Oxygenation-sensitive CMR for assessing vasodilator-induced changes of myocardial oxygenation

Journal of Cardiovascular Magnetic Resonance ROesxeayrcghenation-sensitive CMR for assessing vasodilator-induced changes of myocardial oxygenation Matthias Vhringer Jacqueline A Flewitt Jordin D Green Rohan Dharmakumar Jiun Wang Jr John V Tyberg Matthias G Friedrich 0 0 Stephenson Cardiovascular MR Centre at the Libin Cardiovascular Institute of Alberta, Department of Cardiac Sciences, University of Calgary , Calgary, AB , Canada Background: As myocardial oxygenation may serve as a marker for ischemia and microvascular dysfunction, it could be clinically useful to have a non-invasive measure of changes in myocardial oxygenation. However, the impact of induced blood flow changes on oxygenation is not well understood. We used oxygenation-sensitive CMR to assess the relations between myocardial oxygenation and coronary sinus blood oxygen saturation (SvO2) and coronary blood flow in a dog model in which hyperemia was induced by intracoronary administration of vasodilators. Results: During administration of acetylcholine and adenosine, CMR signal intensity correlated linearly with simultaneously measured SvO2 (r2 = 0.74, P < 0.001). Both SvO2 and CMR signal intensity were exponentially related to coronary blood flow, with SvO2 approaching 87%. Conclusions: Myocardial oxygenation as assessed with oxygenation-sensitive CMR imaging is linearly related to SvO2 and is exponentially related to vasodilator-induced increases of blood flow. Oxygenation-sensitive CMR may be useful to assess ischemia and microvascular function in patients. Its clinical utility should be evaluated. - Background Myocardial oxygenation, reflecting the balance or imbalance between oxygen demand and supply, is an important diagnostic target in various clinical settings[1], but may be especially useful for assessing ischemia and microvascular function. Presently available diagnostic tools are invasive, use exogenous contrast agents and/or radiation, are only useful in particular coronary territories, or have a limited spatial resolution[2,3]. Moreover, they do not provide direct measures of ischemia. T2*-sensitive, Blood-Oxygen-Level-Dependent Cardiovascular MR (BOLD-CMR) uses the paramagnetic properties of deoxygenated hemoglobin as an endogenous contrast mechanism and is thus oxygenation-dependent[4]. In oxygenation-sensitive CMR images, the signal intensity of any soft tissue is inversely correlated with its absolute content of deoxygenated hemoglobin and is therefore theoretically sensitive to changes in blood volume and oxygen supply-demand balance[5]. Such sequences have been used routinely for functional brain imaging[6] and similar approaches have been applied to the heart and peripheral perfusion beds [7-10]. Although these T2* measurement and T2*-mapping techniques have been shown to have high BOLD sensitivity, they had limited clinical use thus far because of long acquisition times and relatively low signal-to-noise ratios. Additionally, magnetic field inhomogeneities, blood flow and cardiac motion may all impair image quality. Recently, BOLD-sensitive, steady-state free precession (SSFP) techniques with much more consistent image quality have been introduced [11-13] and applied in experimental models of coronary artery stenosis[14,15]. To date, however, SSFP BOLD-CMR has not been validated against simultaneous measurements of myocardial oxygenation changes. Moreover, BOLD-weighted SSFP imaging has not been compared against other approaches such as T2* mapping. We hypothesized that SSFP BOLD-CMR can accurately and consistently detect changes of myocardial oxygenation in vivo. Methods We used a canine model with selective intracoronary vasodilator infusion. In order to cover the full range of physiological flow changes we applied graded infusions of the endothelium-dependent vasodilator, acetylcholine, as well as the endothelium-independent vasodilator, adenosine. Seven mongrel dogs (weight 15 to 25 kg) were studied; all experiments were conducted in accordance with the most recent policies and "Guide to the Care and Use of Experimental Animals" by the Canadian Council on Animal Care. The local animal care and use board approved the study protocol and sample size. Under general anesthesia, a midline sternotomy was performed and a 2-mm MR-compatible flow probe (Transonic Systems Inc., Ithaca, NY) placed around the proximal left circumflex (LCX) coronary artery. Under fluoroscopic control, a 2.7-F infusion catheter (Tracker18 Hi-Flow, Boston Scientific Ltd., Cork, Ireland) was introduced into the LCX through a diagnostic coronary catheter (JL 2.5, Torcon NB Advantage Catheter, Cook , Denmark). The tip of the infusion catheter was placed a few millimeters proximal to the flow probe while ensuring that there were no visible side branches located between the infusion catheter and flow probe. In addition, a 4-F balloon catheter (Berman Angiographic Balloon Catheter, Arrow, Reading, PA, USA) was introduced into the coronary sinus (CS) for blood sampling. Blood gases were analyzed using a portable analyzer (STAT PROFILE Critical Care Xpress, Nova Biomedical, Waltham, MA, USA). All procedures including the CMR scan were performed in adjacent rooms with the dogs being placed on an MR-compatible cradle that allowed for a quick and easy transport to and from the MR system. All CMR scans were performed in a clinical 1.5-T MRI system (MAGNETOM Avanto, Siemens Healthcare, Erlangen, Germany) with a 6-element phased-array coil resting on the chest and another below the spine. After acquiring localizer planes and performing manual regional shimming, BOLD-CMR was performed in a single mid-ventricular short-axis view at baseline (BL 1-3) and during intracoronary vasodilator infusion into the LCX. Acetylcholine (ACh) was infused in three increasing doses as previously described [16]: 0.1 g/min (ACh 1), 1 g/min (ACh 2) and 10 g/min (ACh 3). Adenosine (Ade) was infused at the following rates: 30 g/min (Ade 1), 150 g/min (Ade 2) and 300 g/min (Ade 3). Measurements were performed in the following sequence: BL 1, ACh 1-3, BL 2, Ade 1-3, and BL 3. At the end of the protocol, we acquired a series of images during first-pass perfusion using a single-shot GRE-EPI sequence after intracoronary injection of 0.05 mmol/kg gadopentetate dimeglumine (Magnevist, Bayer, Germany) for accurately identifying the LCX perfusion territory and confirming the correct position of the intracoronary catheter. SSFP BOLD-CMR was performed with a T2*-sensitive cine SSFP sequence as previously described [15]. Scan parameters were: FOV = 228 280 mm; matrix size = 125 192; in-plane resolution = 1.8 1.6 mm; slice thickness = 5 mm; TR/TE = 5.8 ms/2.9 ms; flip angle = 90; readout bandwidth = 275 Hz/Px; signal averages = 1; the duration of the typical breath-hold was 15 s. In addition, a segmented multi-echo gradient echo (GRE) sequence was used (echo train length: 8; TE = 2.6, 4.8, 7.0, 9.3, 11.5, 13.7, 16.0, and 18.2 ms) using a mono-pola (...truncated)