Phase Tomography Research Articles

Development of 1-D X-ray beam diffuser for multilayer monochromator and its application to X-ray phase tomography

Development of 1-D X-ray beam diffuser for multilayer monochromator and its application to X-ray phase tomography

Enhanced Detection of Gastrointestinal Malignancies using Machine Learning-Optimized Liquid Biopsy: A Mini Review.

Gastrointestinal (GI) cancers represent some of the most common and lethal malignancies globally, underscoring the urgent need for improved diagnostic strategies. Traditional diagnostic methods, while effective to some degree, are often invasive and unsuit-able for regular screenings. This review article explores integrating machine learning (ML) with liquid biopsy techniques as a revolutionary approach to enhance the detection and monitoring of GI cancers. Liquid biopsies offer a non-invasive alternative for cancer detection through the analysis of circulating tumor DNA (ctDNA) and other biomarkers, which when combined with ML, can significantly improve diagnostic accuracy and patient outcomes. We conducted a comprehensive review of recent advancements in liquid biopsy and ML, focusing on their synergistic potential in the early detection of GI cancers. The review addresses the application of next-generation sequencing and digital droplet PCR in enhancing the sensitivity and specificity of liquid biopsies. Machine learning algorithms have demonstrated remarkable ability in navigating complex datasets and identifying diagnostically significant patterns in ctDNA and other circu-lating biomarkers. Innovations such as machine learning-enhanced "fragmentomics" and tomographic phase imaging flow cytometry illustrate significant strides in non-invasive cancer diagnostics, offering enhanced detection capabilities with high accuracy. The integration of ML in liquid biopsy represents a transformative step in the early detection and personalized treatment of GI cancers. Future research should focus on overcoming current limitations, such as the heterogeneity of tumor-derived genetic materials and the standardization of liquid biopsy protocols, to fully realize the potential of this technol-ogy in clinical settings.

Density of fresh wall of acute aortic dissection with synchrotron-based x-ray phase tomography.

The mechanisms behind the onset of acute aortic dissection have not been fully elucidated. We developed dynamic Synchrotron-based X-ray phase tomography to quantitatively study the dynamics of biological samples and applied it to the fresh aortic wall in acute type-A aortic dissection. Fresh, ring-shaped aortas undergoing aortic repair in acute type-A aortic dissection were measured in a container filled with normal cold saline within 24 hours of surgery. As a control, we obtained five formalin-fixed normal ascending aortas from autopsies (female : 2, 59.7 (SD : 5.5) years). To evaluate the quantitative morphological change, we estimated the density at five each step stretched by 2 mm per step. The fresh specimens were analyzed pathologically about the area ratio of elastic fibre. Samples were obtained from five patients (1 man and 4 women, 59.4 (SD: 8.7) years) The overall density of the tunica media in the fresh aorta was 1.062(SD : 0.006) g/cm3 and differed significantly between the dissected and non-dissected portion (1.05(SD : 0.004) vs 1.066(SD : 0.004) g/cm3, respectively; p = 0.0122). When the fresh aortic wall was stretched and became thinner, the density of the tunica media remained unchanged. Compared with pathological findings, area ratio of the elastic fibre of the tunica media were lower in non-dissected portion than normal (48.6 (SD : 7.1) % v.s 60.5 (SD : 5.7) %, p < 0.001). Dynamic-XPCT can trace the deformation process that occurs in situ in fresh aorta in acute type-A aortic dissection. We confirmed that densitometric property of the aortic wall in acute type-A aortic dissection was unchanged during the stretching process.

Phase space reconstruction technique for beam optimization in the Low Energy High Intensity Proton Accelerator (LEHIPA)

The Low Energy High Intensity Proton Accelerator (LEHIPA) has been commissioned to the design energy of 20 MeV at BARC, India. The low energy beam transport (LEBT) channel of LEHIPA consists of two solenoids and drift spaces for matching the 50 keV proton beam from the ECR ion source (ECR-IS) to the RFQ. The ion beam extracted from the ECR-IS also contains molecular species like H2 + and H3 +. Proton fraction in the beam is found to degrade slowly with time due to surface contamination of the plasma chamber and this reduction in proton beam current has implications for longitudinal and transverse beam dynamics. Hence, it becomes important to characterise the beam at different proton fraction levels to understand the end-to-end beam dynamics of LEHIPA. Computed Tomography (CT) technique has been used for the beam phase-space reconstruction in LEHIPA, using a Python program, incorporating the feature of filtering secondary species from the beam profiles measured using slit scanners in the LEBT. Simultaneous Algebraic Reconstruction Technique (SART) is used in the reconstruction since it is identified to be a better technique for a limited number of projections. The reconstruction program is benchmarked with the TraceWin beam dynamics code and implemented on the measured beam profiles to recreate the phase space distribution at the beginning of LEBT. Further, the tomographic reconstruction method is compared with the solenoid scan method and the rms emittance values are found to be in good agreement. The measured tomographic phase space distribution has then been used as TraceWin input for LEHIPA end-end beam dynamics simulations and the LEHIPA beam line parameters are re-optimized for minimum beam emittance growth. This paper presents the simulations of CT technique, benchmarking simulations with TraceWin, phase space reconstruction with measured beam profiles and beam dynamics studies of LEHIPA using the reconstructed beam distributions.

Label-free morpho-molecular phenotyping of living cancer cells by combined Raman spectroscopy and phase tomography

Accurate, rapid and non-invasive cancer cell phenotyping is a pressing concern across the life sciences, as standard immuno-chemical imaging and omics require extended sample manipulation. Here we combine Raman micro-spectroscopy and phase tomography to achieve label-free morpho-molecular profiling of human colon cancer cells, following the adenoma, carcinoma, and metastasis disease progression, in living and unperturbed conditions. We describe how to decode and interpret quantitative chemical and co-registered morphological cell traits from Raman fingerprint spectra and refractive index tomograms. Our multimodal imaging strategy rapidly distinguishes cancer phenotypes, limiting observations to a low number of pristine cells in culture. This synergistic dataset allows us to study independent or correlated information in spectral and tomographic maps, and how it benefits cell type inference. This method is a valuable asset in biomedical research, particularly when biological material is in short supply, and it holds the potential for non-invasive monitoring of cancer progression in living organisms.

Optimization of contrast and dose in x-ray phase-contrast tomography with a Talbot-Lau interferometer

X-ray phase-contrast imaging has become a valuable tool for biomedical research due to its improved contrast abilities over regular attenuation-based imaging. The recently emerged Talbot-Lau interferometer can provide quantitative attenuation, phase-contrast and dark-field image data, even with low-brilliance x-ray tube sources. Thus, it has become a valid option for clinical environments. In this study, we analyze the effects of x-ray tube voltage and total number of images on the contrast-to-noise ratio (CNR) and dose-weighted CNR (CNRD) calculated from tomographic transmission and phase-contrast data of a phantom sample. Constant counting statistics regardless of the voltage was ensured by adjusting the image exposure time for each voltage setting. The results indicate that the x-ray tube voltage has a clear effect on both image contrast and noise. This effect is amplified in the case of phase-contrast images, which is explained by the polychromatic x-ray spectrum and the dependence of interferometer visibility on the spectrum. CNRD is additionally affected by the total imaging time. While submerging the sample into a water container effectively reduces image artefacts and improves the CNR, the additional attenuation of the water must be compensated with a longer exposure time. This reduces dose efficiency. Both the CNR and CNRD are higher in the phase-contrast images compared to transmission images. For transmission images, and phase-contrast images without the water container, CNRD can be increased by using higher tube voltages (in combination with a lower exposure time). For phase-contrast images with the water container, CNRD is increased with lower tube voltages. In general, the CNRD does not strongly depend on the number of tomographic angles or phase steps used.

Deep learning for accelerating Radon inversion in single-cells tomographic phase imaging flow cytometry

Tomographic phase microscopy (TPM) in flow cytometry is an emerging imaging technology completely stain-free and able to provide unique 3D biophysical data. However, the required throughput for TPM would correspond to a very huge amount of volumetric data to be processed. To date, the most accurate Radon inversion algorithms for tomographic reconstructions are based on iterative methods aided by regularizations, thus providing high performance but at the cost of a demanding computation time. To balance the trade-off between reconstruction performance and reconstruction speed, learning-based approaches have been successfully introduced, thus demonstrating high accuracy in solving the Radon inversion problem. However, the complexity of the proposed models remains a bottleneck due to the high computational resources still required for the training phase. Here we show that, employing a multi-scale fully convolutional Context Aggregation Network (CAN) model, a significant speeding-up of the Radon inversion computation can be achieved. Compared to other conventional encoder-decoder networks such as U-Net, the proposed method has proven to be accurate, faster and particularly suitable for on-board processing. Moreover, we show the generalization capacity of CAN by training the network with simulated tomographic data and testing the learned model on experimental tomographic data.

Label-free liquid biopsy through the identification of tumor cells by machine learning-powered tomographic phase imaging flow cytometry

Image-based identification of circulating tumor cells in microfluidic cytometry condition is one of the most challenging perspectives in the Liquid Biopsy scenario. Here we show a machine learning-powered tomographic phase imaging flow cytometry system capable to provide high-throughput 3D phase-contrast tomograms of each single cell. In fact, we show that discrimination of tumor cells against white blood cells is potentially achievable with the aid of artificial intelligence in a label-free flow-cyto-tomography method. We propose a hierarchical machine learning decision-maker, working on a set of features calculated from the 3D tomograms of the cells’ refractive index. We prove that 3D morphological features are adequately distinctive to identify tumor cells versus the white blood cell background in the first stage and, moreover, in recognizing the tumor type at the second decision step. Proof-of-concept experiments are shown, in which two different tumor cell lines, namely neuroblastoma cancer cells and ovarian cancer cells, are used against monocytes. The reported results allow claiming the identification of tumor cells with a success rate higher than 97% and with an accuracy over 97% in discriminating between the two cancer cell types, thus opening in a near future the route to a new Liquid Biopsy tool for detecting and classifying circulating tumor cells in blood by stain-free method.

Partially coherent broadband 3D optical transfer functions with arbitrary temporal and angular power spectra.

Optical diffraction tomography is a powerful technique to produce 3D volumetric images of biological samples using contrast produced by variations in the index of refraction in an unlabeled specimen. While this is typically performed with coherent illumination from a variety of angles, interest has grown in partially coherent methods due to the simplicity of the illumination and the computation-free axial sectioning provided by the coherence window of the source. However, such methods rely on the symmetry or discretization of a source to facilitate quantitative analysis and are unable to efficiently handle arbitrary illumination that may vary asymmetrically in angle and continuously in the spectrum, such as diffusely scattered or thermal sources. A general broadband theory may expand the scope of illumination methods available for quantitative analysis, as partially coherent sources are commonly available and may benefit from the effects of spatial and temporal incoherence. In this work, we investigate partially coherent tomographic phase microscopy from arbitrary sources regardless of angular distribution and spectrum by unifying the effects of spatial and temporal coherence into a single formulation. This approach further yields a method for efficient computation of the overall systems' optical transfer function, which scales with O(n 3), down from O(mn 4) for existing convolutional methods, where n 3 is the number of spatial voxels in 3D space and m is the number of discrete wavelengths in the illumination spectrum. This work has important implications for enabling partially coherent 3D quantitative phase microscopy and refractive index tomography in virtually any transmission or epi-illumination microscope.

Organomorphic Silicon Carbide Reinforcing Preform Formation Mechanism

Development of the organomorphic ceramic-matrix composites (CMCs), where the reinforcing preform is built using polymer fibers subject essentially to hot pressing, was motivated by a desire to obtain much higher structural uniformity as well as to reduce the number of the process steps involved in the production of CMCs. This paper addresses the peculiarities of the organomorphic silicon carbide preform formation process. Using X-ray phase analysis, tomography, mass and IR spectroscopy, and thermomechanical and X-ray microanalysis, both the properties of the initial fibers of polycarbosilane (PCS)—the silicon carbide fiber precursor—and their transformation in the preform while heated to 1250 °C under constant pressing at 10–100 kPa were studied. Analysis of the data obtained showed the organomorphic SiC preform relative density at a level of 0.3–0.4 to be ensured by self-bonding of the silicon carbide preform, resulting from the fact that during the low-temperature part of pyrolysis, easily polymerizing substances are released leaving a high coke residue, thus cementing the preform. Another possible factor of SiC framework self-bonding is the destruction of the polymer fibers during pyrolysis of various PCS preforms differing in their methylsilane composition (for example, dimethylsilane), where deposition of silicon carbide on the contacting fibers starts as early as at 450–500 °C.

Towards true volumetric refractive index investigation in tomographic phase microscopy at the cellular level

Tomographic phase microscopy (TPM) is one of the most widely used quantitative phase imaging techniques. It is a non-invasive, label-free technique that allows high-resolution imaging. It enables the morphology of a living cell to be captured without chemical treatment, by using the refractive index (RI) as a contrast parameter to image internal structures. In this paper, we demonstrate that due to the fact that biological cells are highly heterogeneous structures, it is crucial to use full volumetric data to calculate the average RI values of biological samples and their organelles. To prove our point, we present tomographic reconstructions of 3 cell types: neuroblastoma SH-SY5Y, adenocarcinoma A549 and leukemia HL-60, from which we calculate the average RI value from the 3D volumetric data and compare it to the average RI values calculated from each 2D section.

Loss Minimized Data Reduction in Single-Cell Tomographic Phase Microscopy Using 3D Zernike Descriptors

Tomographic phase microscopy (TPM) in flow cytometry is one of the most promising computational imaging techniques for the quantitative 3-dimensional (3D) analysis of unstained single cells. Continuous cells’ flow, combined with the stain-free mode, can assure the high-throughput collection of quantitative and informative 3D data. TPM promises to allow rapid cells’ screening by a nondestructive technique and with statistically relevant data. The current leading-edge research aimed at developing TPM systems in flow cytometry has already demonstrated the possibility of acquiring thousands of single-cell tomograms. Nevertheless, a key unsolved problem exists about the efficient storage and easy handling of such a huge amount of 3D data that prevents rapid analysis for cell diagnosis. Here, we show, for the first time, an effective encoding strategy of single-cell tomograms that can completely overcome this critical bottleneck. Essentially, by using the 3D version of Zernike polynomials, we demonstrate that the 3D refractive index distribution of a cell can be straightforwardly encoded in 1D with negligible information loss (&lt;1%), thus greatly streamlining the data handling and storage. The performance analysis of the proposed method has been first assessed on simulated tomographic cell phantom, while the experimental validation has been extensively proofed on tomographic data from experiments with different cell lines. The results achieved here imply an intriguing breakthrough for TPM that promises to unlock computational pipelines for analyzing 3D data that were unattainable until now.

An Image Construction Based on the VTLF Fusion Technique for Oil–Water Two-Phase Flow of Noncontact Electrical Impedance Tomography

Electrical tomography (ET) has attracted extensive attention in the petroleum multiphase flow field due to its advantages of nonintrusion, nonradiation, and simple equipment structure. Some contactless impedance imaging methods, such as capacitively coupled electrical resistance tomography (CCERT) and phase-based dielectric spectroscopy tomography, or displacement current phase tomography (DCPT), have been studied for the parameter measurement of two-phase flow. The mentioned research mainly focuses on the separate imaging by real part, imaginary part, or phase information of the voltage/current. However, the distribution change of media in the region of interest (RoI), especially the presence of high permittivity and conductivity materials, involves the distribution change of conductivity, permittivity, and loss factor. Developing a proper image fusion technique based on the three components can provide more comprehensive image information for tomographic imaging. This article proposed an improved logic filter image fusion algorithm based on a variable threshold (VTLF), and a noncontact electrical impedance tomography (NEIT) system for imaging water–oil two-phase flow is carried out. Image reconstruction is implemented by the VTLF image fusion algorithm under different media distributions, conductivities, and excitation frequencies, and the image fusion algorithms of a weighted average (WA) and logic filter (LF) are also used for comparison. The simulation results show that the reconstructed images by the proposed fusion method of VTLF had the lowest artifacts and the highest correlation coefficient (above 0.7) compared with other methods. Also, in the experiment, the VTLF also got the best performance, so it is effective to improve the reconstruction accuracy.

3D scattering microphantom sample to assess quantitative accuracy in tomographic phase microscopy techniques

In this paper we present a structurally-complex biomimetic scattering structure, fabricated with two-photon polymerization, and utilize this object in order to benchmark a computational imaging system. The phantom allows to tailor the scattering by modifying its degrees of freedom i.e. refractive index contrast and scattering layer dimensions and incorporates a 3D imaging quality test, representing a single cell within tissue. While the sample may be used with multiple 3D microscopy techniques, we demonstrate the impact of scattering on three tomographic phase microscopy (TPM) reconstruction methods. One of these methods assumes the sample to be weak-scattering, while the other two take multiple scattering into account. The study is performed at two wavelengths (visible and near-infrared), which serve as a scaling factor for the scattering phenomenon. We find that changing the wavelength from visible into near-infrared impacts the applicability of TPM reconstruction methods. As a result of reduced scattering in near-infrared region, the multiple-scattering-oriented techniques perform in fact worse than a method aimed for weak-scattering samples. This implies a necessity of selecting proper approach depending on sample’s scattering characteristics even in case of subtle changes in the object-light interaction.

Stain-free identification of cell nuclei using tomographic phase microscopy in flow cytometry

Quantitative phase imaging has gained popularity in bioimaging because it can avoid the need for cell staining, which, in some cases, is difficult or impossible. However, as a result, quantitative phase imaging does not provide the labelling of various specific intracellular structures. Here we show a novel computational segmentation method based on statistical inference that makes it possible for quantitative phase imaging techniques to identify the cell nucleus. We demonstrate the approach with refractive index tomograms of stain-free cells reconstructed using tomographic phase microscopy in the flow cytometry mode. In particular, by means of numerical simulations and two cancer cell lines, we demonstrate that the nucleus can be accurately distinguished within the stain-free tomograms. We show that our experimental results are consistent with confocal fluorescence microscopy data and microfluidic cyto-fluorimeter outputs. This is a remarkable step towards directly extracting specific three-dimensional intracellular structures from the phase contrast data in a typical flow cytometry configuration.

Identification of drug-resistant cancer cells in flow cytometry combining 3D holographic tomography with machine learning

Identifying drug-resistant cancer cells is of fundamental importance to afford disease and find the most effective therapies for the patients. Recently, label-free imaging flow cytometry has been deeply investigated in cell recognition. In particular, the combination of flow cytometry and machine learning allows for achieving high accuracy in cell identification and high throughput. Despite the encouraging results, the potentialities of digital holography (DH) in flow-cytometry modality have not been exploited in full. Up to now, only 2D phase maps have been used in all previously reported research about the use of DH for analyzing flowing cells. Here we show that having access to the whole 3D information of each flowing cell can improve the cells identification. We used digital holographic flow cytometry to collect images of flowing cells and reconstructed their 3D tomographic phase. And for the first time, we extracted scores of meaningful morphometric features from the 3D and 2D phase maps through machine learning methods and finally compare their classification performance. The results show that 3D features can achieve higher classification accuracy with respect to sole 2D analysis demonstrating that 3D morphology information can yield advantages in recognizing drug-resistant endometrial cancer cells, thus allowing a significant step forward in performance of label-free cell classification.

Numerical refractive index correction for the stitching procedure in tomographic quantitative phase imaging.

Tomographic quantitative phase imaging (QPI) lacks an absolute refractive index value baseline, which poses a problem when large dense objects extending over multiple fields of view are measured volume by volume and stitched together. Some of the measurements lack the natural baseline value that is provided by the mounting medium with a known refractive index. In this work, we discuss the problem of the refractive index (RI) baseline of individual reconstructed volumes that are deprived of access to mounting medium due to the extent of the object. The solution of this problem is provided by establishing the RI offsets based on the overlapping regions. We have proven that the process of finding the offset RI values may be justifiably reduced to the analogous procedure in the 2D baseline correction (2D-BC). Finally, we proposed the enhancement of the state-of-the-art 2D-BC procedure previously introduced in the context of 2D QPI. The processing is validated at the examples of a synthetic dataset and a liver organoid.

Finding intracellular lipid droplets from the single-cell biolens' signature in a holographic flow-cytometry assay.

In recent years, intracellular LDs have been discovered to play an important role in several pathologies. Therefore, detection of LDs would provide an in-demand diagnostic tool if coupled with flow-cytometry to give significant statistical analysis and especially if the diagnosis is made in full non-invasive mode. Here we combine the experimental results of in-flow tomographic phase microscopy with a suited numerical simulation to demonstrate that intracellular LDs can be easily detected through a label-free approach based on the direct analysis of the 2D quantitative phase maps recorded by a holographic flow cytometer. In fact, we demonstrate that the presence of LDs affects the optical focusing lensing features of the embracing cell, which can be considered a biological lens. The research was conducted on white blood cells (i.e., lymphocytes and monocytes) and ovarian cancer cells. Results show that the biolens properties of cells can be a rapid biomarker that aids in boosting the diagnosis of LDs-related pathologies by means of the holographic flow-cytometry assay for fast, non-destructive, and high-throughput screening of statistically significant number of cells.

Uniqueness theorems for tomographic phase retrieval with few coded diffraction patterns

3D tomographic phase retrieval under the Born approximation for discrete objects supported on a n × n × n grid is analyzed. It is proved that n projections are sufficient and necessary for unique determination by computed tomography with full projected field measurements and that n + 1 coded projected diffraction patterns are sufficient for unique determination, up to a global phase factor, in tomographic phase retrieval. Hence n + 1 is nearly, if not exactly, the minimum number of diffractions patterns needed for 3D tomographic phase retrieval under the Born approximation.

Numerical analysis of the effect of reduced temporal coherence in quantitative phase microscopy and tomography.

We present the numerical analysis of the effect of the temporarily partially coherent illumination on the phase measurement accuracy in digital holography microscopy (DHM) and optical diffraction tomography (ODT), as reconstruction algorithms tend to assume purely monochromatic conditions. In the regime of reduced temporal coherence, we simulate the hologram formation in two different optical setups, representing classical off-axis two-beam and grating common-path configurations. We consider two ODT variants: with sample rotation and angle-scanning of illumination. Besides the coherence degree of illumination, our simulation considers the influence of the sample normal dispersion, shape of the light spectrum, and optical parameters of the imaging setup. As reconstruction algorithms we employ Fourier hologram method and first-order Rytov approximation with direct inversion and nonnegativity constraints. Quantitative evaluation of the measurement results deviations introduced by the mentioned error sources is comprehensively analyzed, for the first time to the best of our knowledge. Obtained outcomes indicate low final DHM/ODT reconstruction errors for the grating-assisted common-path configuration. Nevertheless, dispersion and asymmetric spectrum introduce non-negligible overestimated refractive index values and noise, and should be thus carefully considered within experimental frameworks.