Highlights of the 16th annual scientific sessions of the Society for Cardiovascular Magnetic Resonance

Many advances presented this year dealt with technical developments with the objective of increasing the spatial and/or temporal resolution of CMR images as well as decreasing image acquisition time. One major focus was how to implement new methods in order to acquire real-time non triggered images allowing much more simple and robust imaging as well as permitting the use of CMR to guide percutaneous interventions [13]. Aneja et al. demonstrated that real-time CMR using non-optimized sequences with lower spatial and temporal resolution already derives similar results to traditional breath-held images in patients with regular cardiac rhythm [14]. Although this approach was limited to a small number of patients and using standard sequences only, the principles of real-time imaging are rapidly gaining acceptance as CMR increases its clinical utility.

With that objective in mind, an increasing number of abstracts focused on new sequences that take advantage of undersampling the k-space with spatiotemporal modeling such as compressed sensing [15] to increase current resolution limits. Using sparse and incoherent sampling with multiple iterative reconstructions Schmidt et al. showed that these new techniques provide high spatial (2.4 × 1.7 × 6.0 mm) as well as temporal resolution (33 ms) within 1 heartbeat acquisition in real time [16]. Introducing a new regional sparsity reconstruction method, Chen et al. also demonstrated the feasibility of compressed sensing for obtaining in vivo perfusion images, allowing for increase in ventricular coverage or spatial resolution as previous techniques have shown [17, 18].

Non-cartesian acquisitions have also seen a gain in use and moved from the preliminary work in coronary artery imaging to other applications in CMR [19]. Sharif et al. used an ungated radial acquisition method to acquire first-pass perfusion images simultaneously combined with cine images for assessment of perfusion and wall motion within the same scan [20]. Yang et al. also used a combination of spiral trajectories with compressed sensing to obtain full ventricular coverage during first pass perfusion with the acquisition of 8 slices in 480 ms with 1.75 × 1.75 mm in-plane resolution allowing for heart rates up to 125 beats per minute during stress imaging [21]. In another abstract by Petersson et al. the authors demonstrated the use of spiral k-space trajectories to acquire 4D flow stacks with a significant 62% reduction in scan time while maintaining similar image quality compared to traditional readout methods [22].

Another highlight of the meeting was in the developments of new methods to overcome old problems in CMR such as dark rim artifacts, off-resonance effects and banding [23]. The Early Career Award in Basic Sciences was awarded to Sharif et al. for an abstract that proposed to reduce dark-rim artifacts during first-pass perfusion by using a single-shot radial sequence in comparison to the routine Cartesian technique with a significant decrease in both artifact score as well as in its maximum width (Figure 2) [24]. Problems with 3 T MRI were also addressed by Wu et al. who proposed a shorter TR for SSFP cine imaging through a reduced RF pulse duration and readout showing that the new approach resulted in images more robust to field in homogeneity at this higher field strength with minimum reduction in spatial resolution while allowing for elimination of frequency scouts and more rapid breath-hold times [25]. Another way to improve susceptibility artifacts without the need to use time consuming frequency adjustments was presented by Rothstein et al. who proposed a new 3D shimming algorithm which resulted in similar results for volumes and function evaluation keeping the same levels of banding artifacts as frequency offsets but with a more simplified workflow [26].

With the rapidly increasing body of new clinical data on T1 mapping, more understanding of the fundamental principles of the sequences used to generate these maps is essential [27]. A comprehensive series of lectures on T1 mapping in the heart was presented by Peter Kellman, Martin Ugander, Erica Dall’ Armellina and Andrew Flett (http://​scmr.​org/​Education/​CMR-online-video-on-demand-lectures/​3072/​3151/​3154.​html).This provided a clear explanation of the pros and cons of each of the different sequences together with the importance of both accuracy and precision of measurement. The potential clinical applications of T1 mapping including the measurement of extracellular volume fraction to visualize the extent and severity of abnormalities within the myocardium were outlined in detail.

Almost 10% of the submitted abstracts for the meeting involved T1 mapping. Many sessions discussed the current use of T1 maps and the need to standardize and optimize the acquisition techniques to pre- and post-contrast maps, the use of bolus versus continuous infusion of gadolinium as well as the different approaches to heart rate variability and types of sequences. Fitts et al. presented a recent arrhythmia-insensitive rapid T1 mapping sequence to overcome some of the limitations of the most widely used modified Look-Locker inversion recovery (MOLLI) techniques [28]. The sequence is based on saturation recovery magnetization and uses centric k-space ordering resulting in smaller coefficients of variation and higher T1 values in patients with high heart rates or arrhythmias. In another study using phantoms, Slavin et al. also showed that using a saturation recovery pulse but with a balanced SSFP readout they could find more accurate T1 values compared to MOLLI due to its insensitivity to beat-to-beat variations and lack of T2 dependence [29]. In a comparison of different sequences including the classical MOLLI sequence, a shortened (sh)MOLLI acquisition with 9 heartbeats [30], a 3′5 MOLLI and a saturation recovery single shot acquisition sequence (SASHA) the authors reported that the 3′5 MOLLI sequence was more closely associated to the original MOLLI T1 values compared to the two other sequences while the shMOLLI technique underestimated T1 values by 13% and the SASHA sequence overestimated the T1 values by 17%. Finally, a 3D technique acquired with free breathing was also presented by Henningsson et al. with the average scan time of 9:37 minutes to cover the whole left ventricle with a resolution of 1.5 × 1.5 × 8 mm³ using SASHA resulting in an increase in SNR compared to 2D-SASHA images with similar T1 values [31].

The development of rapid prototyping (or 3D printing) has allowed highly accurate models of vascular structures to be generated [32]. Acevedo-Bolton et al. described how complex intracranial aneurysms imaged with magnetic resonance could be recreated in order to test the effectiveness of flow-diverting stents [33]. This work not only enables the operator to gain valuable experience on a ‘dry run’ but allows assessment of the pre- and post-procedure flow dynamics using 4D flow.

Three other oral presentations described the utility of the quiescent-interval single shot or ‘QISS’ sequence for non-contrast MR angiography. The technique (with various alterations including radial vs. Cartesian geometry k-space acquisition) shows excellent performance and correlates very well with contrast exams in patients with peripheral arterial disease. These angiographic techniques have the potential to be combined with more functional assessment (such as rest and stress calf perfusion) to assess ischemia and response to treatment [34]. Clement-Guinaudeauet al presented data on a calf muscle perfusion index which correlated very well with peak walking time and claudication onset time [35]. Using a simple MR-compatible device to allow exercise stress in the scanner bore, this has useful potential as a surrogate endpoint for clinical trials.

With the publication of CE-MARC [41] and MR-IMPACT 2 [2] vasodilator stress CMR has been shown in prospectively designed studies to be a reasonable alternative to stress single-photon emission computed tomography (SPECT) for the detection of significant coronary artery disease. It has additionally been shown that performing stress CMR to identify patients who benefit from coronary angiography offers significant savings when compared to directly performing coronary angiography. John Greenwood adds to this by presenting new data from the CE-MARC trial suggesting that strategies the utilize stress CMR first followed by invasive coronary angiography if the CMR study is abnormal or inconclusive may be more cost-effective when compared to a SPECT first strategy [42]. Similarly, Sebastian Kelle showed the cost-effectiveness of strategies for the assessment of stable coronary artery disease that utilize dobutamine stress to identify patients who might benefit from a coronary angiogram rather than sending all patients directly to cardiac catheterization [43]. Using a Markov analysis, Steffen Petersen presented data suggesting that the cost-effectiveness of a diagnostic strategy was dependent on the pretest probability of having coronary artery disease. His analysis suggested stress CMR strategies were most cost effective in patients with a higher pretest probability and that cardiac computed tomography strategies were more cost-effective in patients with a low to intermediate pretest probability [44].

An important limitation of stress CMR is the incomplete coverage of the left ventricle and may prevent adequate assessment of ischemic burden. Sven Plein compared the ischemic and scar burden as determined by a standard clinical stress CMR technique to SPECT [45]. He found that CMR measures significantly less scar but more ischemia than SPECT possibly due to the incomplete coverage of perfusion CMR. In a prospective, multicenter trial, Robert Manka showed that whole heart, three-dimensional myocardial perfusion imaging (Figure 3) correlated highly with fractional flow reserve [46].It remains to be seen whether such a technique will improve the quantification of ischemic burden and whether the burden of ischemia as measured by CMR can identify patients who will benefit from revascularization in addition to aggressive medical therapy alone.

The role of CMR in the assessment and risk stratification of non-ischemic heart disease is well established [47]. Currently available data detail the ability of CMR to risk stratify patients with a variety of non-ischemic heart diseases such as dilated cardiomyopathy, myocardial siderosis, and hypertrophic cardiomyopathy (HCM) [48]. However, large outcome studies remain challenging. Ismail et al. presented data on 711 consecutive patients with HCM with a total of 2852 patient-years of follow-up [49]. Although the amount of fibrosis was a significant univariate predictor of sudden cardiac death risk, on multivariate analysis, only left ventricular ejection fraction remained a significant predictor. The understanding of the HCM phenotype is becoming ever more complex [48, 50]. Deva explained the incremental predictive value of inferoseptal crypts [51] and Flett proposed a previously unrecognized pattern of relative apical hypertrophy which may be an early form of HCM as the group found many features of HCM in patients with unexplained T-wave inversion [52]. Christopher Kramer unveiled the details of a proposed large scale, multiparametric, multicenter, observational registry with the goal of better defining the role of CMR in the risk stratification of the HCM patient.

As previously mentioned, T1 mapping featured highly throughout the sessions. Sado et al. presented findings of low pre-contrast T1 values in Anderson-Fabry disease and iron overload which may help to define pre-clinical phenotypes more accurately [53]. Whereas the assessment of diastolic function has traditionally been cumbersome and time-consuming with CMR, Simpson et al. have developed a new high spatial and temporal resolution tissue phase contrast technique which allows much shorter acquisitions and brings this nearer to clinical prime-time [54]. In an invited lecture, Michael Markl reviewed promising techniques such as tissue phase mapping with respiratory navigator gating, 2-dimensional and 3-dimensional Displacement Encoding with Stimulated Echoes (DENSE), and strain encoded imaging (SENC) for assessing myocardial mechanics and quantifying the various components of myocardial strain (http://​scmr.​org/​Education/​CMR-online-video-on-demand-lectures/​3072/​3157/​3160.​html). During the same session, Alicia Maceira, Amit Patel, and Sophie Mavrogeni suggested that CMR with late gadolinium enhancement imaging plays an important role for the detection of cardiac involvement and risk stratification of patients with systemic disorders such as amyloidosis, vasculitis, and other rheumatologic disorders such as systemic lupus erythematosis, sarcoidosis, rheumatoid arthritis, and systemic sclerosis (http://​scmr.​org/​Education/​CMR-online-video-on-demand-lectures/​3072/​3157/​3161.​html).

Francisco Leyva reviewed the role of CMR for identifying patients that may benefit from defibrillator implantation and resynchronization therapy (http://​scmr.​org/​Education/​CMR-online-video-on-demand-lectures/​3072/​3163/​3164.​html). Several investigators have shown that the extent of late gadolinium enhancement is strongly associated with need for future ICD therapy [55]. The burden of late gadolinium enhancement is also associated with likelihood of responding to resynchronization therapy; in fact, using late gadolinium enhancement images to guide left ventricular lead placement improves outcomes following biventricular pacing [56].

The continued development of real-time imaging for interventional CMR procedures is an important area of research. Tse et al presented preliminary animal data on a voltage-based electroanatomic tracking system which can be used to measure intracardiac electrograms both in and out of the scanner bore, including during imaging sequences. Radiofrequency signals from patches on the skin allow the operator to track the navigation of a catheter tip without the need for co-registration [57]. Ramanan et al. went on to describe how catheter ablation lesions could be mapped acutely using a combination of T2 and T1* mapping with late gadolinium enhancement [58]. This has the potential for verifying the effectiveness of ablation by differentiating the actual lesion area from surrounding oedema. On a different subject, Ainslie et al. reported interim data from the multiparametric CMR assessment on RV apical versus septal pacing study (MAPS)which aims to determine the optimal site for right ventricular pacing using CMR, echo, cardiopulmonary exercise testing and biomarkers [59]. Patients with persistent atrial fibrillation requiring AV node ablation and permanent pacemaker implants were fitted with an MR-conditional St. Jude pacemaker and two RV endocardial leads placed at the RV apex and in the septum (one in the ventricular port and one in the atrial port of the device). Using a crossover study design with ventricular pacing at each of the sites, CMR measurements were taken at three time points. Early results have shown that not only is cardiac scanning safe in these patients but that there is higher aortic and pulmonary flow as well as less dyssynchrony with septal pacing.

CMR continues to show its superiority over other imaging modalities in the field of congenital cardiovascular disease. This year cases for one of the congenital heart case review sessions were competitively selected from open competition with blind and independent ratings from several expert reviewers. The resultant session was popular, interactive and of high quality. The benefits of CMR in congenital heart disease, not least with regards accurate anatomical delineation, were well illustrated. In addition, refinements in flow imaging are improving our understanding not only in complex congenital disease but also in more routine cases. Muzzarelli et al. presented their data on the optimal measurement of aortic flow in bicuspid aortic valve disease [60] and Hart et al. demonstrated how 4D flow can give a much greater insight into the understanding of haemodynamics of the Fontan circulation during the breathing cycle [61]. Although acquisitions can be long and complex post-processing is required, 4D flow techniques have the potential to develop into a useful clinical tool [62].

As in previous years, a highly popular component of the program involved a series of 10 case review sessions, each discussing a range of interesting cases including complex, challenging problems and more mundane day-to-day cases. A lively, interactive debate was generated between delegates and experts in the field. The SCMR website publishes a regular featured ‘case of the week’ and the best 5 of these cases from 2012 were presented in full (http://​scmr.​org/​caseoftheweek.​html).