3D perfusion mapping in post-infarct mice using myocardial contrast echocardiography

Abstract

Myocardial contrast echocardiography (MCE) was used to construct three-dimensional maps of perfusion defects in closed-chest mice, with and without myocardial infarction (MI) induced by permanent coronary ligation. Contiguous, short-axis MCE cine images spanning the heart from apex to base were acquired at 1 mm elevations in each mouse. MCE images at each elevation were color-coded to indicate relative perfusion and were compared with postmortem histology. A strong correlation (R > 0.93) in the size of perfused areas was observed between in vivo measurements and the results of conventional ex vivo tissue staining. 3D multislice and 3D surface renderings of perfusion distribution were created and these perfusion maps also matched well with postmortem histology. These methods provide for the noninvasive determination of the total ischemic region placed at risk by coronary occlusion: this is a critical variable in assessing the potential of novel therapeutic agents to reduce MI size in murine models of ischemia/reperfusion injury.

Introduction

The mouse is a preferred species for studying the pathophysiology of cardiovascular diseases including myocardial infarction (MI), not only because the mouse is a widely available research subject, but also because it can be easily modified by genetic and pharmacologic manipulation (Guo et al 1998, Yang et al 2000, Condorelli et al 2002). Indeed, the noninvasive, in vivo study of mouse cardiac function has gained great interest in the cardiovascular research community. However, the small size (< 8 mm × 6 mm) and rapid rate (400 to 600 bpm) of the mouse heart demand an imaging modality with both high temporal resolution and high spatial resolution.

Imaging modalities currently being utilized to study myocardial perfusion include traditional nuclear-based modalities [single photon emission computed tomography (SPECT) and positron emission tomography (PET)], cardiac magnetic resonance imaging (MRI) and contrast-enhanced ultrasound imaging, also referred to as myocardial contrast echocardiography or MCE (Kaul 2001). SPECT and PET are generally limited by the spatial resolution required for assessing regional myocardial function and perfusion in the mouse. However, MRI offers excellent spatial resolution and contrast and serial studies of mouse cardiac physiology after MI using MRI (Ross et al. 2002) have proven to be quite feasible. One of the fundamental limitations of MRI with respect to myocardial perfusion is its slow acquisition frame rate relative to the rapid heart rate of mice, which makes it essentially impossible to image the first pass of contrast agents through the mouse myocardium at conventional field strengths. Additionally, there is a high capital equipment cost (US $1M to US $2M for premium systems), high maintenance cost (contributing to a cost per cardiac scan of approximately US $100) and low throughput (each comprehensive cardiac exam takes 45 to 60 min). All these issues limit the general utility of MRI for assessing mouse myocardial function and perfusion on a routine basis. Ultrasound is a favorable alternative because it has a higher acquisition frame rate and is a more affordable noninvasive imaging modality. For small animal models of myocardial ischemia, a major concern with ultrasound imaging is the low spatial resolution of many conventional systems. To overcome this drawback, ultrasound scanners with high frequency mechanically swept transducers are now available that provide high spatial resolution in small animals, with lateral resolution of approximately 60 to 100 μm (depending on the selected center frequency) and axial resolution about 40 to 60 μm (Foster et al 2002, Martin-McNulty et al 2005). The acquisition frame rate of this type of system can exceed 100 frames per second and is therefore marginally acceptable for real-time studies of mouse hearts (which beat at approximately 10 cycles per second). Unfortunately, while this type of system is capable of remarkable spatial resolution in B-Mode, it is not currently capable of high sensitivity and specificity contrast imaging. This is largely because most high sensitivity and high specificity contrast imaging modes require multiple pulses along identical beam line directions and this requirement is practically impossible to satisfy when using a mechanically swept transducer. For this reason, we chose to use a commercial, human phased array scanner with excellent contrast sensitivity and specificity (Sequoia 512, Siemens Medical Solutions, Mountain View, CA, USA). Unfortunately, the penalty for using such a system is inferior spatial resolution (approximately 200 μm lateral and 100 μm axial resolution). Nevertheless, numerous laboratories have already successfully demonstrated the use of human scanners in the study of mouse cardiac physiology (Condorelli et al 2002, Patten et al 2002, Scherrer-Crosbie et al 1999).

MCE has been developed quantitatively to measure myocardial blood flow (Wei et al 1998, Scherrer-Crosbie et al 1999, Kaul 2001). MCE involves infusion of a microbubble suspension which acts as a contrast agent to opacify the blood pool and enhance the left ventricle (LV) in the echocardiogram. Infusion of microbubbles as ultrasound contrast agents is physically harmless to animals and human beings (Lang et al. 1987). Physiological consequences resulting from MCE are very rare. It has further been demonstrated that MCE provides an adequate means for quantifying myocardial perfusion during venous infusion of microbubbles (Wei et al. 1998). Scherrer-Crosbie et al. (1999) have shown that MCE can accurately detect perfusion defects during two-dimensional (2D) short-axis scans of ischemic mouse LV at the midventricular level.

Although 2D analysis is still common for investigating myocardial structure and function by echocardiography, a comprehensive 3D analysis of the heart by MCE is needed to determine the total ischemic region-at-risk for infarction. Noninvasive measures of total region-at-risk become particularly important in studies of experimental myocardial infarction, where the efficacy of therapeutic intervention is best demonstrated by reductions in infarct size. Infarct size can be as measured as “% LV Mass” or as “% Region-at-Risk”, with the latter being preferred since the fraction of the LV mass rendered ischemic by coronary occlusion can vary significantly from one animal to the next. 3D MCE has already been shown accurately to determine region-at-risk and myocardial infarct volumes in an open-chest canine model (Linka et al. 1997). However, few quantitative 3D MCE studies have been reported involving closed-chest models of MI in mice. The difficulties mainly lie in three aspects. First, a small animal model requires much higher temporal and spatial resolution than a large animal model. Second, performing MCE in closed-chest animal models of MI is inherently more technically demanding than in open-chest models, both in terms of surgical procedures and image acquisition. Third, the temporal and spatial registration required for accurate 3D reconstruction of a fast-beating mouse heart is more challenging than for slower-beating or larger hearts.

The aim of the current study was to determine whether MCE provides an adequate tool quantitatively to assess 3D myocardial ischemia in an intact mouse model. Toward this end, we used MCE to image in vivo myocardial perfusion in both ischemic and nonischemic mouse hearts. Contiguous short-axis MCE images spanning the entire heart from apex to base were then assembled into multislice 3D stacks and surface-rendered perfusion maps to provide comprehensive volumetric views of the distribution of perfusion. For both 2D cross-sections and 3D volumetric renderings, comparisons were made between in vivo ultrasound imaging assessment and postmortem histology.