Anna Khimchenko

Extending two-dimensional histology into the third dimension through conventional micro computed tomography.

Histological examination achieves sub-micrometer resolution laterally. In the third dimension, however, resolution is limited to section thickness. In addition, histological sectioning and mounting sections on glass slides introduce tissue-dependent stress and strain. In contrast, state-of-the-art hard X-ray micro computed tomography (μCT) systems provide isotropic sub-micrometer resolution and avoid sectioning artefacts. The drawback of μCT in the absorption contrast mode for visualising physically soft tissue is a low attenuation difference between anatomical features. In this communication, we demonstrate that formalin-fixed paraffin-embedded human cerebellum yields appropriate absorption contrast in laboratory-based μCT data, comparable to conventional histological sections. Purkinje cells, for example, are readily visible. In order to investigate the pros and cons of complementary approaches, two- and three-dimensional data were manually and automatically registered. The joint histogram of histology and the related μCT slice allows for a detailed discussion on how to integrate two-dimensional information from histology into a three-dimensional tomography dataset. This methodology is not only rewarding for the analysis of the human cerebellum, but it also has relevance for investigations of tissue biopsies and post-mortem applications. Our data indicate that laboratory-based μCT as a modality can fill the gap between synchrotron radiation-based μCT and histology for a variety of tissues. As the information from haematoxylin and eosin (H&E) stained sections and μCT data is related, one can colourise local X-ray absorption values according to the H&E stain. Hence, μCT data can correlate and virtually extend two-dimensional (2D) histology data into the third dimension.

Tomographic brain imaging with nucleolar detail and automatic cell counting

Brain tissue evaluation is essential for gaining in-depth insight into its diseases and disorders. Imaging the human brain in three dimensions has always been a challenge on the cell level. In vivo methods lack spatial resolution, and optical microscopy has a limited penetration depth. Herein, we show that hard X-ray phase tomography can visualise a volume of up to 43 mm3 of human post mortem or biopsy brain samples, by demonstrating the method on the cerebellum. We automatically identified 5,000 Purkinje cells with an error of less than 5% at their layer and determined the local surface density to 165 cells per mm2 on average. Moreover, we highlight that three-dimensional data allows for the segmentation of sub-cellular structures, including dendritic tree and Purkinje cell nucleoli, without dedicated staining. The method suggests that automatic cell feature quantification of human tissues is feasible in phase tomograms obtained with isotropic resolution in a label-free manner.

X-ray phase microtomography with a single grating for high-throughput investigations of biological tissue

The high-throughput 3D visualisation of biological specimens is essential for studying diseases and developmental disorders. It requires imaging methods that deliver high-contrast, high-resolution volumetric information at short sample preparation and acquisition times. Here we show that X-ray phase-contrast tomography using a single grating can provide a powerful alternative to commonly employed techniques, such as high-resolution episcopic microscopy (HREM). We present the phase tomography of a mouse embryo in paraffin obtained with an X-ray single-grating interferometer at I13-2 Beamline at Diamond Light Source and discuss the results in comparison with HREM measurements. The excellent contrast and quantitative density information achieved non-destructively and without staining using a simple, robust setup make X-ray single-grating interferometry an optimum candidate for high-throughput imaging of biological specimens as an alternative for existing methods like HREM.

Implementation of a double-grating interferometer for phase-contrast computed tomography in a conventional system nanotom® m

Visualizing the internal architecture of large soft tissue specimens within the laboratory environment in a label-free manner is challenging, as the conventional absorption-contrast tomography yields a poor contrast. In this communication, we present the integration of an X-ray double-grating interferometer (XDGI) into an advanced, commercially available micro computed tomography system nanotom® m with a transmission X-ray source and a micrometer-sized focal spot. The performance of the interferometer is demonstrated by comparing the registered three-dimensional images of a human knee joint sample in phase- and conventional absorption-contrast modes. XDGI provides enough contrast (1.094 ± 0.152) to identify the cartilage layer, which is not recognized in the conventional mode (0.287 ± 0.003). Consequently, the two modes are complementary, as the present XDGI set-up only reaches a spatial resolution of (73 ± 6) μm, whereas the true micrometer resolution in the absorption-contrast mode has been proven. By providing complimentary information, XDGI is especially a supportive quantitative method for imaging soft tissues and visualizing weak X-ray absorbing species in the direct neighborhood of stronger absorbing components at the microscopic level.

Non-destructive phase contrast hard x-ray imaging to reveal the three-dimensional microstructure of soft and hard tissues

X-ray imaging in the absorption contrast mode is an established method of visualising calcified tissues such as bone and teeth. Physically soft tissues such as brain or muscle are often imaged using magnetic resonance imaging (MRI). However, the spatial resolution of MRI is insufficient for identifying individual biological cells within three-dimensional tissue. X-ray grating interferometry (XGI) has advantages for the investigation of soft tissues or the simultaneous three-dimensional visualisation of soft and hard tissues. Since laboratory microtomography (μCT) systems have better accessibility than tomography set-ups at synchrotron radiation facilities, a great deal of effort has been invested in optimising XGI set-ups for conventional μCT systems. In this conference proceeding, we present how a two-grating interferometer is incorporated into a commercially available nanotom m (GE Sensing and Inspection Technologies GmbH) μCT system to extend its capabilities toward phase contrast. We intend to demonstrate superior contrast in spiders (Hogna radiata (Fam. Lycosidae) and Xysticus erraticus (Fam. Thomisidae)), as well as the simultaneous visualisation of hard and soft tissues. XGI is an imaging modality that provides quantitative data, and visualisation is an important part of biomimetics; consequently, hard X-ray imaging provides a sound basis for bioinspiration, bioreplication and biomimetics and allows for the quantitative comparison of biofabricated products with their natural counterparts.

Automatic histology registration in application to x-ray modalities

Registration of microscope images to Computed Tomography (CT) 3D volumes is a challenging task because it requires not only multi-modal similarity measure but also 2D-3D or slice-to-volume correspondence. This type of registration is usually done manually which is very time-consuming and prone to errors. Recently we have developed the first automatic approach to localize histological sections in μCT data of a jaw bone. The median distance between the automatically found slices and the ground truth was below 35 μm. Here we explore the limitations of the method by applying it to three tomography datasets acquired with grating interferometry, laboratory-based μCT and single-distance phase retrieval. Moreover, we compare the performance of three feature detectors in the proposed framework, i.e. Speeded Up Robust Features (SURF), Scale Invariant Feature Transform (SIFT) and Affine SIFT (ASIFT). Our results show that all the feature detectors performed significantly better on the grating interferometry dataset than on other modalities. The median accuracy for the vertical position was 0.06 mm. Across the feature detector types the smallest error was achieved by the SURF-based feature detector (0.29 mm). Furthermore, the SURF-based method was computationally the most efficient. Thus, we recommend to use the SURF feature detector for the proposed framework.

Histology-validated x-ray tomography for imaging human coronary arteries

Heart disease is the number one cause of death worldwide. To improve therapy and patient outcome, the knowledge of anatomical changes in terms of lumen morphology and tissue composition of constricted arteries is crucial for designing a localized drug delivery to treat atherosclerosis disease. Traditional tissue characterization by histology is a pivotal tool, although it brings disadvantages such as vessel morphology modification during decalcification and slicing. X-ray tomography in absorption and phase contrast modes yields a deep understanding in blood vessel anatomy in healthy and diseased stages: measurements in absorption mode make visible highly absorbing tissue components including cholesterol plaques, whereas phase contrast tomography gains better contrast of the soft tissue components such as vessel walls. Established synchrotron radiation-based micro-CT techniques ensure high performance in terms of 3D visualization of highly absorbing and soft tissues.

Hierarchical imaging of the human knee

Among the clinically relevant imaging techniques, computed tomography (CT) reaches the best spatial resolution. Sub-millimeter voxel sizes are regularly obtained. For investigations on true micrometer level lab-based μCT has become gold standard. The aim of the present study is the hierarchical investigation of a human knee post mortem using hard X-ray μCT. After the visualization of the entire knee using a clinical CT with a spatial resolution on the sub-millimeter range, a hierarchical imaging study was performed using a laboratory μCT system nanotom m. Due to the size of the whole knee the pixel length could not be reduced below 65 μm. These first two data sets were directly compared after a rigid registration using a cross-correlation algorithm. The μCT data set allowed an investigation of the trabecular structures of the bones. The further reduction of the pixel length down to 25 μm could be achieved by removing the skin and soft tissues and measuring the tibia and the femur separately. True micrometer resolution could be achieved after extracting cylinders of several millimeters diameters from the two bones. The high resolution scans revealed the mineralized cartilage zone including the tide mark line as well as individual calcified chondrocytes. The visualization of soft tissues including cartilage, was arranged by X-ray grating interferometry (XGI) at ESRF and Diamond Light Source. Whereas the high-energy measurements at ESRF allowed the simultaneous visualization of soft and hard tissues, the low-energy results from Diamond Light Source made individual chondrocytes within the cartilage visual.

Three-dimensional registration of synchrotron radiation-based micro-computed tomography images with advanced laboratory micro-computed tomography data frommurine kidney casts

Malfunction of oxygen regulation in kidney and liver may lead to the pathogenesis of chronic diseases. The underlying mechanisms are poorly understood. In kidney, it is hypothesized that renal gas shunting from arteries to veins eliminates excess oxygen. Such shunting is highly dependent on the structure of the renal vascular network. The vascular tree has so far not been quantified under maintenance of its connectivity as three-dimensional imaging of the vessel tree down to the smallest capillaries, which in mouse model are smaller than 5 μm in diameter, is a challenging task. An established protocol uses corrosion casts and applies synchrotron radiation-based micro-computed tomography (SRμCT), which provides the desired spatial resolution with the necessary contrast. However, SRμCT is expensive and beamtime access is limited. We show here that measurements with a phoenix nanotomrm (General Electric, Wunstorf, Germany) can provide comparable results to those obtained with SRμCT, except for regions with small vessel structures, where the signal-to-noise level was significantly reduced. For this purpose the nanotom®m measurement was compared with its corresponding measurement acquired at the beamline P05 at PETRA III at DESY, Hamburg, Germany.

Automatic deformable registration of histological slides to μCT volume data: DEFORMABLE REGISTRATION FROM HISTOLOGY TO VOLUME DATA

Localizing a histological section in the three‐dimensional dataset of a different imaging modality is a challenging 2D‐3D registration problem. In the literature, several approaches have been proposed to solve this problem; however, they cannot be considered as fully automatic. Recently, we developed an automatic algorithm that could successfully find the position of a histological section in a micro computed tomography (μCT) volume. For the majority of the datasets, the result of localization corresponded to the manual results. However, for some datasets, the matching μCT slice was off the ground‐truth position. Furthermore, elastic distortions, due to histological preparation, could not be accounted for in this framework.