TOMCAT Research
Here is a summary of the major research activities within the TOMCAT team. Click on each item to be directed to a more detailed project description.
Beamline-Related Research
Dynamic in-vivo lung imaging at the micrometer scale
In vivo tomographic slice of a newborn rat’s lung with alveolar resolution.
With the recent development and realization of in vivo tomographic X-ray microscopy for the study of lung physiology at the micrometer scale [1] we are now able to study a variety of phenomena that take place in the mammalian lung, or more precisely, in a murine and rat animal model. Following that, we are currently pursuing two parallel lines of research. On the one hand, we are constantly improving our established imaging technique in terms of acquisition speeds, dose reduction and post-processing. Here we follow the ultimate goal towards a free-breathing microscopic lung imaging animal model, but the methods remain transferable to other in vivo studies as well. On the other hand, we have now the availability to study for the first time the detailed (microscopic) mechanisms behind various lung diseases. As the first example, we are currently investigating the origins of ventilator-induced lung injury (VILI), which is being one of the major reasons for the high mortality in ventilator-treated patients with acute respiratory distress syndrome (ADRS) – a pathological condition where ventilator treatment is highly necessary.
Apart from that, our interests also extend to the study of other lung diseases such as emphysema and fibrosis. Our results will have direct implications on the current knowledge and understanding when dealing with lung diseases in clinics. For instance, a fundamental understanding of VILI would be crucial for developing better ventilation strategies for patients.
Apart from that, our interests also extend to the study of other lung diseases such as emphysema and fibrosis. Our results will have direct implications on the current knowledge and understanding when dealing with lung diseases in clinics. For instance, a fundamental understanding of VILI would be crucial for developing better ventilation strategies for patients.
X-ray phase contrast tomographic imaging and analysis of the lung at the micrometer scale
Phase contrast tomographic reconstructions of intact mouse lungs at the micrometer scale.
The gas exchange between the ambient air and the blood takes place in the mammalian lung. The area at the distal part of the organ is subdivided into functional units called acini, which start distally of the terminal bronchioles. Despite of their importance, limited knowledge exists about their three-dimensional (3D) dynamics and development, as the borders between acini are not detectable in classical two-dimensional (2D) histology slices. Furthermore the histological fixation makes any kind of dynamical studies infeasible. Recent developments in synchrotron based X-ray tomographic microscopy helped us to overcome these limitations. Volumetric lung datasets of fresh post-mortem murine lungs with micrometer spatial and sub-second temporal resolutions at different pressure levels are nowadays routinely acquired using synchrotron propagation based phase contrast X-ray tomographic microscopy at the TOMACT beamline of the Swiss Light Source [1]. The aim of this project is to investigate the 3D structural changes of the acinus at the micrometer scale at different pressure levels and days of lung development. Due to the complexity of the lung microstructure and the size of individual acini a specialized 3D segmentation algorithm [2], as well as a technique to enlarge the field of view has been developed [3]. Our methods can run in parallel on multiple datasets. By analyzing a large number of samples we aim to extract the acinar skeleton and address the question of the acinar deformation patterns by quantifying local structural differences at different pressure levels with statistical significance.
Reconstruction of the mouse brain vascular networks with high-resolution synchrotron radiation X-ray tomographic microscopy
A schematic overview of the project: from image acquisition to processing of TB-sized dataset of the mouse brain.
Brain vessels play an important role in the process of maintaining normal brain function. An in-depth knowledge of the vascular structure and topology is essential for better understanding the pathophysiological cerebral processes. Within the context of the Human Brain Project (HBP), this project aims to reconstruct, in a non-destructive way, the entire vascular system of the mouse brain with high-resolution. Synchrotron-radiation X-ray phase-contrast tomographic microscopy at the Swiss Light Source of the Paul Scherrer Institute (Switzerland) is used as a non-invasive key technology for fast image acquisition of mouse brain samples, previously perfused with contrast agent and kept in steady-state conditions. Current sample preparation procedure suggests an optimal perfusion procedure based on indian ink [1]. However, new sample preparation methods are explored within the project. To fully reconstruct the complete cerebrovascular network of the mouse brain with 1μm resolution, several local tomographic scans need to be acquired, reconstructed and stitched together [2] in order to cover the whole brain volume, thus leading to TB-sized datasets. All the pioneering efforts to address the challenging task of analysing such large datasets are pointing towards new horizons in the investigation of large biological samples with 3D high spatial resolution.
X-ray grating interferometry for phase-contrast imaging at the Swiss Light Source
DPC setup at the TOMCAT beamline of the Swiss Light Source.
With the development and realization of differential phase contrast (DPC) imaging based on grating interferometry (GI) [1], which provides high sensitivity to electron-density variations within soft tissues, we are now able to study a variety of phenomena in the medical field such as the amyloid plaque distribution in mouse brains [2] and the evolution of 3D substructures developing during the early stage of tumour formation [3]. Most recently, data acquisition and post-processing have been optimised at TOMCAT, thus enabling a full phase volume to be acquired in 32 min. As detector, a sCMOS camera with a 16-bit nominal dynamic range is attached to a 1:1 optical microscope which results in a 6.5 μm pixel size and 1.3 cm FOV for imaging. Our research is currently focused on the further improvements of the DPC setup in terms of acquisition speeds, dose reduction and post-processing. In addition to that, the ongoing efforts aim to improve the hard components (new gratings manufacture and setup versatility). On the other hand, we are also interested to experimentally exploit the capabilities of DPC imaging in a wider range of applications from medicine to material science specifically using synchrotron light.
Impact of phase-contrast X-ray imaging in cochlear micro-anatomy investigation
Tomographic slices of: (a) ROI in the cochlea, (b) incus and (c) stapes. Images acquired at the TOMCAT beamline, PSI Switzerland.
The human cochlea is composed of about two and three-fourth turns, but unusual anatomy with cochlear three turns has been described [1]. It is surrounded by a compact bony structure and represents the hardest bone in the body with a trilamellar arrangement with islands of modified cartilage and high-mineral content, which increases the stiffness of the bony labyrinth [2]. There is now much interest on cochlear anatomy due to surgical approaches that electrically stimulate the auditory nerve (cochlear implantation, CI). In cochlear implantation (CI), the large variations in cochlear lengths, angles between turns, and position in the skull base can influence the straightforwardness for the insertion of a CI electrode particularly passing the first turn. Should we use personalized electrodes in such a way that they minimize trauma during insertion? This is important, because trauma against the medial wall of the ST (scala tympani) can damage the spiral ganglion, which is the prime target of electric stimulation by the electrodes of the implant. While a state of the art micro CT imaging device such as the Scanco μCT 100 can provide resolutions of less than 4 μm to answer the above-stated questions, its ability to image soft tissue is limited. For this reason, 1 μm resolution images of human cochlea with and without implants and ear ossicles will be acquired with synchrotron-based tomographic microscopy at the Swiss Light Source (SLS) of the Paul Scherrer Institut (Switzerland). This project aims to optimize imaging of bony and soft tissue of the middle and inner ear anatomy and characterize human cochlear micro-anatomy, by integrating it fully into a multimodal microatlas. This investigation will provide insightful information on morphology of normal cochlea anatomy and its variations before and after CI, permitting the study of cochlear trauma after CI and the creation of a multimodal microatlas of the cochlea.
Virtual Reading of a Large Ancient Handwritten Science Book
![Several examples of tomography reconstructed inner page portions revealing words and sentences (top), compared to visible pictures (bottom) [4].](/sites/default/files/styles/primer_full_xl/public/import/x-ray-tomography-group/ResearchEN/project4_patera.png?itok=49_FfIpy)
Several examples of tomography reconstructed inner page portions revealing words and sentences (top), compared to visible pictures (bottom) [4].
The aim of this project is to implement a new X-ray tomography ”virtual reading” technique in order to read inside a large ancient handwritten book without opening it. The development of this technique is primarily inspired by the Venice Time Machine (VTM) project (EPFL, 2015). This is an ongoing collaboration between the Ecole Polytechnique Fédérale de Lausanne (EPFL) and two institutions in Venice: the University Ca’ Foscari and the ”Archivio di Stato”. The Archivio is an historical collection containing almost 100 kilometers of handwritten documents covering ten centuries of the administrative and legal life of Venice. But, as for all ancient collections, their exploitation by scholars is problematic for conservation and logistic reasons: without massive digitization, deciphering, indexing and storage, they are almost unusable. Within the context of a digitalization, X-ray imaging is used to analyze specimen without opening them [1, 2]. Tomographic reading is feasible thanks to the iron present in ancient inks (iron gall) over one millennium – whereas carbon or organic inks do not provide sufficient x-ray contrast. For this reason, a phase approache is also explored [3, 4].
Heart Imaging Project
3D rendering of a human fetal heart (15 weeks) acquired with X-ray tomography at 20 keV.
Heart failure, as well as sudden cardiac death associated with arrhythmias, are the main causes of mortality, and they are mostly associated with alteration in the muscle tissue (e.g. the presence of important fibrosis or scar). This is why the microscopic organization, as well as its integration in macro anatomy, needs to be completely understood to get a better understanding and develop adequate treatment strategies.
Synchrotron X-ray Phase Contrast micro-tomography, as implemented at the TOMCAT beamline (PSI), provide essential and unique 3D reconstructions of complete hearts (see figure) for understanding cardiac microstructure and its influencing factors, both on small animals or human fetal hearts.
The aim of our current Heart Imaging Project is to quantify cardiac structures and tissues for the first time at micrometre scale, non-destructively, without the use of contrast agents and in 3D. Insights on the cardiac organ morphology like the myofibres, vessels and trabeculation organization, and how they change with disease, are crucial for future clinical applications, not only to understand heart remodelling and the different patterns in cardiovascular disease, but also to provide more personalized and adapted treatments.
In this project, we are working in collaboration with the University Pompeu Fabra (UPF Barcelona), the Hospital Clinic and IDIBAPS (Barcelona), the Institute Cardiovascular Sciences (UCL London) and University Hospital Centre Zagreb.
Cardiac Microstructure Analysis with Phase Contrast Imaging
High resolution orthoslices, segmentation of collagen and manual segmentation of a small group of cardiomyocytes in porcine cardiac tissue.
Although the heart has been studied for centuries, its microscopic organization it is yet not completely understood. The cardiac functionality is influenced by the cardiomyocytes (contractile cells) and other microstructures such as the collagen network, which can be affected by several remodelling cardiac diseases. Synchrotron Radiation based X-Ray Phase Contrast Imaging (PCI) is a non-destructive, three-dimensional, time-efficient and high-resolution technique with the potential to resolve cardiac microstructures and allow the analysis of disease-induced changes. In a first study, 0.65 μm pixel size datasets were acquired and used to segment the collagen network and individual cardiomyocytes for the first time in X-ray images (see Figures).
The aim of this research project is to focus now on the use of free-propagation phase contrast imaging at high resolution on rat biopsies. The different cardiac microstructures, such as the collagen network, individual cardiomyocytes and microvasculature, will be segmented to quantify the cardiac remodelling at cellular level induced by selected cardiac diseases, such as hypertension. For this purpose, existing software for segmentation and 3D rendering such as Avizo or Ilastic are used, but self-developed tools are also necessary.
The introduced study is enclosed in a large project involving several academic and clinical institutions, where the anatomy, myofibre structure and vasculature of the heart are assessed in order to better understand the cardiac changes produced in several cardiovascular diseases.
The aim of this research project is to focus now on the use of free-propagation phase contrast imaging at high resolution on rat biopsies. The different cardiac microstructures, such as the collagen network, individual cardiomyocytes and microvasculature, will be segmented to quantify the cardiac remodelling at cellular level induced by selected cardiac diseases, such as hypertension. For this purpose, existing software for segmentation and 3D rendering such as Avizo or Ilastic are used, but self-developed tools are also necessary.
The introduced study is enclosed in a large project involving several academic and clinical institutions, where the anatomy, myofibre structure and vasculature of the heart are assessed in order to better understand the cardiac changes produced in several cardiovascular diseases.
Full-Field Transmission X-ray Microscopy using a Photon Counting Pixel Detector
X-ray image of Star pattern made of gold acquired using the full-field transmission X-ray microscope with the Merlin detector (515 x 515 pixels of 55 microns). The photon energy was set to 7.5 keV to achieve an X-ray magnification of roughly 725x and effective pixel size of 76 nm.
Full-field transmission X-ray microscopy (TXM) can be used to investigate many materials and systems in biology, material science and energy sciences [1—3]. On the one hand, the performance of the TXM is determined by the quality of the X-ray optics. In particular, the outermost zone width of a Fresnel zone plate (FZP) used as objective lens determines the spatial resolution and its diffraction efficiency usually limits the speed of image acquisition. On the other hand, the spatially resolving detector used to acquire the magnified image of the sample is also a key element of the TXM system. To date, they usually consist of combination of a scintillating material and a visible, high numerical aperture microscope to collect the fluorescent visible light produced by the impinging X-ray photons. From the point of view of photon detection efficiency, low noise and high dynamic range, the use of single photon counting pixel detectors would be highly convenient. However, due to their large pixel sizes (> 55 μm) and the only moderate X-ray magnification (100x—300x) of the current TXM systems, the use of single photon counting pixel detectors has not been broadly considered. Within this project, we aim to combine the optimum diffractive X-ray optics with the long experimental hutches available to achieve large X-ray magnifications and make use of the advantages of photon counting pixel detectors.
The hard x-ray phase contrast full-field nanoscope
Gold Siemens Star radiography obtained at 12 keV with an effective pixel size of 60 nm. The innermost spokes defects are clearly visible.
At modern third generation synchrotron sources, voxel sizes in the micrometer range are routinely achieved. However, isotropic 100 nm barrier is reached and surpassed by only a few instruments. At the TOMCAT beamline of the Swiss Light Source, the multimodal endstation (which offers tomographic capabilities in the micron range) is equipped with a full field, hard X-ray nanoscope.
This full-field transmission X-ray microscope (TXM) is composed of custom optical components. A condenser (a custom designed beamshaper) produces a top-flat illumination in the focal plane. After the sample, the Fresnel Zone Plate (FZP), with a focal length chosen according to the energy, acts as an objective lens. To exploit phase contrast, phase rings can be inserted in the back-focal distance of the FZP to generate positive or negative Zernike phase contrast. The detector, placed downstream at about 10 m, records the magnified image of the sample.
The latest improvements in optical and detector designs as well as stability of the hardware lead to a field-of-view from 50 to 80 μm2 with a pixel size down to 60nm in the 8-20 keV energy range.
In the case of the broadband radiation use, tomographic investigations can be performed within few minutes. The setup can be used in absorption or phase contrast mode for a wide range of applications, including biology, geology, materials science and paleontology.
The ongoing efforts aim to further improve the hard components (optics and setup versatility) as well as data acquisition speed, spatial resolution and sensitivity.
This full-field transmission X-ray microscope (TXM) is composed of custom optical components. A condenser (a custom designed beamshaper) produces a top-flat illumination in the focal plane. After the sample, the Fresnel Zone Plate (FZP), with a focal length chosen according to the energy, acts as an objective lens. To exploit phase contrast, phase rings can be inserted in the back-focal distance of the FZP to generate positive or negative Zernike phase contrast. The detector, placed downstream at about 10 m, records the magnified image of the sample.
The latest improvements in optical and detector designs as well as stability of the hardware lead to a field-of-view from 50 to 80 μm2 with a pixel size down to 60nm in the 8-20 keV energy range.
In the case of the broadband radiation use, tomographic investigations can be performed within few minutes. The setup can be used in absorption or phase contrast mode for a wide range of applications, including biology, geology, materials science and paleontology.
The ongoing efforts aim to further improve the hard components (optics and setup versatility) as well as data acquisition speed, spatial resolution and sensitivity.
Monte Carlo simulation of gratings-based interferometry clinical systems
The aim of this project is to simulate phase contrast imaging (PCI) devices for human sized objects in 3 dimensions. These simulations could be used as a tool for optimization of PCI devices as well as for the investigation of the influence setup errors on the image quality. Previous work by S. Peter [1] showed, using the EGSnrc C++ framework, that modified Monte Carlo techniques are feasible candidates to perform this kind of simulations. The use of Monte Carlo techniques potentially opens many additional opportunities for the investigation of PCI devices, like dose calculations and investigations of the build up of differential phase contrast signals. The codes developed in the aforementioned work serve as a starting point in this project. In a first phase the need to implement new physics in the code has to be investigated, which will be done in both a theoretical as well as computational manner, e.g. by the setup of test scenarios. In a later stage of the project the resulting codes should then be extended from the current purely 2 dimensional simulations into 3 dimensional ones. In addition, the size of the simulated problems has to be increased from a few micro meter sized objects to objects with a size of the order of centimeters. To achieve the significant speedup of the algorithm, which is needed for the increased sample sizes and the additional dimension, several variation reduction techniques will have to be investigated for their feasibility for the simulation of a phase contrast imaging device.
Fast and flexible reconstruction algorithms and workflows for high-throughput tomographic imaging
Over the past few years, at the TOMCAT beamline at the Swiss Light Source a new state-of-the-art endstation devoted to tomographic microscopy with sub-second temporal resolution has been established, with new scientific results being published in different fields. The new GigaFRoST detector can stream the data with rates up to multiple GB/s in a continuous manner and therefore enables experiments not possible so far. To fully exploit these recent technological achievements, the IT infrastructure needs to be matched to these high and sustained data rates. The computational tools necessary for the post-processing of raw tomographic projections have however generally not experienced the same efficiency increase as the experimental facilities, hindering optimal exploitation of this new potential.In the TOMCAT group, we developed a fast, flexible and user-friendly post-processing pipeline overcoming this efficiency mismatch and delivering reconstructed tomographic datasets just few seconds after the data have been acquired, enabling efficient post-processing of TBs of tomographic data. If the acquired datasets are though strongly underconstrained, consisting of only few noisy projections as it is typically the case when the time resolution is pushed to the limit, routines based on iterative algorithms and a priori information are required to satisfactorily reconstruct 3D volumes. In our experience, dedicated procedures for each different problem are though necessary. Taking advantage of the recent development of efficient projectors, we are therefore aiming at a highly flexible parallel iterative reconstruction framework, where the different algorithm components (e.g. projectors, regularizers and solvers) can be easily interchanged and therefore the reconstruction schemes easily optimized for each single experiment. In addition to enabling the reconstruction of a large variety of datasets, this flexible solution will provide invaluable insight and hands-on experience important for defining future directions.
Lossy Compression for CT Datasets
A dataset for tomographic reconstruction as obtained in laboratory setups typically consists of several hundred projective X-ray images of considerable size, quickly amounting to dozens or hundreds of gigabytes per experiment. For different reasons, among them archiving, it is desirable to keep the originals. Their transfer and storage impose a substantial – also financial – burden on the IT infrastructure, making it highly desirable to reduce the size of the files. Unfortunately, previous studies have shown that lossless compression has only limited impact due to hardly compressible image noise. On the other hand, lossy compression has long since become standard for many other imaging applications such as geographic information systems, trading some lost information for better compressibility.
The core idea of lossy compression is to subtly modify an image such that it can be compressed well while keeping the visual appearance as much as possible. For average compression ratios, this will typically lead to tiny ‘ghost’ structures around more prominent image features or smoothing that the human eye will miss. In the case at hand, however, where the images serve as input to the first stage of a longer processing pipeline, such modifications may lead to artefacts in the final tomographic reconstructions. This project aims at evaluating different lossy compression schemes and at investigating the impact of their respective artefacts onto reconstruction quality.
The core idea of lossy compression is to subtly modify an image such that it can be compressed well while keeping the visual appearance as much as possible. For average compression ratios, this will typically lead to tiny ‘ghost’ structures around more prominent image features or smoothing that the human eye will miss. In the case at hand, however, where the images serve as input to the first stage of a longer processing pipeline, such modifications may lead to artefacts in the final tomographic reconstructions. This project aims at evaluating different lossy compression schemes and at investigating the impact of their respective artefacts onto reconstruction quality.
Algorithms for sub-second X-ray tomographic microscopy of liquid water dynamics in polymer electrolyte fuel cells
Data courtesy: J. Eller, PSI.
Fossil fuels are the foundation of our global energy system. Their expected reduced availability and effect on climate challenge mankind to seek for alternative energy sources. Polymer Electrolyte Fuel Cells (PEFC) are a key technology in future decarbonized energy systems, but improvements in efficiency, performance, durability and cost are still needed. Sub-optimal water management is a major limiting factor for increasing the power density of PEFC.
In this project, aim is to develop sub-second tomographic microscopy for PEFC in order to both visualize and quantify the liquid water dynamics in the gas diffusion layers (GDL), the key component regulating water management. Tomographic challenges arise from the sensitivity of PEFC to X-ray radiation and, limited signal-to-noise ratio at the short scanning times required for the investigation of PEFC during transient operation. Development of advanced imaging procedures is required. The efficiency and quality of the acquisition setup at the TOMCAT beamline at the Swiss Light Source will be significantly improved with the installation of a new microscope with a high numerical aperture. Advanced phase retrieval algorithms and iterative schemes with time regularization will be developed to boost tomographic microscopy for these strongly undersampled and noisy datasets.
In this project, aim is to develop sub-second tomographic microscopy for PEFC in order to both visualize and quantify the liquid water dynamics in the gas diffusion layers (GDL), the key component regulating water management. Tomographic challenges arise from the sensitivity of PEFC to X-ray radiation and, limited signal-to-noise ratio at the short scanning times required for the investigation of PEFC during transient operation. Development of advanced imaging procedures is required. The efficiency and quality of the acquisition setup at the TOMCAT beamline at the Swiss Light Source will be significantly improved with the installation of a new microscope with a high numerical aperture. Advanced phase retrieval algorithms and iterative schemes with time regularization will be developed to boost tomographic microscopy for these strongly undersampled and noisy datasets.
Laboratory imaging research
Dual Phase Grating Interferometer
Sketch of experimental setup.
This project explores the possibility of achieving applicable grating interferometer (GI) for table-top applications using only phase-shift gratings. The motivation is to remove the necessity of absorption gratings in conventional GI, which reduces the dose and flux efficient, and in the mean time bypass the fabrication difficulties for manufacturing high quality absorption gratings. The proposed interferometer consists of two phase gratings of small pitches that are placed close to each other. The whole system can be considered as two interferometers in cascade. This configuration generates a large fringe than can be resolved directly without the need of an analyser grating. Moreover, the proposed method provides a practical and convenient way of changing the dark-field sensitivity compared to conventional GI. By simply tuning the integrating distance in the millimeter range can sense structural information of the sample in different length scale.
Omnidirectional Dark-Field Imaging with Circular Unit Cell Gratings
Sketch of experimental setup.
The dark-field signal in X-ray grating interferometry is highly directional, specifically for anisotropic samples the signal intensity depends on the angle between the grating lines and the sample. In this project a novel interferometric design is developed that allows 2D omnidirectional dark-field sensitivity. This is achieved by using fine pitch circular gratings that are repeated to cover the full field of view. The interference fringe generated by the dedicated grating is acquired and resolved with sufficient resolution, and then analysed. The gratings can be arranged in different configurations to optimize between the sensitivity and the final image resolution. The method has been validated [1] and benchmarked [2] at the TOMCAT beamline. On-going developments include the optimization of the analysis framework to increase directional sensitivity, translation of the imaging technique to a table-top setup, and ultimately the utilization of the method in tensor tomography experiments.
Single Shot Differential Phase Contrast Imaging with Single Photon Sensitive Detectors
Sketch of experimental setup.
This project focuses on the development of single shot grating based differential phase contrast imaging methods that do not require an analyser (absorption) grating. The removal of the analyser grating (known as G2) is favourable from both fabrication and photon utilization point of views. The proposed ‘G2-less’ imaging is achieved by utilizing a pixel interpolation method that allows resolution enhancement, which in turn is used to record interference fringe with very fine pitch directly. The resolution enhancement is achieved by exploiting the charge sharing that takes place in direct conversion detectors with small pixel sizes (25 micrometres). By exposing in a single photon sensitive regime the location of the incoming photon can be estimated with a precision higher than that dictated by the pixel size. The recorded fringe is then analysed by a custom phase retrieval algorithm based on the Hilbert transform. We have already managed to successfully demonstrate the applicability of the method with both strip (GOTTHARD) [1] and pixel (MOENCH) [2] detectors. Further developments, include the optimization of the imaging conditions and experiments with larger detector modules.
Design and construction of an X-ray phase-contrast mammography prototype for the in-vivo investigation of early breast cancer
Left: GI on the Philips Microdose Mammography Setup. Right: Gold grating with p2=1.3 μm, a height of 16 μm and a duty cycle of 0.5.
This project consists of the development of a grating-based phase contrast mammography prototype for the in-vivo investigation of breast cancer. Clinically, we aim at improving the diagnostic power of mammography by exploiting the additional information provided by differential phase and dark-field signals. To this end, our approach is to design and build a grating interferometer (GI) that can be fitted into a Philips Microdose Mammography setup, which already fulfills the requirements of a clinical setting [1, 2]. To define the parameters of this GI, we developed an optimization method based on maximizing the sensitivity to phase and dark-field changes taking into account the geometric and grating fabrication constraints. The phase sensitivity was defined as the minimum detectable electron density gradient, whereas the dark-field sensitivity was expressed as the corresponding signal-to-noise ratio (SNR) [3]. In addition, we have investigated alternatives to retrieve and reconstruct the new additional signals that can be compatible with the Philips setup acquisition mode [4].
Differential phase contrast for X-ray tubes above 100 kVp
A 1D edge-on grating interferometer setup at PSI East.
Differential phase contrast and dark-field X-ray imaging have been developed over the last fifteen years. The applications have been extended from synchrotron sources to table-top systems with a Talbot-Lau geometry. Challenges in the fabrication of the optical components limited the deployment to sources typically used in mammography and cartilage screening, with an acceleration voltage below 40 kV. Applying these techniques to general purpose medical investigations or material analysis requires a re-evaluation of the performance and feasibility of grating interferometers on table-top sources, including a quantitative analysis of the response of the differential phase and dark-field signals related to the electron density and microstructural features of the sample.
Ex-vivo study of suspicious microcalcifications in breast tissue biopsies
Absorption and scattering signals of microcalcification in breast biopsy tissue.
Breast cancer is the second most frequently diagnosed cancer in the world and it is the first leading cause of cancer-related deaths in women in less developed regions. The identification of (clustered) microcalcifications plays an important role in early detection of pre-malignant and malignant lesions. Microcalcifications comprise tiny calcium deposits possibly located in areas of accelerated cell turnover, suggestive of precancerous changes or early invasive breast cancer. The hypothesis of this project is that the chemical composition of microcalcifications as well as their internal microstructure are associated with the microenvironment in which the microcalcifications are formed. By probing these information, a descriptor could be established with the goal to improve breast cancer diagnosis. The project aims to understand the structural features of microcalcifications using X-ray small-angle scattering (SAXS) and ptychography, and eventually investigate the potential of using grating interferometer to classify ex-vivo microcalcifications using their absorption and dark-field (scattering) signals.
X-ray Phase Contrast Microtomography for Improved Pathology
Reconstructed slice of X-ray phase contrast microtomography from a human breast biopsy tissue, suspected of being affected by tumour development.
Examination and visualization of biological tissues in Pathology is an essential diagnosis method in medical practice. Currently used histological techniques rely on chemical fixation, staining, very thin sectioning and subsequent visible microscopy inspection of the tissue specimens extracted from the patient’s body. In this project, we aim to investigate and promote the use of X-ray phase contrast microtomography as a complementary method for histopathological techniques. Exploiting the higher sensitivity of X-ray phase contrast is particularly suited for biological soft tissues, for which ordinary X-ray absorption does not provide enough image contrast. The proposed approach may substantially reduce or remove completely the staining requirements of the specimens, thus allowing their examination in conditions closer to their natural state. In addition, X-ray microtomography does not destroy the tissue as the sectioning is virtually done a posteriori on its three-dimensional tomographic image reconstruction. A new laboratory X-ray grating interferometry setup is being assembled using novel state-of-art X-ray source and detector with expected fields of view up to 2.0 cm and high spatial resolutions (<10 µm). The new system will be optimized using actual specimens from the Kantonsspital Baden (Baden, Switzerland). The three-dimensional information obtained from the X-ray phase contrast microtomo-graphy will be later compared the images obtained by conventional histological techniques. If successful, the proposed X-ray phase contrast system could become a disruptive new tool for Pathology.
Feasibility study of an X-ray phase contrast breast CT scanner
Sketch of a grating interferometer based breast CT scanner.
The aim of this project is to assess the feasibility of a grating interferometry (GI) based phase contrast breast CT system. While GI is an established tool for phase contrast X-ray CT imaging and is used for a wide range of applications, its transfer to clinically relevant systems in not obvious. Such a scanner needs to comply with strict dose requirements, while retaining a high level of sensitivity and accuracy. This requires operating the interferometer at higher X-ray energies, which poses challenges to all hardware components, in particular to high energy gratings and detectors. Additionally, the physical model behind high energy GI is not yet fully understood, further complicating any performance assessments. Implementing a grating interferometer on a fast rotating gantry demands high stability and consequently fast and robust signal retrieval methods.
Thus, a general framework has to be developed that considers all mechanical, physical and hardware based limitations. This includes a performance analysis of signal retrieval methods and a comparison of absorption and phase contrast based imaging, as well as an analytical and simulation based understanding of the image formation process and dose deposition. From the hardware point of view the feasibility of a highly stable CT scanner is assessed, as well as the resulting requirements on and capabilities of dedicated hardware components.
Micro and Nano Fabrication
Fabrication of gratings for phase contrast X-ray imaging
Examples of gratings
Grating interferometry is proved to be one of the most promising techniques for phase contrast X-ray imaging. Typical grating interferometer consists of a phase shifting grating (G1), analyzing grating (G2) and an optional absorbing source grating (G0). Usually, required grating period is in a range of few microns. The height of the lines of G1 grating should provide a certain phase shift, while the height of the grating lines of G2 should be sufficient to suppress radiation of defined energy. In both cases, structures with heights of tens (or even hundreds) of micrometers are required. However, the realization of structures with so high aspect ratios yet having sufficient quality over the large area is demanding. We develop fabrication procedures which enable such gratings. We produce G1 gratings in Si by reactive ion etching using Bosch technique [1] or by metal assisted chemical etching [2]. The absorbing G0 and G2 gratings are produced by filling Si templates with metal utilizing electroplating, metal casting [3] or atomic layer deposition [4]. Larger structures can alternatively be produced by laser cutting in W foils.
MAGIC: Metal Assisted chemical etching of Gratings for X-ray InterferometriC systems
MACE grating
Grating fabrication is the main bottleneck so far preventing grating-based X-ray phase-contrast interferometry (GI) from being applied at high energies and large field of view. Metal Assisted Chemical Etching (MACE) is an electroless chemical etching technique that has been largely used to create high aspect ratio nanostructures in silicon substrates. With respect to the other wet etching techniques, MACE showed better performance in terms of anisotropy and feature size. However, MACE still suffers from some limitations such as the control of the catalyst stability for etching high aspect ratio structures in the micro-scale and off-mask undesired porosity. This research project wants to answer the fundamental open questions about MACE in order to fully explore the range of application and the limits of this technique for grating fabrication and to point out the real performances in terms of etching selectivity, rate, aspect ratio, feature size and X-ray optical performances in combination with other proper technique for the realization of the absorber gratings. The main goal of MAGIC is to use MACE for grating fabrication with characteristics that fulfil high energy and large field of view applications. Understanding how MACE works with the proper knowledge about semiconductor physics and metal nanostructures is the key to optimize this process and to make it reliable as a grating fabrication technology.
Fabrication of Silicon Structures for Speckle-based X-ray Phase Contrast Imaging
Phase contrast X-ray imaging is an excellent technique for the investigation of samples in biology, materials science and energy science. In recent years, speckle-based X-ray imaging has been proven as a new approach to obtain the phase and dark-field information of the object under study. This technique uses an optical element (typically, an ordinary sandpaper foil with a fine grain size) to produce a speckled pattern in the X-ray image. The changes in the speckled pattern after introducing the sample can be used to detect the refraction and the scattering signals of the sample. In this work, we investigate the use of a tailored optical element made of silicon structures in the submicrometer range to achieve higher spatial resolution and address other challenges of the method such as the non-uniformity of the phase sensitivity over the whole field of view of the X-ray imaging setup. The silicon nanostructures are created by metal assisted chemical etching (MACE) using gold as catalyst.