Machine Teaching Allows for Rapid Development of Automated Systems for Retinal Lesion Detection From Small Image Datasets.

BACKGROUND AND OBJECTIVE: Machine teaching, a machine learning subfield, may allow for rapid development of artificial intelligence systems able to automatically identify emerging ocular biomarkers from small imaging datasets. We sought to use machine teaching to automatically identify retinal ischemic perivascular lesions (RIPLs) and subretinal drusenoid deposits (SDDs), two emerging ocular biomarkers of cardiovascular disease. MATERIALS AND METHODS: IRB approval was obtained. Four small datasets of SD-OCT B-scans were used to train and test two distinct automated systems, one identifying RIPLs and the other identifying SDDs. An open-source interactive machine-learning software program, RootPainter, was used to perform annotation and training simultaneously over a 6-hour period. RESULTS: For SDDs at the B-scan level, test-set accuracy = 92%, sensitivity = 100%, specificity = 88%, positive predictive value (PPV) = 82%, and negative predictive value (NPV) = 100%. For RIPLs at the B-scan level, test-set accuracy = 90%, sensitivity = 60%, specificity = 93%, PPV = 50%, and NPV = 95%. CONCLUSION: Machine teaching demonstrates promise within ophthalmic imaging to rapidly allow for automated identification of novel biomarkers from small image datasets. [Ophthalmic Surg Lasers Imaging Retina 2024;55:XX-XX.].

Adaptation to ex vivo culture reduces human hematopoietic stem cell activity independently of cell cycle.

Loss of long-term hematopoietic stem cell (LT-HSC) function ex vivo hampers the success of clinical protocols reliant on culture. However, the kinetics and mechanisms by which this occurs remain incompletely characterized. Here, through time-resolved scRNA-Seq, matched in vivo functional analysis and the use of a reversible in vitro system of early G1 arrest, we define the sequence of transcriptional and functional events occurring during the first ex vivo division of human LT-HSCs. We demonstrate that the sharpest loss of LT-HSC repopulation capacity happens early on, between 6 and 24 hours of culture, before LT-HSCs commit to cell cycle progression. During this time window, LT-HSCs adapt to the culture environment, limiting global variability in gene expression and transiently upregulating gene networks involved in signaling and stress responses. From 24 hours, LT-HSC progression past early G1 contributes to the establishment of differentiation programmes in culture. However, contrary to current assumptions, we demonstrate that loss of HSC function ex vivo is independent of cell cycle progression. Finally, we show that targeting LT-HSC adaptation to culture by inhibiting early activation of JAK/STAT signaling improves HSC long-term repopulating function ex vivo. Collectively, our study demonstrates that controlling early LT-HSC adaptation to ex vivo culture, for example via JAK inhibition, is of critical importance to improve HSC gene therapy and expansion protocols.

Exploring Diagnostic Precision and Triage Proficiency: A Comparative Study of GPT-4 and Bard in Addressing Common Ophthalmic Complaints.

In the modern era, patients often resort to the internet for answers to their health-related concerns, and clinics face challenges to providing timely response to patient concerns. This has led to a need to investigate the capabilities of AI chatbots for ophthalmic diagnosis and triage. In this in silico study, 80 simulated patient complaints in ophthalmology with varying urgency levels and clinical descriptors were entered into both ChatGPT and Bard in a systematic 3-step submission process asking chatbots to triage, diagnose, and evaluate urgency. Three ophthalmologists graded chatbot responses. Chatbots were significantly better at ophthalmic triage than diagnosis (90.0% appropriate triage vs. 48.8% correct leading diagnosis; p < 0.001), and GPT-4 performed better than Bard for appropriate triage recommendations (96.3% vs. 83.8%; p = 0.008), grader satisfaction for patient use (81.3% vs. 55.0%; p < 0.001), and lower potential harm rates (6.3% vs. 20.0%; p = 0.010). More descriptors improved the accuracy of diagnosis for both GPT-4 and Bard. These results indicate that chatbots may not need to recognize the correct diagnosis to provide appropriate ophthalmic triage, and there is a potential utility of these tools in aiding patients or triage staff; however, they are not a replacement for professional ophthalmic evaluation or advice.

Swept-source optical coherence tomography angiography metrics of retinal ischaemic perivascular lesions in patients being evaluated for carotid artery stenosis and controls.

BACKGROUND/AIMS: Retinal microvascular ischaemia may produce localised middle retinal disruption with corresponding scotoma, a phenomenon termed paracentral acute middle maculopathy (PAMM). Small chronic middle retinal atrophic lesions termed retinal ischaemic perivascular lesions (RIPLs) appear qualitatively similar to PAMM lesions and have recently been hypothesised to result specifically from PAMM. However, no studies have quantitatively demonstrated an ischaemic origin of RIPLs. We quantitatively investigated the pathophysiology of RIPLs and their relationship with PAMM with swept-source optical coherence tomography angiography (SS-OCTA). METHODS: A total of 14 controls and 25 patients being evaluated for carotid artery stenosis (CAS) were enrolled. SS-OCTA imaging of each eye was taken. Projection-resolved en face 6 mm × 6 mm superficial capillary plexus (SCP) and deep capillary plexus (DCP) images were quantitatively analysed with two algorithms for changes in vessel linear density (VLD) and vessel tortuosity (VT) at RIPLs relative to both the immediately surrounding macula and the entire macula, as well as between eyes with RIPLs and eyes without RIPLs. RESULTS: All controls and 22 of 25 CAS patients were included in the analysis. RIPLs demonstrated a localised decrease in DCP VLD in CAS patients and controls. RIPLs tended to show a localised decrease in SCP VLD in CAS patients but a localised increase in controls. No changes in VT were found. Eyes with RIPLs had VLD and VT similar to their RIPL-free fellow eyes. CONCLUSION: RIPLs are associated with quantifiable local, but not global, ischaemia, supporting the idea of shared pathophysiology with classic PAMM lesions along a continuum of ischaemia severity.

Flexible positioning of a large area detector using an industrial robot.

The DIAD beamline for Dual Imaging and Diffraction at Diamond Light Source has opted to use an industrial robot to position its Dectris Pilatus 2M CdTe diffraction detector. This setup was chosen to enable flexible positioning of the detector in a quarter-sphere around the sample position whilst reliably holding the large weight of 139 kg of detector, detector mount and cabling in a stable position. Metrology measurements showed that the detector can be positioned with a linear repeatability of <19.7 µm and a rotational repeatability of <16.3 µrad. The detector position stays stable for a 12 h period with <10.1 µm of movement for linear displacement and <3.8 µrad for rotational displacement. X-ray diffraction from calibration samples confirmed that the robot is sufficiently stable to resolve lattice d-spacings within the instrumental broadening given by detector position and beam divergence.

Beamline K11 DIAD: a new instrument for dual imaging and diffraction at Diamond Light Source.

The Dual Imaging and Diffraction (DIAD) beamline at Diamond Light Source is a new dual-beam instrument for full-field imaging/tomography and powder diffraction. This instrument provides the user community with the capability to dynamically image 2D and 3D complex structures and perform phase identification and/or strain mapping using micro-diffraction. The aim is to enable in situ and in operando experiments that require spatially correlated results from both techniques, by providing measurements from the same specimen location quasi-simultaneously. Using an unusual optical layout, DIAD has two independent beams originating from one source that operate in the medium energy range (7-38 keV) and are combined at one sample position. Here, either radiography or tomography can be performed using monochromatic or pink beam, with a 1.4 mm × 1.2 mm field of view and a feature resolution of 1.2 µm. Micro-diffraction is possible with a variable beam size between 13 µm × 4 µm and 50 µm × 50 µm. One key functionality of the beamline is image-guided diffraction, a setup in which the micro-diffraction beam can be scanned over the complete area of the imaging field-of-view. This moving beam setup enables the collection of location-specific information about the phase composition and/or strains at any given position within the image/tomography field of view. The dual beam design allows fast switching between imaging and diffraction mode without the need of complicated and time-consuming mode switches. Real-time selection of areas of interest for diffraction measurements as well as the simultaneous collection of both imaging and diffraction data of (irreversible) in situ and in operando experiments are possible.

Data processing methods and data acquisition for samples larger than the field of view in parallel-beam tomography.

Parallel-beam tomography systems at synchrotron facilities have limited field of view (FOV) determined by the available beam size and detector system coverage. Scanning the full size of samples bigger than the FOV requires various data acquisition schemes such as grid scan, 360-degree scan with offset center-of-rotation (COR), helical scan, or combinations of these schemes. Though straightforward to implement, these scanning techniques have not often been used due to the lack of software and methods to process such types of data in an easy and automated fashion. The ease of use and automation is critical at synchrotron facilities where using visual inspection in data processing steps such as image stitching, COR determination, or helical data conversion is impractical due to the large size of datasets. Here, we provide methods and their implementations in a Python package, named Algotom, for not only processing such data types but also with the highest quality possible. The efficiency and ease of use of these tools can help to extend applications of parallel-beam tomography systems.

Physics of animal health: on the mechano-biology of hoof growth and form.

Global inequalities in economic access and agriculture productivity imply that a large number of developing countries rely on working equids for transport/agriculture/mining. Therefore, the understanding of hoof conditions/shape variations affecting equids' ability to work is still a persistent concern. To bridge this gap, using a multi-scale interdisciplinary approach, we provide a bio-physical model predicting the shape of equids' hooves as a function of physical and biological parameters. In particular, we show (i) where the hoof growth stress originates from, (ii) why the hoof growth rate is one order of magnitude higher than the proliferation rate of epithelial cells and (iii) how the soft-to-hard transformation of the epithelium is possible allowing the hoof to fulfil its function as a weight-bearing element. Finally (iv), we demonstrate that the reason for hoof misshaping is linked to the asymmetrical design of equids' feet (shorter quarters/long toe) together with the inability of the biological growth stress to compensate for such an asymmetry. Consequently, the hoof can adopt a dorsal curvature and become 'dished' overtime, which is a function of the animal's mass and the hoof growth rate. This approach allows us to discuss the potential occurrence of this multifaceted pathology in equids.

High-energy, high-resolution, fly-scan X-ray phase tomography.

High energy X-ray phase contrast tomography is tremendously beneficial to the study of thick and dense materials with poor attenuation contrast. Recently, the X-ray speckle-based imaging technique has attracted widespread interest because multimodal contrast images can now be retrieved simultaneously using an inexpensive wavefront modulator and a less stringent experimental setup. However, it is time-consuming to perform high resolution phase tomography with the conventional step-scan mode because the accumulated time overhead severely limits the speed of data acquisition for each projection. Although phase information can be extracted from a single speckle image, the spatial resolution is deteriorated due to the use of a large correlation window to track the speckle displacement. Here we report a fast data acquisition strategy utilising a fly-scan mode for near field X-ray speckle-based phase tomography. Compared to the existing step-scan scheme, the data acquisition time can be significantly reduced by more than one order of magnitude without compromising spatial resolution. Furthermore, we have extended the proposed speckle-based fly-scan phase tomography into the previously challenging high X-ray energy region (120 keV). This development opens up opportunities for a wide range of applications where exposure time and radiation dose are critical.

Residual Lattice Strain and Phase Distribution in Ti-6Al-4V Produced by Electron Beam Melting.

Residual stress/strain and microstructure used in additively manufactured material are strongly dependent on process parameter combination. With the aim to better understand and correlate process parameters used in electron beam melting (EBM) of Ti-6Al-4V with resulting phase distributions and residual stress/strains, extensive experimental work has been performed. A large number of polycrystalline Ti-6Al-4V specimens were produced with different optimized EBM process parameter combinations. These specimens were post-sequentially studied by using high-energy X-ray and neutron diffraction. In addition, visible light microscopy, scanning electron microscopy (SEM) and electron backscattered diffraction (EBSD) studies were performed and linked to the other findings. Results show that the influence of scan speed and offset focus on resulting residual strain in a fully dense sample was not significant. In contrast to some previous literature, a uniform α- and ß-Ti phase distribution was found in all investigated specimens. Furthermore, no strong strain variations along the build direction with respect to the deposition were found. The magnitude of strain in α and ß phase show some variations both in the build plane and along the build direction, which seemed to correlate with the size of the primary ß grains. However, no relation was found between measured residual strains in α and ß phase. Large primary ß grains and texture appear to have a strong effect on X-ray based stress results with relatively small beam size, therefore it is suggested to use a large beam for representative bulk measurements and also to consider the prior ß grain size in experimental planning, as well as for mathematical modelling.

Revealing internal flow behaviour in arc welding and additive manufacturing of metals.

Internal flow behaviour during melt-pool-based metal manufacturing remains unclear and hinders progression to process optimisation. In this contribution, we present direct time-resolved imaging of melt pool flow dynamics from a high-energy synchrotron radiation experiment. We track internal flow streams during arc welding of steel and measure instantaneous flow velocities ranging from 0.1 m s-1 to 0.5 m s-1. When the temperature-dependent surface tension coefficient is negative, bulk turbulence is the main flow mechanism and the critical velocity for surface turbulence is below the limits identified in previous theoretical studies. When the alloy exhibits a positive temperature-dependent surface tension coefficient, surface turbulence occurs and derisory oxides can be entrapped within the subsequent solid as result of higher flow velocities. The widely used arc welding and the emerging arc additive manufacturing routes can be optimised by controlling internal melt flow through adjusting surface active elements.

Superior techniques for eliminating ring artifacts in X-ray micro-tomography.

Synchrotron-based X-ray micro-tomography systems often suffer severe ring artifacts in reconstructed images. In sinograms the artifacts appear as straight lines or stripe artifacts. These artifacts are caused by the irregular response of a detecting system giving rise to a variety of observed types of stripes: full stripes, partial stripes, fluctuating stripes, and unresponsive stripes. The use of pre-processing techniques such as distortion correction or phase retrieval blurs and enlarges these stripes. It is impossible for a single approach to remove all types of stripe artifacts. Here, we propose three techniques for tackling all of them. The proposed techniques are easy to implement; do not generate extra stripe artifacts and void-center artifacts; and give superior quality on challenging data sets and in comparison with other techniques. Implementations in Python and a challenging data set are available for download.

X-ray phase-contrast imaging with engineered porous materials over 50†keV.

X-ray phase-contrast imaging can substantially enhance image contrast for weakly absorbing samples. The fabrication of dedicated optics remains a major barrier, especially in high-energy regions (i.e. over 50 keV). Here, the authors perform X-ray phase-contrast imaging by using engineered porous materials as random absorption masks, which provides an alternative solution to extend X-ray phase-contrast imaging into previously challenging higher energy regions. The authors have measured various samples to demonstrate the feasibility of the proposed engineering materials. This technique could potentially be useful for studying samples across a wide range of applications and disciplines.

Three-dimensional visualization of extracellular matrix networks during murine development.

The extracellular matrix (ECM) plays a crucial role in embryogenesis, serving both as a substrate to which cells attach and as an active regulator of cell behavior. However, little is known about the spatiotemporal expression patterns and 3D structure of ECM proteins during embryonic development. The lack of suitable methods to visualize the embryonic ECM is largely responsible for this gap, posing a major technical challenge for biologists and tissue engineers. Here, we describe a method of viewing the 3D organization of the ECM using a polyacrylamide-based hydrogel to provide a 3D framework within developing murine embryos. After removal of soluble proteins using sodium dodecyl sulfate, confocal microscopy was used to visualize the 3D distribution of independent ECM networks in multiple developing tissues, including the forelimb, eye, and spinal cord. Comparative analysis of E12.5 and E14.5 autopods revealed proteoglycan-rich fibrils maintain connections between the epidermis and the underlying tendon and cartilage, indicating a role for the ECM during musculoskeletal assembly and demonstrating that our method can be a powerful tool for defining the spatiotemporal distribution of the ECM during embryogenesis.

Identifying the Cause of Rupture of Li-Ion Batteries during Thermal Runaway.

As the energy density of lithium-ion cells and batteries increases, controlling the outcomes of thermal runaway becomes more challenging. If the high rate of gas generation during thermal runaway is not adequately vented, commercial cell designs can rupture and explode, presenting serious safety concerns. Here, ultra-high-speed synchrotron X-ray imaging is used at >20 000 frames per second to characterize the venting processes of six different 18650 cell designs undergoing thermal runaway. For the first time, the mechanisms that lead to the most catastrophic type of cell failure, rupture, and explosion are identified and elucidated in detail. The practical application of the technique is highlighted by evaluating a novel 18650 cell design with a second vent at the base, which is shown to avoid the critical stages that lead to rupture. The insights yielded in this study shed new light on battery failure and are expected to guide the development of safer commercial cell designs.

Evaluating scintillator performance in time-resolved hard X-ray studies at synchrotron light sources.

The short pulse duration, small effective source size and high flux of synchrotron radiation is ideally suited for probing a wide range of transient deformation processes in materials under extreme conditions. In this paper, the challenges of high-resolution time-resolved indirect X-ray detection are reviewed in the context of dynamic synchrotron experiments. In particular, the discussion is targeted at two-dimensional integrating detector methods, such as those focused on dynamic radiography and diffraction experiments. The response of a scintillator to periodic synchrotron X-ray excitation is modelled and validated against experimental data collected at the Diamond Light Source (DLS) and European Synchrotron Radiation Facility (ESRF). An upper bound on the dynamic range accessible in a time-resolved experiment for a given bunch separation is calculated for a range of scintillators. New bunch structures are suggested for DLS and ESRF using the highest-performing commercially available crystal LYSO:Ce, allowing time-resolved experiments with an interframe time of 189 ns and a maximum dynamic range of 98 (6.6 bits).

Almost a spider: a 305-million-year-old fossil arachnid and spider origins.

Spiders are an important animal group, with a long history. Details of their origins remain limited, with little knowledge of their stem group, and no insights into the sequence of character acquisition during spider evolution. We describe a new fossil arachnid, Idmonarachne brasierigen. et sp. nov. from the Late Carboniferous (Stephanian,ca 305-299 Ma) of Montceau-les-Mines, France. It is three-dimensionally preserved within a siderite concretion, allowing both laboratory- and synchrotron-based phase-contrast computed tomography reconstruction. The latter is a first for siderite-hosted fossils and has allowed us to investigate fine anatomical details. Although distinctly spider-like in habitus, this remarkable fossil lacks a key diagnostic character of Araneae: spinnerets on the underside of the opisthosoma. It also lacks a flagelliform telson found in the recently recognized, spider-related, Devonian-Permian Uraraneida. Cladistic analysis resolves our new fossil as sister group to the spiders: the spider stem-group comprises the uraraneids and I. brasieri While we are unable to demonstrate the presence of spigots in this fossil, the recovered phylogeny suggests the earliest character to evolve on the spider stem-group is the secretion of silk. This would have been followed by the loss of a flagelliform telson, and then the ability to spin silk using spinnerets. This last innovation defines the true spiders, significantly post-dates the origins of silk, and may be a key to the group's success. The Montceau-les-Mines locality has previously yielded a mesothele spider (with spinnerets). Evidently, Late Palaeozoic spiders lived alongside Palaeozoic arachnid grades which approached the spider condition, but did not express the full suite of crown-group autapomorphies.

Radial lens distortion correction with sub-pixel accuracy for X-ray micro-tomography.

Distortion correction or camera calibration for an imaging system which is highly configurable and requires frequent disassembly for maintenance or replacement of parts needs a speedy method for recalibration. Here we present direct techniques for calculating distortion parameters of a non-linear model based on the correct determination of the center of distortion. These techniques are fast, very easy to implement, and accurate at sub-pixel level. The implementation at the X-ray tomography system of the I12 beamline, Diamond Light Source, which strictly requires sub-pixel accuracy, shows excellent performance in the calibration image and in the reconstructed images.

I12: the Joint Engineering, Environment and Processing (JEEP) beamline at Diamond Light Source.

I12 is the Joint Engineering, Environmental and Processing (JEEP) beamline, constructed during Phase II of the Diamond Light Source. I12 is located on a short (5 m) straight section of the Diamond storage ring and uses a 4.2 T superconducting wiggler to provide polychromatic and monochromatic X-rays in the energy range 50-150 keV. The beam energy enables good penetration through large or dense samples, combined with a large beam size (1 mrad horizontally × 0.3 mrad vertically). The beam characteristics permit the study of materials and processes inside environmental chambers without unacceptable attenuation of the beam and without the need to use sample sizes which are atypically small for the process under study. X-ray techniques available to users are radiography, tomography, energy-dispersive diffraction, monochromatic and white-beam two-dimensional diffraction/scattering and small-angle X-ray scattering. Since commencing operations in November 2009, I12 has established a broad user community in materials science and processing, chemical processing, biomedical engineering, civil engineering, environmental science, palaeontology and physics.