Self-Navigated Cine DENSE MRI for Free-Breathing Myocardial Strain Imaging

Myocardial strain imaging by echocardiography and magnetic resonance imaging (MRI) is increasingly used to assess cardiac function. Cine displacement-encoded stimulated-echoes (DENSE) MRI is a strain imaging technique that is accurate, reproducible and amenable to rapid displacement and strain analysis. With these properties, the clinical applications of cine DENSE are expanding. Recent studies demonstrated the potential of cine DENSE to improve implementation of cardiac resynchronization therapy in heart failure patients and detect subclinical ventricular dysfunction in childhood obesity. Studies in patients with acute myocardial infarction proved the prognostic value of strain by cine DENSE.
Conventionally, cine DENSE requires breath-holds for data acquisition, which is challenging in many patient populations, such as heart failure, pediatrics, and cardiomyopathy with dyspnea. A reliable free-breathing method is important for reproducible and efficient imaging in these patients. Although a conventional diaphragm-based navigator method (dNAV) was previously applied to enable free-breathing cine DENSE imaging, there are numerous disadvantages including the requirement of extra scout scans, variable imaging quality, sensitivity to respiratory pattern change, and reduced reproducibility. The field of cardiac MRI has shifted to self-navigated free-breathing methods where respiratory motion is typically extracted from the imaging data itself for motion compensation. The overall goal of this dissertation was to investigate the artifact sources in free-breathing cine DENSE and to develop and evaluate a self-navigated method that addressed these artifacts correspondingly.
A match-making reconstruction framework was developed to effectively compensate for two major sources of artifacts, namely the striping artifacts due to residual T1-relaxation echo and blurring due to inter-heartbeat motion. The framework was validated through experimental data in phantom and healthy subjects. The phantom experiments demonstrated the concepts of the framework where minimal residual energy of the T1-relaxation echo (rT1E) indicated little motion between the phase-cycling data and subsequent motion correction with image-based navigators reduced blurring artifacts. A preliminary evaluation in healthy subjects showed that the framework can reduce free-breathing artifacts better than the conventional diaphragm navigator method.
An acquisition algorithm was designed to diminish rT1E adaptively in real-time. The algorithm calculates the rT1E values of new phase-cycling pairs and always repeats the k-space data portion with the highest rT1E during the acquisition until the rT1E is low enough or maximal imaging time is reached. Experiments in healthy subjects were performed to determine the stopping criteria. The results demonstrated that the rT1E decreased efficiently.
The presented self-navigated free-breathing method was evaluated in healthy volunteers and patients with heart disease. Free-breathing datasets were acquired with the adaptive algorithm and the image reconstruction was performed with compensation for both inter- and intra-heartbeat motion using stimulated-echo image-based navigators (ste-iNAV). The methods were found to achieve better image quality and more reproducible strain imaging than the dNAV method.
The self-navigated acquisition and reconstruction methods address the artifacts sources in free-breathing cine DENSE effectively and achieve reproducible segmental myocardial strain. Unlike dNAV, the method does not require extra set-up scans. These methods hold promise for reliable free-breathing strain imaging in patients with heart disease.