Oxygenation-sensitive cardiovascular magnetic resonance

Journal of Cardiovascular Magnetic Resonance Oxygenation-sensitive cardiovascular magnetic resonance Matthias G Friedrich 0 1 3 Theodoros D Karamitsos 2 0 Departments of Cardiac Sciences and Radiology, University of Calgary , Calgary , Canada 1 Montreal Heart Institute, Departments of Cardiology and Radiology, Universite de Montreal , Montreal, QC , Canada 2 Oxford Centre for Clinical Magnetic Resonance Research, Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, John Radcliffe Hospital , Oxford , UK 3 Montreal Heart Institute, Departments of Cardiology and Radiology, Universite de Montreal , Montreal, QC , Canada Oxygenation-sensitive cardiovascular magnetic resonance (CMR) is a non-contrast technique that allows the non-invasive assessment of myocardial oxygenation. It capitalizes on the fact that deoxygenated hemoglobin in blood can act as an intrinsic contrast agent, changing proton signals in a fashion that can be imaged to reflect the level of blood oxygenation. Increases in O2 saturation increase the BOLD imaging signal (T2 or T2*), whereas decreases diminish it. This review presents the basic concepts and limitations of the BOLD technique, and summarizes the preclinical and clinical studies in the assessment of myocardial oxygenation with a focus on recent advances. Finally, it provides future directions and a brief look at emerging techniques of this evolving CMR field. Cardiovascular magnetic resonance; Blood-oxygen level-dependent; Microcirculation; Ischemia; Oxygenation - Review Contrast Generation in Oxygenation-Sensitive CMR It was Linus Pauling who first described an effect of the oxygenation state of hemoglobin on its magnetic properties [1]. With respect to its behavior during magnetic resonance experiments, the de-oxygenation of hemoglobin causes the molecule to act as an intrinsic paramagnetic contrast agent resulting in pronounced spin-spin interaction. This accelerates the decay of transverse magnetization and thus shortens the spin-spin relaxation time. The term T2* describes the rate of loss of the transverse signal when using sequences without refocusing radiofrequency pulses. Regional iron deposition, hemoglobin degradation products in tissue hemorrhage or deoxygenated hemoglobin accelerate this process by their paramagnetic properties, a phenomenon called the BOLD (Blood-Oxygen-LevelDependent) effect. This phenomenon can be exploited to perform BOLD-sensitive or, depending on the context, oxygenation-sensitive imaging with the aim to detect changes in tissue oxygenation. In 1990, Ogawa et al. demonstrated that oxygenationsensitive MR can be used to detect consequences of very small blood flow changes in the brain resulting from external stimuli [2]. In humans, typically several averages and color-coded maps are used to visualize these changes [3], a technique which is widely known as functional brain MR. Accordingly, myocardial deoxygenation or ischemia is characterized by a net relative increase of de-oxygenated hemoglobin in the capillary blood and thus leads to T2* shortening, which can be visualized by T2* maps or by a regional signal loss in T2*-weighted MR images (i.e. images acquired by protocols sensitive to decreased regional field homogeneity). Conversely, a decrease of the proportion of de-oxygenated hemoglobin (for example by inducing vasodilation without a matching increase of myocardial oxygen demand) causes a relative decrease of de-oxygenated hemoglobin and leads to an increase of T2* in this territory and hence to an increased signal intensity in oxygenation-sensitive images. It is important to keep in mind that the observed changes reflect changes in (mainly the venous compartment of the) capillary bed and therefore strictly do not represent the actual cell (i.e. cardiomyocytes). Yet the oxygenation of the capillary blood directly reflects the balance of oxygen supply and demand and therefore can be understood as a direct marker of tissue oxygenation. Technical aspects of oxygenation-sensitive CMR While image quality is frequently affected in echo-planar sequences [4] and other, T2-weighted sequences [5] by motion or susceptibility artifacts (especially along the lung-heart interface), steady-state-free-precession sequences have shown promising results, allowing for a simultaneous acquisition of morphological, functional and oxygenationsensitive data of the heart [6]. Of note, the oxygenationsensitive contrast in images acquired by this sequence is directly dependent on the repetition time TR [7]. With an appropriate TR, the obtained image quality is much more consistent when compared to other BOLD-sensitive sequences, though at the expense of reduced sensitivity to the BOLD effect and thus lower oxygenation contrast. Further studies may have to fine-tune the balance between sensitivity to oxygenation changes and susceptibility artifacts. Importantly, the magnitude of the BOLD effect (i.e. the measurable effect size) largely depends on the strength of the static magnetic field. While the overall susceptibility to field inhomogeneity at higher field strengths may cause artifacts, it also accounts for a higher sensitivity to changes induced by paramagnetic effects. Higher field strengths thus improve the sensitivity of MR to the BOLD effect. Data by Dharmakumar et al. indicate that the sensitivity to detect changes in myocardial oxygenation may increase by a factor of about 2.5 when moving from 1.5T to 3T [8]. With respect to a clinical application in patients with suspected myocardial ischemia, it is very important to be aware of the significant limitations of current diagnostic techniques to verify the hemodynamic relevance of coronary artery disease. Because of a lack of diagnostic targets on the (cellular) level of actual ischemia, imaging techniques use surrogate markers such as stress-inducible dysfunction (echocardiography), changes of blood inflow characteristics (echocardiography, first-pass perfusion CMR) or metabolic changes (nuclear cardiology techniques). All these are surrogates and thus cannot directly reflect an ischemic response of myocardial tissue, while oxygenation-sensitive CMR offers exactly that. Experimental and pre-clinical studies In 1993, Atalay et al. used a Langendorff heart model to show that the signal intensity in images sensitive to the BOLD effect (in that paper called susceptibility-dependent) is closely correlated to blood oxygenation [9]. Wendland et al. induced hypoxia in rats and demonstrated changes in the blood and the myocardium caused by the BOLD effect [10]. In July 1996, Niemi et al. observed in healthy volunteers, that the increase of myocardial blood flow as induced by the infusion of the vasodilator dipyridamole was closely correlated with a myocardial signal intensity increase in images obtained by a T2*-sensitive single-shot echo planar imaging (EPI) sequence [11]. Yet, another study, published in the same journal issue by (...truncated)