Extraction of the 3D local orientation of myocytes in human cardiac tissue using X-ray phase-contrast micro-tomography and multi-scale analysis
Graphical abstract
Introduction
Cardiovascular diseases remain one of today’s most serious health problems, which motivates the search for a deeper understanding of myocardial function. The relation between mechanical function, hemodynamics, perfusion, diffusion, transfer rate and the adaptive structural changes emerging in cardiac diseases remain poorly understood. In addition, a better understanding of the anatomy of 3D tissue would be a great help to further the understanding of cardiac pathophysiology. This would require producing a 3D multi-scale atlas of the human heart from the cell level to that of the organ, unavailable today.
Cardiac myocytes are approximately cylindrical, 10 µm in diameter and 100 µm long. Their membrane is about 8 nm thick (Iaizzo, 2009). Today, it is not possible to image a whole adult human heart at a micrometric resolution (<3 µm), thereby providing the individual myocyte arrangement (for the sake of simplification, we refer to these fibres as myocytes arrangements) within the tissue. Histological slices deliver high-resolution (µm) 2D images of ex vivo samples. Polarised light imaging (PLI, Jouk et al., 2007) provides the 3D orientation of populations of myocytes (100 × 100 ×500 µm3) of ex vivo hearts based on their birefringence property. Magnetic resonance cardiac diffusion tensor imaging (cDTI) measures the magnitude and direction of intra-myocardial water diffusion (Ferreira et al., 2014) (2.1 × 1.7 × 6 mm3), but various assumptions on the tissue’s micro-architecture (Frindel, Robini, Croisille, Zhu, 2009, Ferreira, Kilner, McGill, Nielles-Vallespin, Scott, Ho, McCarthy, Haba, Ismail, Gatehouse, de Silva, Lyon, Prasad, Firmin, Pennell, 2014, Bernus, Radjenovic, Trew, LeGrice, Sands, Magee, Smaill, Gilbert, 2015) are still necessary to interpret in vivo measurements and cardiac muscle mobility in normal and diseased hearts (Wang et al., 2012). High-resolution MRI setups (Baltes et al., 2009) can achieve 30 × 30 × 300 µm3, but this is larger than the size of myocytes. At the same time, μCT imaging has been used by Stephenson et al. (2012) to image rat and rabbit hearts at a resolution within the 13–41 µm range to estimate heart morphology. Similarly, van Deel et al. (2016) imaged in vivo mice hearts to extract quantitative metabolic markers at an isotropic 40 µm resolution. Recently, some preliminary fibre analysis has been conducted in the atrial section of canine hearts by Aslanidi et al. (2013) at an isotropic 36 µm resolution and in fresh ex vivo human hearts by Zhao et al. (2015) at an isotropic resolution of 49 µm. However, these techniques are only based on X-ray absorption and they all use an isotropic resolution, which is higher than the lateral dimensions of a cardiomyocyte. A review of the literature examining studies dedicated to the measurement of the local 3D cardiomyocyte orientation was conducted (Appendix). It was restricted to studies providing measurements of the helix angle range through the cardiac wall. It can be noted that this range differs with the imaging modality, the protocol used and the heart model (mouse, rat, dog, human). However, maximum values are observed when very high spatial resolution techniques are used (Streeter, Bassett, 1966, Streeter, Spotnitz, Patel, Ross, Sonnenblick, 1969, LeGrice, Smaill, Chai, Edgar, Gavin, Hunter, 1995, Gilbert, D., A.P., White, Tanner, A.V., Dobrzynski, Bernus, Radjenovic, 2012).
In this study, our aim was to access the local 3D myocyte arrangements within fresh human post-mortem heart samples. To handle the above-mentioned limitations, we used the synchrotron radiation phase micro-computed tomography setup (SR-PCT) at the European Synchrotron Radiation Facility (ESRF) beamline ID19 to image the samples using phase contrast imaging, at a voxel size of 3.5 µm. X-ray phase contrast imaging yields several orders of magnitude higher sensitivity in light materials compared to standard X-ray attenuation-based imaging (Momose et al., 1995). Phase contrast SR-PCT is increasingly used to characterize soft tissue due to its histology-like contrast (Pierre, Langer, Boistel, Cloetens, 2007, Boistel, Aubin, Cloetens, Langer, Gillet, Josset, Pollet, Herrel, 2011, Marinescu, Langer, Durand, Olivier, Chabrol, Rositi, Chauveau, Cho, Nighoghossian, Berthezène, Peyrin, Wiart, 2013, Lang, Zanette, Dominietto, Langer, Rack, Schulz, Le Duc, David, Mohr, Pfeiffer, Muller, Weitkamp, 2014). The first acquisitions were made on a healthy fetal rabbit heart (Baličević et al., 2015) and a healthy human heart (Varray et al., 2013). The isotropic 3D spatial resolution of SR-PCT makes it possible to represent a myocyte by several voxels and to process the data to estimate the transmural myocyte arrangements (in terms of orientation) at several scales. The reconstructed tomographic images were previously compared to corresponding histological images, to confirm that the cardiomyocytes were visible (Mirea et al., 2015).
In this paper, the 3D-acquired tomograms were processed to extract the local fibre arrangement by measuring the transmural (helix and transverse) fibre angles. We developed a multi-scale scheme to extract the orientation of the fibres at various spatial resolutions and to evaluate the accuracy of the angular estimation. Finally, we used a fibre tracking technique to display the fibres arrangements of several cardiac tissue samples.
The paper is organised as follows. First, we present the cardiac tissue samples and the acquisition procedure. Second, the multi-scale method used to extract the 3D cardiomyocyte arrangements is described. Third, the results obtained at different resolutions in both simulations and experiments are highlighted. A discussion concludes the paper.
Section snippets
Extraction and preparation of the human heart samples
Eight fresh human heart samples were supplied by the Medico-Legal Institute of Lyon IML HCL (N° DC-2012-1588), taken from the hearts of two males aged 32 and 35 years. These patients died a violent death and had no past medical history, no treatment and no addiction. Less than 24 h after death, the hearts were removed and preserved in a 10% formalin solution to ensure good preservation. A slight shrinkage effect due to the formalin solution was present but limited by the short time between
Extraction of the local orientation of myocytes
In SR-PCT images, the orientation of myocytes is directly related to the fibres’ local orientation. We propose a dedicated methodology to extract this 3D orientation and compute both θ and ϕ values at the same time. We first computed the module of the 3D Fourier transform (FT) of one 3D sub-volume Vsub of the reconstructed volume:
In the spatial domain, the stack of myocytes has a particular arrangement in beams or lines or “fibres” (Fig. 7a). In the Fourier domain (Fig. 7b), such
Method validation
We tested our local fibres orientation extraction method using 2D histological heart tissue sections extracted immediately in the vicinity of the LV posterior sample (Fig. 9). We also selected two image sections extracted from the 3D X-ray reconstructed volumes of sample #6 (LV_post_S1_2), one with homogeneous tissue (mainly myocytes) and one containing other biological material (vessel, cleavage planes, etc.). We computed the orientation of the myocytes, on each 2D image using our method. We
Discussion
In this paper, the fibres structure of the human heart is investigated at a high resolution using both simulated and real data. The data are X-ray SR-PCT images of eight LV cardiac tissue samples, acquired at ID19, ESRF, and reconstructed at an isotropic voxel size of 3.5 µm. By extracting the fibres orientation in the Fourier domain, the low-pass shape of the myocyte arrangements was identified and gave the local direction of the fibres through the LV wall. The myocyte orientation was
Acknowledgements
The authors thank Cécile Olivier, Lihui Wang, David Rousseau, and Matthieu Ozon for their help for acquiring data at the ESRF and Pierre Croisille and Yue-Min Zhu for fruitful discussions. They also thank the ESRF for support throughout the MD-701 experiment and for the assistance of Alexander Rack and Anne Bonnin from the ESRF ID19 group. This work was done within the framework of LabEx PRIMES (ANR-11-LABX-006).
References (40)
- et al.
Structural–functional correlates of the 3-dimensional arrangement of the myocytes making up the ventricular walls
J. Thoracic Cardiovas. Surg.
(2008) - et al.
Probing dynamic myocardial microstructure with cardiac magnetic resonance diffusion tensor imaging
J. Cardiovas. Magn. Reson.
(2014) - et al.
Comparison of diffusion tensor imaging by cardiovascular magnetic resonance and gadolinium enhanced 3d image intensity approaches to investigation of structural anisotropy in explanted rat hearts
J. Cardiovas. Magn. Reson.
(2015) - et al.
In vivo cardiovascular magnetic resonance diffusion tensor imaging shows evidence of abnormal myocardial laminar orientations and mobility in hypertrophic cardiomyopathy
J. Cardiovas. Magn. Res.
(2014) - et al.
Comparison of regularization methods for human cardiac diffusion tensor MRI
Med. Image Anal.
(2009) - et al.
Multiscale modeling and simulation of the cardiac fiber architecture for DMRI
IEEE Trans. Biomed. Eng.
(2012) - et al.
Application of micro-computed tomography with iodine staining to cardiac imaging, segmentation, and computational model development
IEEE Trans. Med. Imaging
(2013) - et al.
Assessment of myofiber orientation in high resolution phase-contrast ct images
- et al.
Micro mri of the mouse brain using a novel 400mhz cryogenic quadrature rf probe
NMR Biomed.
(2009) - et al.
Whispering to the deaf: communication by a frog without external vocal sac or tympanum in noisy environments
PLoS ONE
(2011)
Design–based stereological estimation of the total number of cardiac myocytes in histological sections
Basic Res. Cardiol.
In vivo quantitative assessment of myocardial structure, function, perfusion and viability using cardiac micro-computed tomography
J. Visual. Exp.: JoVE
Forensic Histopathology: Fundamentals and Perspectives
Practical cone-beam algorithm
J. Optical Soc. Am. A
Visualization and quantification of whole rat heart laminar structure using high-spatial resolution contrast-enhanced mri
Am. J. Physiol. - Heart Circulat. Physiol.
Left ventricular fibre architecture in man
Br. Heart J.
Ex vivo 3d diffusion tensor imaging and quantification of cardiac laminar structure
Magn. Reson. Med.
Handbook of Cardiac Anatomy, Physiology, and Devices
Analysis of the fiber architecture of the heart by quantitative polarized light microscopy. accuracy, limitations and contribution to the study of the fiber architecture of the ventricles during fetal and neonatal life.
Eur. J. Cardiothorac. Surg.
Experimental comparison of grating- and propagation-based hard x-ray phase tomography of soft tissue
J. Appl. Phys.
Cited by (29)
Measurement of local orientation of cardiomyocyte aggregates in human left ventricle free wall samples using X-ray phase-contrast microtomography
2022, Medical Image AnalysisCitation Excerpt :The transmural distributions of the orientation of myocyte aggregates calculated by the classic method and the skeleton method are very similar. These results confirm the preliminary results of Varray et al. (2017), and show that at a global scale, the behaviors of the two methods are comparable. However, even if the skeleton method can provide angular values at a global scale, it needs a preliminary binarization step and it is much more time consuming than the classic method, so it is not designed for this.
Imaging biomarkers for cardiovascular diseases
2019, Handbook of Medical Image Computing and Computer Assisted InterventionUse of the intrusion angle to describe the radial orientation of local cardiomyocytes in the left ventricle
2023, Journal of MorphologyOscillation of the orientation of cardiomyocyte aggregates in human left ventricle free wall
2023, Journal of AnatomyDescription of the Intrusion Angle of Local Cardiomyocyte Aggregates in Human Left Ventricular Free Wall Using X-ray Phase-Contrast Tomography
2023, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)A tomographic microscopy-compatible Langendorff system for the dynamic structural characterization of the cardiac cycle
2022, Frontiers in Cardiovascular Medicine