 T/m/s, using an eight element cardiac phased array surface coil. 3 3 3 3 Safety and quality analysis To evaluate patient safety, a physician was present at the CMR scanner throughout the scan. Heart rhythm was monitored continuously. The patient was asked to report any discomfort and symptoms during the scan procedure. Additionally, a repeat catheterization was performed in 16 patients who underwent CMR examination in the acute phase after MI (9 at 3.0 T, 7 at 1.5 T). Stent position and patency after CMR scanning were visually assessed on the repeat catheterization and compared to the primary PCI. Occurrence of repeat intervention or hospitalization within 30 days of the CMR examination was recorded. 13 1 Furthermore, image quality of the cine and LGE images were scored on a separate workstation (Centricity Radiology v6.1, GE Medical Systems, Zeist, the Netherlands) by four independent observers, who were blinded for MR system and clinical history. To distinguish between image quality that is satisfactory for clinical use or high quality images for research purposes, images were scored on a scale from 1 to 4: 1 not clinically useful; 2 clinically useful, but of insufficient quality for quantitative analysis; 3 clinically useful and of sufficient quality for quantitative analysis; 4 excellent quality. The following definitions were used for evaluation of the cine images: 1 poor quality, extensive artifacts, completely obscuring endocardial borders; 2 moderate quality, assessment of global function is possible, partly using assumptions, but regional wall thickening is not possible in all segments, due to interfering artifacts; 3 good quality, assessment of global and regional function is possible, despite some small artifacts; 4 excellent quality, functional analysis is possible and there is no interference of artifacts. And for the evaluation of LGE images: 1 poor quality, extensive artifacts, infarcted myocardium is not visible; 2 moderate quality, infarcted myocardium is visible, but delineation is not possible in all segments; 3 good quality, infarcted myocardium is easy to distinguish from viable myocardium, despite some small artifacts; 4 excellent quality, no artifacts. Two scores were given for each scan: 1 for the set of cine images and 1 for the set of LGE images. After completing the independent image quality assessment, all 4 observers exchanged their scores for each case and agreed on a consensus score. Statistical analysis t U P Results 1 Table 1 Patient characteristics Number of patients 25 Age (years) 55 ± 9 Men 21 (84) 2 25.5 ± 1.9 Risk factors Diabetes mellitus 0 (0) Hyperlipidaemia 3 (12) Hypertension 6 (24) Smoking 16 (64) Family history of CAD 8 (32) Maximum peak CK-MB (U/L) 308 (197–464) Infarct-related artery LAD 18 (72) RCx 3 (12) RCA 4 (16) Values are presented as number (ratio in %), mean ± standard deviation or median (25th–75th percentile) ECG electrodes were positioned and displaced until an optimal ECG signal was acquired. Although it took more time to obtain a stable ECG waveform with a clearly delineated R wave at 3.0 T than at 1.5 T (10–15 min vs. <2 min respectively), it did not affect image quality, since scans were repeated when trigger problems occurred. Stent safety and artifacts at 3.0 T No patient reported any discomfort or symptoms during the CMR scan procedure. In addition, there were no clinical events during, or shortly after scanning at 3.0 T, and none of the patients underwent a repeat intervention or hospitalization for any reason within 30 days. 2 Table 2 Type and number of coronary stents used per CMR examination Field strength and timing Number of patients a 3.0 T acute phase 9 1× Multi-Link Vision 1 2× Multi-Link Vision 2 1× Driver 3.0 T chronic phase 3 1× Multi-Link Vision 2 b 1 1× Driver 1.5 T acute phase 4 1× Prokinetic 1 2× Prokinetic 2 1× Lekton Motion 1 1× Lekton Motion/2× Multi-Link Vision 1 1× Multi-Link Vision 1 1× AVE 1 1× Multi-Link Zeta 1 b 1.5 T chronic phase 2 1× Lekton Motion 1 2× Lekton Motion 1 1× AVE 1 1× Multi-Link Zeta 1 b a b 1 Fig. 1 A B  1 Stent safety and artifacts at 1.5 T Also during scanning at 1.5 T, no patient reported any discomfort or symptoms, and there were no clinical events during, or shortly after scanning. None of the patients underwent a repeat intervention or hospitalization for any reason within 30 days after CMR scanning. 2 P 1 1 Evaluation of image quality 2 3 P Fig. 2 A C B D  Table 3 Consensus score of image quality of cine and late gadolinium-enhanced images for 1.5 T and 3.0 T  Cine SSFP images Late gadolinium-enhanced images 1.5 T 3.0 T 1.5 T 3.0 T 1. Not clinically useful 0 (0%) 0 (0%) 0 (0%) 2 (12%) 2. Clinically useful, no quantitative analysis 0 (0%) 8 (44%) 0 (0%) 7 (41%) 3. Clinically useful, quantitatively analyzable 0 (0%) 4 (22%) 1 (6%) 8 (47%) 4. Excellent quality 18 (100%) 6 (33%) 17 (94%) 0 (0%) Values are presented as absolute numbers (percentage) Discussion In this paper we report on our initial experience with CMR scanning at 3.0 T in a clinical situation. We found that it is safe and feasible to perform CMR scanning at 3.0 T in the acute and chronic phase after MI in patients treated with primary stenting. Image quality of cine imaging at 3.0 T is of sufficient quality for global assessment of left ventricular function, however, quantitative analysis is not possible in almost half of the patients, due to dark band, flow and stent artifacts. Image quality of LGE studies were significantly better at 1.5 T. Therefore further optimization of pulse sequences at 3.0 T is essential. 14 15 16 18 19 21 7 22 3 9 23 Fig. 3 A B A  24 clear image enhancement 25 4 Fig. 4 A B  26 27 28 In conclusion, our study demonstrates that it is safe and feasible to perform CMR scanning at 3.0 T in the acute and chronic phase after MI in patients treated with primary stenting. Although cine imaging at 3.0 T is of sufficient quality for clinical use, quantitative assessment is less reliable compared to 1.5 T, mainly due to dark band, flow and stent artifacts. Further optimization of pulse sequences at 3.0 T is essential to make 3.0 T CMR scanning suitable for clinical cardiology. Limitations The use of different MR systems, from different vendors, with different coils and sequence parameters, of course introduces confounding factors for a comparison. However, the sequences were optimized for their field strength to evaluate safety, feasibility and image quality rather than technical differences. A second limitation is that patients did not undergo a CMR examination at both 1.5 T and 3.0 T. Because the MR systems were on two different locations, it was not feasible to study the same patient twice in the acute phase after MI.