Mouse models of cardiac disease are an important tool to gain understanding of the pathophysiological processes related to the heart, as well as for the development of new treatment strategies. In this respect, Magnetic Resonance Imaging (MRI) has become the gold standard imaging modality, because it combines high spatial resolution imaging with a large variety of soft tissue contrast weightings that can be related to the presence of diseased tissue. In addition, (targeted) MRI contrast agents can be employed to visualize different processes on the molecular level, for example in relation to myocardial infarction and the subsequent cardiac remodeling process. The specificity to discriminate healthy from diseased tissue as well as the sensitivity for detection of MR contrast agents is strongly affected by the specific MRI protocol design. Moreover, the challenging physiology of the mouse heart, especially with respect to its small size and high heart rate, often limits the direct translation of imaging protocols already available from clinical studies. Finally, the growing knowledge on cardiac pathology continuously pushes the development of sophisticated mouse cardiac MRI protocols that allow more detailed measurements of a variety of physiologically relevant cardiac parameters. The overall goal of this thesis was therefore to design and investigate novel imaging strategies in the field of mouse cardiac MRI and their application in models of cardiac disease. Chapter 2 of this thesis contains an extensive overview of currently available protocols for mouse cardiac MRI and more specifically those related to contrast enhanced imaging of myocardial infarction. The remainder of the thesis contains the experimental chapters describing all details on our newly developed mouse cardiac MRI techniques. This chapter shortly summarizes the aims and results with respect to each of these techniques, categorized based on the parameter of interest for which each measurement was specifically designed. Diastolic function Measurement of murine diastolic function requires Cine imaging with an extremely high frame rate ¿ more than 60 frames within a cardiac cycle of 100-120 ms ¿ to be able to discriminate between the two separate filling phases of the heart. In chapter 3, it was shown that using a retrospectively triggered MRI sequence, reconstruction of 80 Cine images was feasible, corresponding to a temporal resolution of around 1.5 ms. This was achieved without using any form of data interpolation. With retrospective triggering, the MRI measurements are not synchronized with the ECG, thereby in theory sampling an infinite number of time points during the cardiac cycle. Correct assignment of each k-line to a specific cardiac frame could be done retrospectively by measuring an additional navigator signal prior to image acquisition, whose signal amplitude varies with cardiac as well as respiratory movements. Because in this case, filling of k-space for each cardiac frame is a stochastic process, simulations were performed to investigate the efficiency of the method with respect to signal averaging, which was found to be almost equal compared to regular prospective triggering. Diabetic cardiomyopathy has a high prevalence in type 2 diabetes patients and is characterized by diastolic dysfunction. With the current technique, we were indeed able to measure a subtle reduction is several diastolic function parameters, which are the E/A-ratio and the E-contribution to total left ventricular filling. Therefore, this technique is a promising tool in experimental studies of diabetic cardiomyopathy and for evaluation of emerging treatment strategies for diastolic dysfunction. Myocardial perfusion Chapter 4 describes the application of first-pass perfusion measurements in a mouse model of myocardial infarction to allow the assessment of the myocardial perfusion status. A first-pass perfusion measurement is performed by venous injection of an MRI contrast agent and monitoring its passage through the left ventricle and myocardial wall. From the signal intensity changes upon passage of the contrast agent, myocardial perfusion values can be determined. The application of this technique in mice requires ultra-fast MRI sequences that can sample the signal intensity-time curves with sufficient temporal resolution. Because this concerns imaging of non-periodic signal changes this is a much different problem compared to the diastolic function experiments described in chapter 3. We showed that using a saturation recovery MRI sequence with segmented k-space read-out in combination with parallel imaging acceleration techniques, a time-series of images could be acquired with a temporal resolution of 1 image for each 3 heart beats. The use of parallel imaging was crucial, since this requires less k-lines for image reconstruction compared to conventional imaging. Furthermore, the use of saturation pulses enhanced the contrast between contrast-enhanced and non-enhanced blood and myocardium. Using this technique, semi-quantitative perfusion values could be determined based on the upslope of the signal intensity-time curves. Experiments in mice with permanent occlusion of the LAD showed a significant decrease of perfusion values in the infarcted myocardium as compared to remote myocardium. In future experiments, this technique will be extended to provide quantitative perfusion values (in mg/l/min), requiring determination of the true arterial input function from a pre-bolus measurement with a smaller contrast agent bolus volume. T1 and T2 relaxation times Pathology is often accompanied by a change in the magnetic properties of the tissue, in particular the T1 and T2 relaxation times. This directly affects the signal intensity on the MR image. Diseased and healthy tissue can therefore be discriminated on MR images, which is one of the main applications of MRI in clinical diagnostics. However, there is much interest in quantitative assesment of T1 and T2 relaxation times, as this improves repeatibility of results in longitudinal studies and reproducibility between research groups. In this thesis, we aimed at developing protocols for both T1 and T2 mapping of the complete mouse heart for application in mouse models of myocardial infarction. Whole-heart coverage is important considering that a priori, the extent of the infarct is unknown. Currently available protocols for T1 mapping are relatvively time-consuming. In chapter 5, a 3D T1 mapping sequence is presented which allows myocardial T1 quantification of the mouse heart within 20 minutes. The retrospective triggering sequence used in chapter 3 proved also useful in this study, because it allows steady-state acqusition with very short repetition times, enabling whole heart coverage. T1 values were derived from measuring a variable flip angle data set and using available MRI signal models. Variable flip angle data showed excellent agreement in cardiac anatomy, allowing pixel-wise determination of T1. In healthy mice, no substantial differences in T1 were found for different heart regions in the 3D volume. Coefficents of repeatibility were determined from measurements at different days, which varied as function of the number of flip angles used in data analysis. In contrast to T1, T2 values could not be acquired using 3D acquisitons or retrospective triggering. Alternatively, chapter 6 describes a multi-slice T2 mapping protocol for the mouse heart based on a ECG-triggered T2 magnetization preparation module with variable TE. Because the preparation module consisted of many consecutive RF pulses, the effect of these pulses on T2 relxation had to be taken into account. Additionally, simulations were used to calculate the effect of T1 relaxation on T2 estimation, which was small as long as the repetition time was kept sufficiently long. Homogeneous T2 maps of healthy mouse heart were obtained, with no substantial differences between different heart regions or slices. In a ischemia/reperfusion model, elevated T2 values were found in the infarcted area, probably as result of edema formation. The extent of the infarction was also measured using late gadolinium enhanced imaging. The degree of correlation of T2 and LGE enhanced regions strongly depended on the signal intensity thresholds derived from remote tissue. Contrast agent accumulation Another application of quantitative T1 and T2¬ mapping is the assessment of the concentration of a contrast agent, which has been targeted to a specific disease site. This is especially valuable in molecular imaging applications where contrast agents report on the presence of specific disease markers related to various cardiac remodeling processes after myocardial infarction. Chapter 7 describes the application of the T1 mapping protocol from chapter 5 to quantify the accumulation of a Gd-based liposomal contrast agent in a model of myocardial infarction. Functional imaging and assessment of wall thickening values were used to determine which regions could be identified as being infarcted. Statistical analysis showed that before contrast agent administration, T¬1¬ values were already elevated in the infarcted regions as compared to remote myocardium, however, a more pronounced change in T1 values was found 24h post-contrast, with significantly lower T1 values in the infarcted areas. Pre-contrast T1 values in control mice were very similar to the study described in chapter 5, proving good reproducibility of T1 quantification using our methods. After the MRI measurement, the hearts were cut into slices, from which the Gd-content was determined in different sections of the heart using inductively coupled plasma mass spectrometry. T1 changes measured using in vivo MRI correlated well with ex vivo measurements of Gd concentration. These are promising results for quantification of contrast agent concentrations in contrast-enhanced MRI of mouse models of cardiac disease. More research has to be performed with regard to changes in contrast agent efficiency as a result of different cellular environments. Our results already indicate that the relaxivity values of liposomal contrast agents are significantly lower in vivo as compared to values obtained from measurements in phantom solutions. Conclusion This thesis has shown that mouse cardiac MRI is capable of assessing a large variety of parameters related to cardiac physiology in the in vivo mouse heart in a non-invasive way. This makes this technique an attractive platform for experimental studies on cardiac disease, as well as developing new treatment strategies.
|Qualification||Doctor of Philosophy|
|Award date||21 Mar 2012|
|Place of Publication||Eindhoven|
|Publication status||Published - 2012|