VALIDATING NOVEL FREE-BREATHING CARDIOVASCULAR MAGNETIC RESONANCE SEQUENCES FOR FUTURE APPLICATIONS OF PERI-OPERATIVE IMAGING OF INDUCIBLE MYOCARDIAL DEOXYGENATION

Abstract

<h3>Background</h3> Perioperative myocardial ischemia is associated with increased morbidity and mortality. General anesthesia (GA) bears many possible triggers for myocardial ischemia including acute blood gas changes. While cardiovascular magnetic resonance (CMR) is not feasible in an everyday intra-operative setting, its ability to measure myocardial oxygenation using oxygen-sensitive (OS) imaging provides a novel possibility to quantify tissue ischemia during the induction of GA in a research setting. However, OS-CMR is acquired during a breath-hold to reduce chest motion and requires up to 10s per image, and hence is not ideal for continuous monitoring. Thus, we optimized two free breathing OS-CMR sequence variants that acquire images every heartbeat to measure rapid changes in myocardial oxygenation. We aimed to compare these new sequences in an awake healthy control population performing voluntary hyperventilation and apnea. The intent of this analysis is to develop and validate this new CMR sequence for future studies implementing OS-CMR to measure myocardial oxygenation during GA. <h3>Methods</h3> Twenty awake healthy participants underwent a CMR study and performed two breathing protocols. First a maneuver representing anesthesia pre-induction protocols was performed with paced breathing at 15bpm for 150s followed by 5 deep breaths and then apnea. Controls were then asked to perform a rapid paced maneuver consisting of 60s of deep hyperventilation at 30bpm for 60s, followed by a second apnea phase. During the procedure images were acquired with the original breath-hold (BH) OS-CMR sequence, the new free-breathing single shot (SS-FB) and real-time (RT-FB) sequences. The first is a single-shot variant that only acquires one image in diastole at a high resolution of 1.5mm2. The second is a real-time variant, that acquires 20 frames per second thus providing functional information throughout the cardiac cycle as well, but at a spatial resolution of 3mm2. Myocardial oxygenation was statistically compared at baseline, post hyperventilation which is a vasoconstricting stimulus, and at the 30s timepoint in apnea that is known to induce vasodilation. <h3>Results</h3> For the first breathing pacing protocol representing GA induction, hyperventilation significantly reduced myocardial oxygenation with all three sequences (SS-FB: -12.1±10.6%, p<0.001, RT-FB: -8.0 ±4.0% p<0.001, BH: -4.2±3.7% p<0.001) and apnea resulted in a rise in myocardial oxygenation (SS-FB: 6.2±7.3% p=0.002, RT-FB: 11.8±4.1% p<0.001, BH: 6.9±5.7% p<0.001). With the rapid pacing maneuver similar results were observed with hyperventilation (SS-FB: -4.4±8.2% p=0.047, RT-FB: -7.0±1.8% p<0.001, BH: -4.3±6.2% p=0.008) and apnea (SS-FB: 9.2±6.5% p=<0.001, RT-FB: 9.1±2.1% p<0.001, BH: 4.5±5.0% p=0.001). The new SS-FB and RT-FB sequences were not inferior to the known gold standard BH sequence for any stage. <h3>Conclusion</h3> In awake controls, alterations in myocardial oxygenation through hyperventilation and apnea can be observed with the novel free-breathing OS-CMR sequences. It is now warranted to implement these new techniques in future studies investigating peri-operative imaging of myocardial ischemia.