Increased contrast for soft tissue

Phase-contrast x-ray tomography

Phase-imaging techniques make use of the real part of a material’s refractive index. At hard x-ray energies, this can be a factor of 1000 greater than that of the complex part, which gives rise to absorption.^{Reference Alberto, Paola and Pekka12} Phase-contrast methods include free-space propagation,^{Reference Zabler, Cloetens, Guigay, Baruchel and Schlenker13} grating interferometry,^{Reference Weitkamp, Diaz, David, Pfeiffer, Stampanoni, Cloetens and Ziegler14} Zernike phase contrast,^{Reference Tkachuk, Duewer, Cui, Feser, Wang and Yun15} and analyzer-based^{Reference Hagen, Maghsoudlou, Totonelli, Diemoz, Endrizzi, Rigon, Menk, Arfelli, Dreossi, Brun, Coan, Bravin, De Coppi and Olivo16} methods. The free-space propagation method is the simplest and requires a coherent, or partially coherent, x-ray source (from a synchrotron, or laboratory x-ray tube having a small spot size) and a relatively large distance between sample and detector. Refraction of x-rays at interfaces within the sample is then manifested as fringes in the radiographs, leading to “edge-enhancement.” Walton et al.^{Reference Walton, Bradley, Withers, Newton, Watson, Austin and Sherratt17} used this technique to elucidate changes in artery wall architecture between the pressurized and unpressurized states, albeit in an ex situ study looking at fixed samples. This type of phase contrast has also been employed for studying novel tissue engineering materials, such as bioglasses.^{Reference Poologasundarampillai, Wang, Li, Nakamura, Bradley, Lee, Stevens, McPhail, Kasuga and Jones18}

More advanced techniques measure the phase shift through the object by encoding wavefront distortions in changes in the intensity profile using a mask. Pfeiffer et al.^{Reference Pfeiffer, Weitkamp, Bunk and David19} showed that this type of imaging can be carried out using conventional low-brilliance x-ray tubes. The resolution is currently limited by difficulties in fabricating fine gratings down to ∼10 μm. Schultz et al. reported ∼16-μm spatial resolution imaging of a human cerebellum.^{Reference Schulz, Weitkamp, Zanette, Pfeiffer, Beckmann, David, Rutishauser, Reznikova and Müller20} A number of competing techniques, including speckle and edge-illumination techniques, are being developed that have simpler setups and can achieve higher resolutions. These techniques have recently been shown to work well with conventional laboratory x-ray sources.^{Reference Hagen, Maghsoudlou, Totonelli, Diemoz, Endrizzi, Rigon, Menk, Arfelli, Dreossi, Brun, Coan, Bravin, De Coppi and Olivo16,Reference Zanette, Zhou, Burvall, Lundström, Larsson, Zdora, Thibault, Pfeiffer and Hertz21,Reference Wang, Kashyap and Sawhney22}

At much higher resolutions, phase-contrast nanotomography using the holography approach has enabled the hierarchical structure of bone to be imaged, with information gained about mineralization and collagen orientations down to ∼60-nm resolution.^{Reference Langer, Pacureanu, Suhonen, Grimal, Cloetens and Peyrin23} X-ray microscopes make use of Zernike phase contrast. Walton et al.^{Reference Walton, Bradley, Withers, Newton, Watson, Austin and Sherratt17} used this technique to visualize small-scale fiber bundles within rat artery walls in 3D at 150-nm resolution (see Figure 2).

Figure 2.

Multiscale x-ray phase-contrast imaging of a rat artery wall. (a) Rendering of the two main layers (media [middle] in yellow, adventitia [outer] in gray) at 1.2-μm resolution, with the concentric lamella in the medial layer shown in yellow. The difference in wall morphology between the unpressurized (0 mm Hg, 0 Pa) and pressurized (110 mm Hg, 14.7 kPa) states is shown in the inset. (b) Higher resolution imaging of three successive lamella (orange) in the media at 150-nm resolution.^{Reference Walton, Bradley, Withers, Newton, Watson, Austin and Sherratt17}

Coherent diffraction imaging (CDI) can theoretically achieve 10-nm resolution or better by using a “lensless” approach.^{Reference Marchesini, Chapman, Hau-Riege, London, Szoke, He, Howells, Padmore, Rosen, Spence and Weierstall24} This has enabled the ultrastructure of collagen to be determined in intact tendons, including the measurement of the periodic spacing of collagen molecules within fibrils.^{Reference Berenguer, Bean, Bozec, Vila-Comamala, Zhang, Kewish, Bunk, Rodenburg and Robinson25} CDI can be used in a scanning mode, termed ptychography, to image relatively large nonisolated objects, and can also be combined with tomography to yield 3D information to study, for example, 3D mineral distributions in dentin,^{Reference Zanette, Enders, Dierolf, Thibault, Gradl, Diaz, Guizar-Sicairos, Menzel, Pfeiffer and Zaslansky26} the lacuno-canalicular network in bone (the network of spaces in which cells reside),^{Reference Dierolf, Menzel, Thibault, Schneider, Kewish, Wepf, Bunk and Pfeiffer27} and the structure of cells^{Reference Giewekemeyer, Hackenberg, Aquila, Wilke, Groves, Jordanova, Lamzin, Borchers, Saksl, Zozulya, Sprung and Mancuso28} and chromosomes.^{Reference Shemilt, Verbanis, Schwenke, Estandarte, Xiong, Harder, Parmar, Yusuf, Zhang and Robinson29}

Hierarchical multiscale imaging

Multiscale tomography brings together images from different scales. These can be uncorrelated, where different features are examined and quantified at different scales, or correlated, where a feature is identified at a coarse scale and then studied in greater detail.^{Reference Burnett, McDonald, Gholinia, Geurts, Janus, Slater, Haigh, Ornek, Almuaili, Engelberg, Thompson and Withers7}

Such multiscale approaches can be carried out within the same instrument by using different instruments having the same imaging modality (x-rays), or by combining information from different modalities (see Figure 3). Uncorrelated multiscale imaging is much easier to undertake because there is no need to locate and co-register the same volume of interest (which may lie beneath the surface) across different instruments. Uncorrelated imaging is useful when statistical data are required at each scale level (e.g., across the structural scales in bone in Figure 1). If the range of resolutions to be spanned is not too large, correlative multiscale imaging can be carried out using a higher magnification region of interest tomography without cutting up the sample.^{Reference Kyrieleis, Titarenko, Ibison, Connolley and Withers49} Walton et al.^{Reference Walton, Bradley, Withers, Newton, Watson, Austin and Sherratt17} linked together nano- and micro-CT imaging of an artery wall (Figure 2), without cutting the sample, by performing the nano-CT near the sample surface. In some cases, x-ray imaging can be used to ensure that time-consuming serial sectioning is focused on a region known to contain the feature of interest.

Figure 3.

Levels of hierarchy associated with image modalities used for analysis of vascular networks. Each order of magnitude in resolution reveals a new functional level (structure, tissue, and cell). Adapted with permission from Reference Reference Heinzer, Krucker, Stampanoni, Abela, Meyer, Schuler, Schneider and Müller68. © 2006 Elsevier. Note: CT, computed tomography; SR, synchrotron radiation; SEM, scanning electron microscopy; ROI, region of interest.

Despite the attractiveness of CT as a part of multiscale imaging schemes, little work has been reported to date, particularly for biological materials, and this is sure to be an area of increased activity going forward. For instance, the possibilities of multiscale imaging of subcellular components by combining soft x-ray microscopy, light microscopy, and 3D SEM techniques are just beginning to be explored.^{Reference Patwardhan, Ashton, Brandt, Butcher, Carzaniga, Chiu, Collinson, Doux, Duke, Ellisman, Franken, Grunewald, Heriche, Koster, Kuhlbrandt, Lagerstedt, Larabell, Lawson, Saibil, Sanz-Garcia, Subramaniam, Verkade, Swedlow and Kleywegt50}

Multifaceted correlative tomography

Recently, 2D imaging techniques have been combined to provide complementary information on the same surface, or near surface, region of interest. Coined “correlative microscopy,”^{Reference Caplan, Niethammer, Taylor and Czymmek51} this approach is now routinely applied in the life sciences, in particular, as a means of combining optical fluorescence microscopy and electron microscopy.^{Reference Giepmans52} Burnett et al.^{Reference Burnett, McDonald, Gholinia, Geurts, Janus, Slater, Haigh, Ornek, Almuaili, Engelberg, Thompson and Withers7} adopted the phrase “correlative tomography” to describe the 3D analogue of this. The approach involves using surface features identified from CT data to locate sub-regions, physically cut from the sample using a focused ion beam, for higher resolution electron imaging or multifaceted analysis. Initially, this was carried out manually, but more recently, semiautomated systems have been developed.^{Reference Merkle, Lechner, Steinbach, Gelb, Kienle, Phaneuf, Unrau, Singh and Chawla53} Correlative approaches have also been developed to link optical imaging with CT. Handshuh et al.^{Reference Handschuh, Baeumler, Schwaha and Ruthensteiner54} developed a workflow to correlate CT, serial section light microscopy and TEM, using osmium stained tissue.

There is an increasing interest in combining high-resolution imaging modalities in a single instrument, with synchrotron beamlines being particularly suited to this. For instance, fluorescence imaging and soft x-ray tomography have been combined to achieve multimodal imaging of whole cells and subcellular components.^{Reference Schneider, Guttmann, Rehbein, Werner and Follath55,Reference Smith, McDermott, Do, Leung, Panning, Le Gros and Larabell56} Liebl et al.^{Reference Liebi, Georgiadis, Menzel, Schneider, Kohlbrecher, Bunk and Guizar-Sicairos57} combined tomographic imaging of trabecular bone with small-angle x-ray scattering tomography such that morphology and crystal orientation can be related. A natural extension of this is to correlate imaging with microscale mechanical tests such that the local properties of bone can be mapped and related to the collagen orientation.

Temporal (4D) imaging

For many years, the noninvasive nature of CT has been exploited to follow the degradation of man-made materials.^{Reference Withers and Preuss58,Reference Maire and Withers59} Attention is starting to focus on natural materials, however, time-lapse x-ray imaging of biological materials is challenging. The mechanical and physical properties are highly dependent on the environment, with dehydration being particularly problematic for soft tissues. Very high doses of x-rays introduce molecular damage that can progressively affect the mechanical properties.^{Reference Barth, Zimmermann, Schaible, Tang, Alliston and Ritchie60}

To date, there are relatively few high-resolution 4D studies of biological materials. Lowe et al.^{Reference Lowe, Garwood, Simonsen, Bradley and Withers61} used native x-ray absorption contrast to image the development of selected tissues during metamorphosis of a painted lady butterfly over many days. By contrast, Walker et al.^{Reference Walker, Schwyn, Mokso, Wicklein, Müller, Doube, Stampanoni, Krapp and Taylor62} used fast synchrotron imaging at much shorter extremes to study the dynamics of the flight motor in a living blow fly, which oscillates at 150 Hz (see Figure 4).

Figure 4.

Four-dimensional (4D) imaging of the flight motor in a blow fly. (a) Rendering of the thorax showing power muscles (yellow to red) and steering muscles (green to blue). (b) Rendering of the steering muscles at three time points (t) for high-amplitude wing motion. Adapted with permission from Reference Reference Walker, Schwyn, Mokso, Wicklein, Müller, Doube, Stampanoni, Krapp and Taylor62. © 2014 PLoS Biology.

Studies are starting to emerge exploiting in situ loading with CT for biological materials. Jang et al.^{Reference Jang, Lin, Seo, Etchin, Merkle, Fahey and Ho63} studied deformation at the micrometer scale in the bone-periodontal ligament during compressive loading, while Wiest et al.^{Reference Wiest, Zabler, Rack, Fella, Balles, Nelson, Schmelzeisen and Hanke64} combined real-time radiography with tomography to study local deformation under fatigue of a two-piece dental implant. Recently, nanotomography has been used to study progressive crack propagation in beetle cuticle in situ under tension (Figure 5), giving insights into failure modes and their link to mechanical properties.^{Reference Patterson, Cordes, Henderson, Mertens, Clarke, Lu, Hartwell, Hornberger, Merkle, Etchin, Tkachuk, Leibowitz, Trapp, Qiu, Zhang, Bale, Withers and Bradley65} A notched sample of the beetle elytron (the hardened forewing) was prepared and mounted in a nanomechanical test stage, which enables tomography scans to be taken while the sample is under load. The cuticle has a layered structure in which the orientation of chitin fiber bundles (macrofibers) in successive layers differs by ∼90°. Cracks grow through this “cross-ply” structure by displacement of neighboring macrofibers within a layer and by failure of individual macrofibers. The tortuous crack morphology therefore closely follows the orientation of the plies.

Figure 5.

Virtual CT section showing progressive crack growth in beetle cuticle (a) with applied load under tension and (b) with a rendering of the crack path in successive layers shown. The layers overlap, but have been offset horizontally for clarity.^{Reference Patterson, Cordes, Henderson, Mertens, Clarke, Lu, Hartwell, Hornberger, Merkle, Etchin, Tkachuk, Leibowitz, Trapp, Qiu, Zhang, Bale, Withers and Bradley65} The data were obtained by x-ray nanotomography.

Summary and perspectives

In view of the many and varied ways in which the hierarchical structures in biological materials give rise to outstanding and unusual properties, it is surprising that little multiscale 3D or time-lapse (4D) imaging has been carried out to date. Most studies have relied upon a combination of ex situ mechanical testing followed by SEM imaging of the fracture surface to try and understand behavior. Recent advances in x-ray imaging technology, coupled with correlative multiscale tomography, promise to change this landscape for studying biological materials and, more broadly, across materials science. Phase-contrast imaging techniques are now available across length scales from hundreds of micrometers to tens of nanometers, but have yet to be linked up within correlative frameworks. There is little doubt that developments in automating the correlation between different instruments and imaging modalities will see a rapid expansion of multiscale, multifaceted correlative imaging approaches in the future, providing a more holistic understanding from macromolecules, through the whole cell architecture and extracellular matrix to the behavior of bulk tissue.