Background

Despite major advances in prevention, diagnosis and therapy over the past few decades, cardiovascular disease remains the leading cause of death in industrialized nations. Many powerful imaging modalities aimed at probing the health of the heart have emerged and have been further developed by both industry and academia. These include, but are not limited to echocardiography, positron emission tomography (PET), single-photon emission computed tomography (SPECT), x-ray angiography, computed tomography (CT), and magnetic resonance (MR). More recently, and owing to the fact that close to 50% of cardiovascular disease can be attributed to coronary heart disease, fractional flow reserve (FFR), a technique that includes invasive pressure measurements under x-ray guidance, has also emerged as a tool to determine the likelihood that a coronary stenosis impedes oxygen delivery to the heart muscle. However, many of the above techniques are either invasive or involve potentially harmful ionizing radiation. Therefore, MR provides an extremely attractive non-invasive alternative that simultaneously provides a high soft tissue contrast without x-ray exposure for neither operator nor patient, and that informs about anatomy, function, blood flow, and even tissue characteristics. However, MR signals are relatively weak, and the heart is subject to both cardiac and respiratory motion. These characteristics prohibit real-time data collection with sufficient temporal resolution, spatial resolution, and volumetric coverage. As a result, cardiac MR data acquisition has to be segmented and synchronized to both the rhythm of the heartbeat and to respiration, and the procedure is commonly performed over a number of consecutive cardiac cycles to avoid blurring in the images and to maximize the diagnostic yield.

Unmet needs

The relatively complex anatomy of the heart mandates meticulous plan scanning procedures and ECG lead placement. This costs time and requires skilled operators. The need for ECG gating or triggering together with strategies aimed at respiratory motion suppression makes cardiac MR data acquisition even more time consuming. For some procedures, where the anatomy of the heart is imaged with high spatial and temporal resolution, data collection efficiency (time during which data are sampled divided by time spent for the scan) can be as low as 2%, provided that most time is spent waiting for the next quiescent period of the heart to occur. However, this makes cardiac MR very time inefficient, prolongs scanning time, decreases patient comfort, and ultimately contributes to higher costs for cardiac MR. Then, and as MR provides the unique opportunity for quantification of tissue characteristics, function, and blood flow, standardization, accuracy, precision, repeatability and reproducibility among different centers, across different vendor platforms, and on different field strengths provide another formidable challenge. Therefore, improved ease-of-use with reduced operator dependency, improved time efficiency with shortened examination times, as well as established standardization procedures are among the unmet needs.

Where do we come from?

Over the past two or three decades, major steps in MR development that helped address some of the above needs have been undertaken by both academia and industry. With the development of performant gradient systems, the until then dormant steady state with free precession (SSFP) approach moved into focus again and permitted an unprecedented quality in that signal and contrast between the myocardium and the blood pool was significantly enhanced in an unprecedented manner. To this day SSFP or balanced SSFP (bSSFP) have become a work horse in cardiac MRI. With the development of coil arrays and multiple receiver channels, parallel imaging became possible two decades ago and helped abbreviate scan times significantly. Then, with the intravenous injection of contrast media together with sophisticated timing of k-space data acquisition, the blood vessels could be visualized with unprecedented clarity and in conjunction with inversion recovery, location, extent, and transmurality of scar tissue could be characterized. Simultaneously, and owing to more performant gradients, parallel imaging and contrast injection, first pass perfusion imaging at rest and during stress became a reality. Finally, quantitative methods that help better characterize ventricular function, blood flow, anatomy, and regional tissue properties have emerged and continue to be developed at a staggering pace.

Where do we go?

Clearly, and with the aim provide access to a larger pool of operators and patients that are increasingly located outside of academic centers, to improve time efficiency of the exam, and to extract actionable, relevant, accurate and precise numbers, further innovation in the domains of data acquisition, reconstruction, and post processing/analysis is mandatory. Non-ECG triggered, uninterrupted free-breathing and fully self-navigated 3D radial data acquisition may help remove a good number of the above-mentioned hurdles. First, plan scanning is reduced to the placement of a 3D volume in the region of the heart while ECG lead placement and individually dependent double oblique scan plane orientation as well as breath-holding are no longer needed. This supports a significantly improved ease-of-use and will likely improve reproducibility. Secondly, time efficiency will be improved by an order of magnitude and will support shortened scanning times and improved patient comfort. Thirdly, with the addition of compressed sensing reconstructions that are guided by self-gating signals extracted from the acquired k-space data or from a more recently introduced Pilot Tone concept, we will be empowered to flexibly interrogate such datasets retrospectively, freely, and in 3D. This will shift the paradigm from “plan and scan” to “scan now and ask questions later”. Consistent with this general idea, “Fingerprinting” and “Multi Tasking” provide original, novel, and exciting opportunities and hold the promise to extract quantitative information to additionally characterize regional tissue properties and blood flow with high accuracy, precision, reproducibility, and flexibility. Clearly, artificial intelligence, new devices, and hardware for interventions as well as continuous and stunning advances in computing power currently provide unprecedented opportunities to further push the envelope, to make new scientific discoveries, to add diagnostic and therapeutic value, and to generally expand the field of cardiac MR.

Acknowledgements

No acknowledgement found.

References

No reference found.