Technological Advances and New Physiological Insights for Reliably Probing Myocardial Oxygenation with Magnetic Resonance Imaging
- Author(s): Yang, Hsin-Jung
- Advisor(s): Dharmakumar, Rohan
- Ennis, Daniel
- et al.
Coronary artery disease (CAD) is the leading cause of death in the developed world. Evaluation of coronary function is critically important in the course of CAD management. Myocardial perfusion imaging (MPI) is the most popular noninvasive means of probing coronary function. While various imaging modalities have been used to image myocardial perfusion, the current clinical standards usually require the use of ionizing radiation or the exogenous contrast agent. The requirements increase the potential risks and render MPI infeeasible in a large segment of patients with CAD.
Blood oxygenation level dependent (BOLD) cardiac MRI (CMR) is a newer approach for imaging myocardial perfusion. BOLD CMR utilizes the intrinsic contrast of the deoxygenated blood to probe myocardial oxygenation without ionizing radiation and contrast agent. Developments in the past twenty years have made significant advance in BOLD CMR and recent studies have shown promising results for the clinical application of the approach. However, BOLD CMR still suffer from multiple challenges, such as inadequate accuracy, imaging confounders, limited imaging speed, and sub optimal stress imaging protocols, all which significantly limit its reliability and clinical applicability. In this dissertation, we propose to overcome these limitations from the technical and physiological vantage points to enable a reliable BOLD CMR exam, which can rapidly accelerate the clinical translation and adoption of the approach.
The first objective of this dissertation is to develop, test and validate a cardiac MRI technique that is robust, reliable and fast. To overcome the challenging cardiac imaging condition at 3T and provide reliable and sensitive BOLD signal, this dissertation utilized multiple techniques, such as motion correction, SR preparation, adiabatic pulses and novel reconstruction methods, to realize a confounder-corrected, free-breathing 3D T2 mapping technique at 3T that can be completed within the period of adenosine stress. The technique was tested and validated with simultaneously acquired 13N-ammonia PET perfusion in a whole-body PET/MR system. Ex-vivo studies demonstrated the proposed approach could be used to overcome the heart-rate dependent changes in T2 between rest and stress. In canines without coronary stenosis, T2 under adenosine stress was significantly greater than at rest, which was consistent with the observed increase in PET perfusion measurements. The changes in BOLD signal between rest and stress was highly correlated to myocardial perfusion reserve. In animals with coronary stenosis, perfusion anomalies were consistently detected between the developed technique and PET. These results demonstrate the proposed approach has the capability to enable rapid and reliable measurements of whole heart perfusion.
The second objective of the dissertation is to explore innovative vasodilation strategies to improve the reliability and applicability of BOLD CMR exams. Two novel ways of inducing vasodilation (Regadenoson and hypercapnia) for BOLD CMR were studied. Regadenoson is a new pharmacological vasodilator that has a prolonged vasodilation effect compared to the standard stress agent (adenosine). Utilizing the longer vasodilation period, we found that a significantly improved stress image quality and reliability can be achieved by imaging at a delayed time point (10 mins after the regadenoson administration). For validation purpose, BOLD and PET images were acquired simultaneously in healthy animals. The delayed acquisitions showed markedly increased myocardial T2 from the BOLD images and increased myocardial blood flow from 13N-ammonia PET perfusion. To further improve the sensitivity and reliability of regadenoson BOLD exam, a repeated BOLD image acquisition strategy was proposed to continuously monitor the dynamics of BOLD signal after regadenoson injection. The BOLD signal changes were used to investigate the coronary dynamics after the peak vasodilation in healthy and diseased subjects. The coronary dynamic parameters (CDPs) were derived and were used to compare to the conventional single time point approach. CDPs showed the capability of extracting greater BOLD response in the healthy subjects and improved performance in disease identification in the animals with impaired coronary arteries. This preliminary study showed the potential of improving BOLD CMR performance with existing stress agents and better-designed strategies to assess coronary vasodilatation.
Hypercapnia (elevated arterial CO2 (PaCO2)) is a known mediator of carotid vasodilation, but its effects on the coronary arteries had been unclear given the lack of tools to accurately and rapidly alter arterial CO2. By prospectively and independently controlling PaCO2, this dissertation first investigated whether a physiologically tolerable hypercapnic stimulus can increase MBF in healthy and diseased animals. The extent of effect on MBF due to hypercapnia was compared to adenosine using quantitative 13N-ammonia PET measurements. In the healthy animals, MBF under hypercapnia and adenosine were positively correlated; and were not different. Under LAD stenosis, MBF increased under hypercapnia and adenosine but the effect was significantly lower in the affected territories. Mean perfusion defect volumes measured with adenosine and hypercapnia were significantly correlated and were not different. The BOLD MR signal response to the controlled hypercapnia was also investigated under similar setup. In intact canines, changes in myocardial BOLD MR signal were equivalent to changes with adenosine. In addition, BOLD response under repeat hypercapnic stimulations were also tested. Reproducible BOLD responses were observed following repeat stimulation between baseline and equivalent hypercapnic states. In the dogs with coronary artery stenosis, BOLD MR signal changes during hypercapnia and adenosine infusion were also not different. The results demonstrate arterial blood CO2 tension is an independent variable of MBF, which can induce MBF and BOLD signal to the same level as standard dose of adenosine.
In summary, this dissertation culminates in key advancements which have the capacity to reliably advance cardiac BOLD in the clinical arena in the assessment of myocardial perfusion in the setting of ischemic heart disease. The technical and physiological advances direly address the key obstacles present with in the current BOLD techniques and build the foundation for a reliable, robust and practical myocardial BOLD exam.