*3.4.2. Interstrain comparisons of global cardiac function*

Apart from early attempts [Wiesmann 2000], little work has been published on interstrain comparisons of global cardiac functional differences in various mouse strains [Constantinides\_SBI 2010, Bucholz 2010]. Instead, studies of cardiac dysfunction have often included quantitative comparisons with normal control or sham mice [Zhang 2008]. Interstrain variability on cardiac function exists in different mouse strains, which may be dependent or independent of strain, but there are certainly developmental factors and other factors that ought to be investigated. Reported results must thus be carefully considered for age, weight, sex, and genetic background. In recent findings from our group (Figure 15) no significant variations in cardiac function were observed from age-, sex-matched (male), normal C57BL/6J and DBA/2J mice. Image-derived hemodynamic indices of function reported in Figure 16 exemplify similar cardiac functional responses from both the right and left ventricular cavities, in agreement with prior reports [Zhang 2008].

Study of the Murine Cardiac Mechanical Function Using Magnetic Resonance Imaging:

Finite differences may exist but careful consideration of other factors (such as age, sex, or

**Figure 16.** Interstrain hemodynamic (EDV, ESV, SV, EF) index comparisons for left (LV) and right (RV)

Apart from tissue displacement encoding techniques other motion tracking [Osman 1999],

Despite the importance of all such techniques they nevertheless necessitate access to dedicated high-field MRI scanners and invariably require development, or use, of complex algorithms and reconstruction software. An easier methodology to assess regional cardiac function for comparison with humans was recently reported by Constantinides et al. [Constantinides\_ISMRM 2011] with the use of dedicated software (Figure 17) based on

*Wall Motion EDepicardial wall diameter epicardial wall diameter* \_ \_ \_ \_ *ES* (1)

 \_ \_ \_ 100. *wall thickness wall thickness*

*ES ED*

 2 2 \_ \_

100. *endocardial diameter endocardial diameter*

*ED ES*

*ED*

\_

*endocardial diameter*

*wall thickness*

2

(3)

(2)

velocity [Streif 2003], and acceleration techniques [Staehle 2011] have been reported.

ventricular performance in male C57BL/6J (n=5) and DBA/2J mice (n=5) based on MRI.

segmented epicardial and endocardial boundaries, according to:

*rEF*

*WallThickening ED*

anesthesia effects) must also be considered in such analyses.

The Current Status, Challenges, and Future Perspectives 369

**Figure 15.** Typical (top two rows) LV surface and finite element, and (bottom two rows) corresponding RV representation of the 4D dynamic surface and finite element model of DBA/2J mice through eight cardiac phases over the entire cardiac cycle.

#### *3.4.2.1. Quantification of regional cardiac function - Comparison of mouse and human*

Despite the usefulness of global cardiac index comparisons, the value and importance of regional cardiac functional analyses in disease is paramount. Further to the use of tagging [Liu 2006] and DENSE [Gilson 2005, Zhong 2010] as non-invasive techniques for assessment of regional cardiac displacement and strain in mice, similar patterns of motional responses were observed in mice and humans with noted finite, but distinct, differences. Particularly, Gilson [Gilson 2005] reports basal displacement but almost no apical displacements in the mouse during the cardiac cycle, in comparison to both apical and basal motion in humans [Moore 2000]. Furthermore, circumferential and radial displacements seem to scale proportionally with values previously reported in humans [Constantinides\_Phantom 2012]. Finite differences may exist but careful consideration of other factors (such as age, sex, or anesthesia effects) must also be considered in such analyses.

368 Practical Applications in Biomedical Engineering

cardiac phases over the entire cardiac cycle.

*3.4.2. Interstrain comparisons of global cardiac function* 

left ventricular cavities, in agreement with prior reports [Zhang 2008].

Apart from early attempts [Wiesmann 2000], little work has been published on interstrain comparisons of global cardiac functional differences in various mouse strains [Constantinides\_SBI 2010, Bucholz 2010]. Instead, studies of cardiac dysfunction have often included quantitative comparisons with normal control or sham mice [Zhang 2008]. Interstrain variability on cardiac function exists in different mouse strains, which may be dependent or independent of strain, but there are certainly developmental factors and other factors that ought to be investigated. Reported results must thus be carefully considered for age, weight, sex, and genetic background. In recent findings from our group (Figure 15) no significant variations in cardiac function were observed from age-, sex-matched (male), normal C57BL/6J and DBA/2J mice. Image-derived hemodynamic indices of function reported in Figure 16 exemplify similar cardiac functional responses from both the right and

**Figure 15.** Typical (top two rows) LV surface and finite element, and (bottom two rows) corresponding RV representation of the 4D dynamic surface and finite element model of DBA/2J mice through eight

Despite the usefulness of global cardiac index comparisons, the value and importance of regional cardiac functional analyses in disease is paramount. Further to the use of tagging [Liu 2006] and DENSE [Gilson 2005, Zhong 2010] as non-invasive techniques for assessment of regional cardiac displacement and strain in mice, similar patterns of motional responses were observed in mice and humans with noted finite, but distinct, differences. Particularly, Gilson [Gilson 2005] reports basal displacement but almost no apical displacements in the mouse during the cardiac cycle, in comparison to both apical and basal motion in humans [Moore 2000]. Furthermore, circumferential and radial displacements seem to scale proportionally with values previously reported in humans [Constantinides\_Phantom 2012].

*3.4.2.1. Quantification of regional cardiac function - Comparison of mouse and human* 

**Figure 16.** Interstrain hemodynamic (EDV, ESV, SV, EF) index comparisons for left (LV) and right (RV) ventricular performance in male C57BL/6J (n=5) and DBA/2J mice (n=5) based on MRI.

Apart from tissue displacement encoding techniques other motion tracking [Osman 1999], velocity [Streif 2003], and acceleration techniques [Staehle 2011] have been reported.

Despite the importance of all such techniques they nevertheless necessitate access to dedicated high-field MRI scanners and invariably require development, or use, of complex algorithms and reconstruction software. An easier methodology to assess regional cardiac function for comparison with humans was recently reported by Constantinides et al. [Constantinides\_ISMRM 2011] with the use of dedicated software (Figure 17) based on segmented epicardial and endocardial boundaries, according to:

$$\text{Vall Motion} = \text{ED}\_{\text{equivalent\\_wall\\_diameter}} - \text{ES}\_{\text{equivalent\\_wall\\_diameter}} \tag{1}$$

$$\text{MallThickening} = 100. \frac{\left(E S\_{\text{wall\\_thickness}} - E D\_{\text{wall\\_thickness}}\right)}{E D\_{\text{wall\\_thickness}}} \tag{2}$$

$$rEF = 100. \frac{\left(\left(ED\_{end\text{oxidized\\_diameter}}^2 - ES\_{end\text{credit\\_diameter}}\right)^2\right)}{ED\_{end\text{credit\\_diameter}}^2} \tag{3}$$

Based on such analyses, bullseye-plots of regional cardiac function were generated in 17 sector representations of the murine and human hearts (from independent studies in two separate Institutions) showing similar patterns in transmural variations in wall motion and thickness, and regional ejection fraction (Figure 18) in mouse and man. Such spatial patterns observed for mouse and human (in agreement with prior tagging work [Moore 2000]) are supported by two-tailed paired t-tests indicating absence of statistical significant differences in the mean values of wall thickness (p=0.07), wall motion (p=0.051), or regional EF (p=0.065) at the 1% significance. Repeated measures ANOVA indicated significant differences in regional mouse and human for wall thickness (p=0.002) and regional EF (p<0.0001) and insignificant differences for wall motion (p=0.016) at the 1% significance. However, despite the similarity in such patterns, quantification of global and regional functional indices (Table 2) shows distinct, finite differences, in agreement with prior reports. Unknown at this stage is whether such differences can be attributed to species variability or endogenous or exogenous parameter dependencies, as they relate to the conduct of such studies and data analyses, or the modus operandi of the human and murine hearts.

Study of the Murine Cardiac Mechanical Function Using Magnetic Resonance Imaging:

**Figure 18.** (Left, middle) Regional parameter quantification and comparisons in the mouse and the human using VITREA. Mean bullseye plots for wall motion, wall thickening, and regional ejection fraction for C57BL/6J mice (n=9) and human (n=8) datasets over the entire cardiac cycle. Schematic diagram of sectoral representation of the heart according to AHA guidelines; (right) Regional cardiac performance inter-comparison of mouse and human, including wall thickness, wall motion, and rEF variation in the various sectors of the murine and human hearts (sector 17 is excluded from presented

results).

The Current Status, Challenges, and Future Perspectives 371

**Figure 17.** (Left to right) Epicardial and endocardial LV contour definition in mouse long and short axis MRI, and 3D LV blood cavity segmentation; 3D ventriculogram rendition using Vitrea from murine MRI; short axis human MRI, and Vitrea reconstruction of wall motion, wall thickening, and regional ejection fraction from a typical mouse dataset.

Therefore, numerous practical benefits are associated with dedicated, state-of-the-art mouse cardiac MR imaging, including the non-invasive nature of the techniques, the inherent capability to map cardiac morphology and function, for both LV and RV chambers, and their motional patterns. High spatial and temporal resolution imaging can thus be achieved, through execution of high-throughput protocols, yielding direct, accurate estimates of global and regional indices of cardiac function, avoiding any assumptions whatsoever or model-based derivation approaches endorsed by other imaging techniques such as ultrasound.

Based on such analyses, bullseye-plots of regional cardiac function were generated in 17 sector representations of the murine and human hearts (from independent studies in two separate Institutions) showing similar patterns in transmural variations in wall motion and thickness, and regional ejection fraction (Figure 18) in mouse and man. Such spatial patterns observed for mouse and human (in agreement with prior tagging work [Moore 2000]) are supported by two-tailed paired t-tests indicating absence of statistical significant differences in the mean values of wall thickness (p=0.07), wall motion (p=0.051), or regional EF (p=0.065) at the 1% significance. Repeated measures ANOVA indicated significant differences in regional mouse and human for wall thickness (p=0.002) and regional EF (p<0.0001) and insignificant differences for wall motion (p=0.016) at the 1% significance. However, despite the similarity in such patterns, quantification of global and regional functional indices (Table 2) shows distinct, finite differences, in agreement with prior reports. Unknown at this stage is whether such differences can be attributed to species variability or endogenous or exogenous parameter dependencies, as they relate to the conduct of such studies and data

**Figure 17.** (Left to right) Epicardial and endocardial LV contour definition in mouse long and short axis MRI, and 3D LV blood cavity segmentation; 3D ventriculogram rendition using Vitrea from murine MRI; short axis human MRI, and Vitrea reconstruction of wall motion, wall thickening, and regional

Therefore, numerous practical benefits are associated with dedicated, state-of-the-art mouse cardiac MR imaging, including the non-invasive nature of the techniques, the inherent capability to map cardiac morphology and function, for both LV and RV chambers, and their motional patterns. High spatial and temporal resolution imaging can thus be achieved, through execution of high-throughput protocols, yielding direct, accurate estimates of global and regional indices of cardiac function, avoiding any assumptions whatsoever or model-based derivation approaches endorsed by other imaging techniques such as

analyses, or the modus operandi of the human and murine hearts.

ejection fraction from a typical mouse dataset.

ultrasound.

**Figure 18.** (Left, middle) Regional parameter quantification and comparisons in the mouse and the human using VITREA. Mean bullseye plots for wall motion, wall thickening, and regional ejection fraction for C57BL/6J mice (n=9) and human (n=8) datasets over the entire cardiac cycle. Schematic diagram of sectoral representation of the heart according to AHA guidelines; (right) Regional cardiac performance inter-comparison of mouse and human, including wall thickness, wall motion, and rEF variation in the various sectors of the murine and human hearts (sector 17 is excluded from presented results).

As an extension of the development and use of such techniques (DENSE, tagging, HARP) has been the tremendous value for regional cardiac functional quantification and direct applicability to transgenic mice and in pathological states (myocardial infarction, heart failure).

Study of the Murine Cardiac Mechanical Function Using Magnetic Resonance Imaging:

controversial results from human trials [Rosenzweig 2006] due to the uncertain and unclear long-term fate, target-destination of the injected cells, their engraftment and viability,

Employing high-resolution MRI in association with metabolism, iron-oxide labeled cells [Stuckey 2006] can be tracked and visualized thereby monitoring their migration patterns

The superb ability of PET to detect ligand-receptor binding at the nano- to picomolar concentrations and the excellent spatial resolution of MR imaging, have stimulated efforts for the construction of hybrid PET/MRI systems. While the first prototype systems have already been completed [Herzog 2010, Pichler 2008], it will be of interest to see if corporate interest will aid establish such hybrid imaging techniques as tools of the arsenal of other

In the short-lived period of mouse cardiac MRI of the past 15 years, tremendous strides have been made for image-based phenotyping of the cardiovascular system. Such were realized in terms of the scalability of equipment, ease of handling and maintaining animals under proper physiological homeostatic conditions, adaptation of conventional imaging techniques, and inception of new, fast imaging acquisition schemes that have revolutionized

In conclusion, despite the usefulness, practicality, and low costs associated with the study of the mouse, important genetic, developmental, morphological, and cardiovascular system differences exist between mouse and man especially as such are unmasked in pathological conditions or under stress. Physiological results indicate that the optimal ISO anesthetic regimen for mouse studies under anesthesia is approximately 1.5% v/v, yielding stable MAP and HR values comparable to those observed in the animal's conscious state, with a minute-tominute variability in MAP and HR of ≤11%. Based on such recordings, the optimal FiO2 appears to be 50%. The additional use of N2O was associated with higher and more stable values of MAP and HR (at a mixture of 25–50% O2 and 75–50% N2O). Arterial pH values are within the physiological range and varied between 7.20 and 7.43. ISO anesthesia at 1.5% v/v is also associated with mild hyperglycemia (+47%) whereas insulin levels remain unaltered. The protocol described for physiological studies of mice under anesthesia has the potential for high reproducibility in diagnostic modalities, including MRI, microCT, ultrasound, and microPET. This regimen can be useful in phenotypic screening and pharmacological studies of cardiac function in mice and can facilitate the transfer of such work to noninvasive imaging platforms,

Basic MRI studies of murine global cardiac structure and function under optimal physiological conditions, in combination with PCA and other image processing techniques,

cardiac, image-based phenotyping with efficient, high-throughput imaging protocols.

with tremendous potential for both basic science and translational research.

scientific interest and excitement remains high.

and ultimate engraftment fate in tissues of interest.

diagnostic tools for clinical practice and for basic science research.

**4.2. Hybrid system imaging** 

**5. Conclusion** 

The Current Status, Challenges, and Future Perspectives 373

Such approaches are, however, associated with a number of limitations and drawbacks, including the necessity for use of complex algorithms and laborious data post-processing (tagging, HARP, and DENSE), the inherently low spatial resolution for strain quantification (tagging vs. DENSE), the T1-tag dependence, the low SNR performance of DENSE (often with a temporal decreasing dependence following ECG-triggering), and the necessity to eliminate the anti-echo and free-induction decay signals in DENSE through proper acquisition adjustments and/or image subtractions.


**Table 2.** Summary of mean global and regional cardiac mechanical functional indices (±sd) of murine and human myocardium.
