A multiscale X-ray phase-contrast tomography dataset of a whole human left lung

Real-world federated learning in radiology: hurdles to overcome and benefits to gain

OBJECTIVE

Federated Learning (FL) enables collaborative model training while keeping data locally. Currently, most FL studies in radiology are conducted in simulated environments due to numerous hurdles impeding its translation into practice. The few existing real-world FL initiatives rarely communicate specific measures taken to overcome these hurdles. To bridge this significant knowledge gap, we propose a comprehensive guide for real-world FL in radiology. Minding efforts to implement real-world FL, there is a lack of comprehensive assessments comparing FL to less complex alternatives in challenging real-world settings, which we address through extensive benchmarking.

MATERIALS AND METHODS

We developed our own FL infrastructure within the German Radiological Cooperative Network (RACOON) and demonstrated its functionality by training FL models on lung pathology segmentation tasks across six university hospitals. Insights gained while establishing our FL initiative and running the extensive benchmark experiments were compiled and categorized into the guide.

RESULTS

The proposed guide outlines essential steps, identified hurdles, and implemented solutions for establishing successful FL initiatives conducting real-world experiments. Our experimental results prove the practical relevance of our guide and show that FL outperforms less complex alternatives in all evaluation scenarios.

DISCUSSION AND CONCLUSION

Our findings justify the efforts required to translate FL into real-world applications by demonstrating advantageous performance over alternative approaches. Additionally, they emphasize the importance of strategic organization, robust management of distributed data and infrastructure in real-world settings. With the proposed guide, we are aiming to aid future FL researchers in circumventing pitfalls and accelerating translation of FL into radiological applications.

3D digital light process bioprinting: Cutting-edge platforms for resolution of organ fabrication

Research in the field of regenerative medicine, which replaces or restores the function of human damaged organs is advancing rapidly. These advances are fostering important innovations in the development of artificial organs. In recent years, three-dimensional (3D) bioprinting has emerged as a promising technology for regenerative medicine applications. Among various techniques, digital light process (DLP) 3D bioprinting stands out for its ability to precisely create high-resolution, structurally complex artificial organs. This review explores the types and usage trends of DLP printing equipment, bioinks, and photoinitiators. Building on this foundation, the applications of DLP bioprinting for creating precise microstructures of human organs and for regenerating tissue and organ models in regenerative medicine are examined. Finally, challenges and future perspectives regarding DLP-based bioprinting, particularly for precision printing applications in regenerative medicine, are discussed.

Deep learning for 3D vascular segmentation in hierarchical phase contrast tomography: a case study on kidney

Automated blood vessel segmentation is critical for biomedical image analysis, as vessel morphology changes are associated with numerous pathologies. Still, precise segmentation is difficult due to the complexity of vascular structures, anatomical variations across patients, the scarcity of annotated public datasets, and the quality of images. Our goal is to provide a foundation on the topic and identify a robust baseline model for application to vascular segmentation using a new imaging modality, Hierarchical Phase-Contrast Tomography (HiP-CT). We begin with an extensive review of current machine-learning approaches for vascular segmentation across various organs. Our work introduces a meticulously curated training dataset, verified by double annotators, consisting of vascular data from three kidneys imaged using HiP-CT as part of the Human Organ Atlas Project. HiP-CT pioneered at the European Synchrotron Radiation Facility in 2020, revolutionizes 3D organ imaging by offering a resolution of around 20 μm/voxel and enabling highly detailed localised zooms up to 1-2 μm/voxel without physical sectioning. We leverage the nnU-Net framework to evaluate model performance on this high-resolution dataset, using both known and novel samples, and implementing metrics tailored for vascular structures. Our comprehensive review and empirical analysis on HiP-CT data sets a new standard for evaluating machine learning models in high-resolution organ imaging. Our three experiments yielded Dice similarity coefficient (DSC) scores of 0.9523, 0.9410, and 0.8585, respectively. Nevertheless, DSC primarily assesses voxel-to-voxel concordance, overlooking several crucial characteristics of the vessels and should not be the sole metric for deciding the performance of vascular segmentation. Our results show that while segmentations yielded reasonably high scores-such as centerline DSC ranging from 0.82 to 0.88, certain errors persisted. Specifically, large vessels that collapsed due to the lack of hydrostatic pressure (HiP-CT is an ex vivo technique) were segmented poorly. Moreover, decreased connectivity in finer vessels and higher segmentation errors at vessel boundaries were observed. Such errors, particularly in significant vessels, obstruct the understanding of the structures by interrupting vascular tree connectivity. Our study establishes the benchmark across various evaluation metrics, for vascular segmentation of HiP-CT imaging data, an imaging technology that has the potential to substantively shift our understanding of human vascular networks.

Computer vision applications for the detection or analysis of tuberculosis using digitised human lung tissue images - a systematic review

OBJECTIVE

To conduct a systematic review of the computer vision applications that detect, diagnose, or analyse tuberculosis (TB) pathology or bacilli using digitised human lung tissue images either through automatic or semi-automatic methods. We categorised the computer vision platform into four technologies: image processing, object/pattern recognition, computer graphics, and deep learning. In this paper, the focus is on image processing and deep learning (DL) applications for either 2D or 3D digitised human lung tissue images. This review is useful for establishing a common practice in TB analysis using human lung tissue as well as identifying opportunities for further research in this space. The review brings attention to the state-of-art techniques for detecting TB, with emphasis on the challenges and limitations of the current techniques. The ultimate goal is to promote the development of more efficient and accurate algorithms for the detection or analysis of TB, and raise awareness about the importance of early detection.

DESIGN

We searched five databases and Google Scholar for articles published between January 2017 and December 2022 that focus on Mycobacterium tuberculosis detection, or tuberculosis pathology using digitised human lung tissue images. Details regarding design, image processing and computer-aided techniques, deep learning models, and datasets were collected and summarised. Discussions, analysis, and comparisons of state-of-the-art methods are provided to help guide future research. Further, a brief update on the relevant techniques and their performance is provided.

RESULTS

Several studies have been conducted to develop automated and AI-assisted methods for diagnosing Mtb and TB pathology from digitised human lung tissue images. Some studies presented a completely automated method of diagnosis, while other studies developed AI-assisted diagnostic methods. Low-level focus areas included the development of a novel μ CT scanner for soft tissue image contract, and use of multiresolution computed tomography to analyse the 3D structure of the human lung. High-level focus areas included the investigation the effects of aging on the number and size of small airways in the lungs using CT and whole lung high-resolution μ CT, and the 3D microanatomy characterisation of human tuberculosis lung using μ CT in conjunction with histology and immunohistochemistry. Additionally, a novel method for acquiring high-resolution 3D images of human lung structure and topology is also presented.

CONCLUSION

The literature indicates that post 1950s, TB was predominantly studied using animal models even though no animal model reflects the full spectrum of human pulmonary TB disease and does not reproducibly transmit Mtb infection to other animals (Hunter, 2011). This explains why there are very few studies that used human lung tissue for detection or analysis of Mtb. Nonetheless, we found 10 studies that used human tissues (predominately lung) of which five studies proposed machine learning (ML) models for the detection of bacilli and the other five used CT on human lung tissue scanned ex-vivo.

Synchrotron X-ray imaging of soft biological tissues - principles, applications and future prospects

Synchrotron-based tomographic phase-contrast X-ray imaging (SRµCT or SRnCT) is a versatile isotropic three-dimensional imaging technique that can be used to study biological samples spanning from single cells to human-sized specimens. SRµCT and SRnCT take advantage of the highly brilliant and coherent X-rays produced by a synchrotron light source. This enables fast data acquisition and enhanced image contrast for soft biological samples owing to the exploitation of phase contrast. In this Review, we provide an overview of the basics behind the technique, discuss its applications for biologists and provide an outlook on the future of this emerging technique for biology. We introduce the latest advances in the field, such as whole human organs imaged with micron resolution, using X-rays as a tool for virtual histology and resolving neuronal connections in the brain.

High-angular-sensitivity X-ray phase-contrast microtomography of soft tissue through a two-directional beam-tracking synchrotron set-up

Two-directional beam-tracking (2DBT) is a method for phase-contrast imaging and tomography that uses an intensity modulator to structure the X-ray beam into an array of independent circular beamlets that are resolved by a high-resolution detector. It features isotropic spatial resolution, provides two-dimensional phase sensitivity, and enables the three-dimensional reconstructions of the refractive index decrement, δ, and the attenuation coefficient, μ. In this work, the angular sensitivity and the spatial resolution of 2DBT images in a synchrotron-based implementation is reported. In its best configuration, angular sensitivities of ∼20 nrad and spatial resolution of at least 6.25 µm in phase-contrast images were obtained. Exemplar application to the three-dimensional imaging of soft tissue samples, including a mouse liver and a decellularized porcine dermis, is also demonstrated.

Deep Learning for 3D Vascular Segmentation in Phase Contrast Tomography

Automated blood vessel segmentation is critical for biomedical image analysis, as vessel morphology changes are associated with numerous pathologies. Still, precise segmentation is difficult due to the complexity of vascular structures, anatomical variations across patients, the scarcity of annotated public datasets, and the quality of images. Our goal is to provide a foundation on the topic and identify a robust baseline model for application to vascular segmentation using a new imaging modality, Hierarchical Phase-Contrast Tomography (HiP-CT). We begin with an extensive review of current machine learning approaches for vascular segmentation across various organs. Our work introduces a meticulously curated training dataset, verified by double annotators, consisting of vascular data from three kidneys imaged using Hierarchical Phase-Contrast Tomography (HiP-CT) as part of the Human Organ Atlas Project. HiP-CT, pioneered at the European Synchrotron Radiation Facility in 2020, revolutionizes 3D organ imaging by offering resolution around 20μm/voxel, and enabling highly detailed localized zooms up to 1μm/voxel without physical sectioning. We leverage the nnU-Net framework to evaluate model performance on this high-resolution dataset, using both known and novel samples, and implementing metrics tailored for vascular structures. Our comprehensive review and empirical analysis on HiP-CT data sets a new standard for evaluating machine learning models in high-resolution organ imaging. Our three experiments yielded Dice scores of 0.9523 and 0.9410, and 0.8585, respectively. Nevertheless, DSC primarily assesses voxel-to-voxel concordance, overlooking several crucial characteristics of the vessels and should not be the sole metric for deciding the performance of vascular segmentation. Our results show that while segmentations yielded reasonably high scores-such as centerline Dice values ranging from 0.82 to 0.88, certain errors persisted. Specifically, large vessels that collapsed due to the lack of hydro-static pressure (HiP-CT is an ex vivo technique) were segmented poorly. Moreover, decreased connectivity in finer vessels and higher segmentation errors at vessel boundaries were observed. Such errors, particularly in significant vessels, obstruct the understanding of the structures by interrupting vascular tree connectivity. Through our review and outputs, we aim to set a benchmark for subsequent model evaluations using various modalities, especially with the HiP-CT imaging database.

Multidimensional Analysis of the Adult Human Heart in Health and Disease Using Hierarchical Phase-Contrast Tomography

Background Current clinical imaging modalities such as CT and MRI provide resolution adequate to diagnose cardiovascular diseases but cannot depict detailed structural features in the heart across length scales. Hierarchical phase-contrast tomography (HiP-CT) uses fourth-generation synchrotron sources with improved x-ray brilliance and high energies to provide micron-resolution imaging of intact adult organs with unprecedented detail. Purpose To evaluate the capability of HiP-CT to depict the macro- to microanatomy of structurally normal and abnormal adult human hearts ex vivo. Materials and Methods Between February 2021 and September 2023, two adult human donor hearts were obtained, fixed in formalin, and prepared using a mixture of crushed agar in a 70% ethanol solution. One heart was from a 63-year-old White male without known cardiac disease, and the other was from an 87-year-old White female with a history of multiple known cardiovascular pathologies including ischemic heart disease, hypertension, and atrial fibrillation. Nondestructive ex vivo imaging of these hearts without exogenous contrast agent was performed using HiP-CT at the European Synchrotron Radiation Facility. Results HiP-CT demonstrated the capacity for high-spatial-resolution, multiscale cardiac imaging ex vivo, revealing histologic-level detail of the myocardium, valves, coronary arteries, and cardiac conduction system across length scales. Virtual sectioning of the cardiac conduction system provided information on fatty infiltration, vascular supply, and pathways between the cardiac nodes and adjacent structures. HiP-CT achieved resolutions ranging from gross (isotropic voxels of approximately 20 µm) to microscopic (approximately 6.4-µm voxel size) to cellular (approximately 2.3-µm voxel size) in scale. The potential for quantitative assessment of features in health and disease was demonstrated. Conclusion HiP-CT provided high-spatial-resolution, three-dimensional images of structurally normal and diseased ex vivo adult human hearts. Whole-heart image volumes were obtained with isotropic voxels of approximately 20 µm, and local regions of interest were obtained with resolution down to 2.3-6.4 µm without the need for sectioning, destructive techniques, or exogenous contrast agents. Published under a CC BY 4.0 license Supplemental material is available for this article. See also the editorial by Bluemke and Pourmorteza in this issue.

A closer look at high-energy X-ray-induced bubble formation during soft tissue imaging

Improving the scalability of tissue imaging throughput with bright, coherent X-rays requires identifying and mitigating artifacts resulting from the interactions between X-rays and matter. At synchrotron sources, long-term imaging of soft tissues in solution can result in gas bubble formation or cavitation, which dramatically compromises image quality and integrity of the samples. By combining in-line phase-contrast imaging with gas chromatography in real time, we were able to track the onset and evolution of high-energy X-ray-induced gas bubbles in ethanol-embedded soft tissue samples for tens of minutes (two to three times the typical scan times). We demonstrate quantitatively that vacuum degassing of the sample during preparation can significantly delay bubble formation, offering up to a twofold improvement in dose tolerance, depending on the tissue type. However, once nucleated, bubble growth is faster in degassed than undegassed samples, indicating their distinct metastable states at bubble onset. Gas chromatography analysis shows increased solvent vaporization concurrent with bubble formation, yet the quantities of dissolved gasses remain unchanged. By coupling features extracted from the radiographs with computational analysis of bubble characteristics, we uncover dose-controlled kinetics and nucleation site-specific growth. These hallmark signatures provide quantitative constraints on the driving mechanisms of bubble formation and growth. Overall, the observations highlight bubble formation as a critical yet often overlooked hurdle in upscaling X-ray imaging for biological tissues and soft materials and we offer an empirical foundation for their understanding and imaging protocol optimization. More importantly, our approaches establish a top-down scheme to decipher the complex, multiscale radiation-matter interactions in these applications.

High-Resolution Phase-Contrast Tomography on Human Collagenous Tissues: A Comprehensive Review

Phase-contrast X-ray imaging is becoming increasingly considered since its first applications, which occurred almost 30 years ago. Particular emphasis was placed on studies that use this technique to investigate soft tissues, which cannot otherwise be investigated at a high resolution and in a three-dimensional manner, using conventional absorption-based settings. Indeed, its consistency and discrimination power in low absorbing samples, unified to being a not destructive analysis, are pushing interests on its utilization from researchers of different specializations, from botany, through zoology, to human physio-pathology research. In this regard, a challenging method for 3D imaging and quantitative analysis of collagenous tissues has spread in recent years: it is based on the unique characteristics of synchrotron radiation phase-contrast microTomography (PhC-microCT). In this review, the focus has been placed on the research based on the exploitation of synchrotron PhC-microCT for the investigation of collagenous tissue physio-pathologies from solely human samples. Collagen tissues' elasto-mechanic role bonds it to the morphology of the site it is extracted from, which could weaken the results coming from animal experimentations. Encouraging outcomes proved this technique to be suitable to access and quantify human collagenous tissues and persuaded different researchers to approach it. A brief mention was also dedicated to the results obtained on collagenous tissues using new and promising high-resolution phase-contrast tomographic laboratory-based setups, which will certainly represent the real step forward in the diffusion of this relatively young imaging technique.

Multidimensional Analysis of the Adult Human Heart in Health and Disease using Hierarchical Phase-Contrast Tomography (HiP-CT)

Cardiovascular diseases (CVDs) are a leading cause of death worldwide. Current clinical imaging modalities provide resolution adequate for diagnosis but are unable to provide detail of structural changes in the heart, across length-scales, necessary for understanding underlying pathophysiology of disease. Hierarchical Phase-Contrast Tomography (HiP-CT), using new (4th) generation synchrotron sources, potentially overcomes this limitation, allowing micron resolution imaging of intact adult organs with unprecedented detail. In this proof of principle study (n=2), we show the utility of HiP-CT to image whole adult human hearts ex-vivo: one 'control' without known cardiac disease and one with multiple known cardiopulmonary pathologies. The resulting multiscale imaging was able to demonstrate exemplars of anatomy in each cardiac segment along with novel findings in the cardiac conduction system, from gross (20 um/voxel) to cellular scale (2.2 um/voxel), non-destructively, thereby bridging the gap between macroscopic and microscopic investigations. We propose that the technique represents a significant step in virtual autopsy methods for studying structural heart disease, facilitating research into abnormalities across scales and age-groups. It opens up possibilities for understanding and treating disease; and provides a cardiac 'blueprint' with potential for in-silico simulation, device design, virtual surgical training, and bioengineered heart in the future.

Preparation of large biological samples for high-resolution, hierarchical, synchrotron phase-contrast tomography with multimodal imaging compatibility

Imaging across different scales is essential for understanding healthy organ morphology and pathophysiological changes. The macro- and microscale three-dimensional morphology of large samples, including intact human organs, is possible with X-ray microtomography (using laboratory or synchrotron sources). Preparation of large samples for high-resolution imaging, however, is challenging due to limitations such as sample shrinkage, insufficient contrast, movement of the sample and bubble formation during mounting or scanning. Here, we describe the preparation, stabilization, dehydration and mounting of large soft-tissue samples for X-ray microtomography. We detail the protocol applied to whole human organs and hierarchical phase-contrast tomography at the European Synchrotron Radiation Facility, yet it is applicable to a range of biological samples, including complete organisms. The protocol enhances the contrast when using X-ray imaging, while preventing sample motion during the scan, even with different sample orientations. Bubbles trapped during mounting and those formed during scanning (in the case of synchrotron X-ray imaging) are mitigated by multiple degassing steps. The sample preparation is also compatible with magnetic resonance imaging, computed tomography and histological observation. The sample preparation and mounting require 24-36 d for a large organ such as a whole human brain or heart. The preparation time varies depending on the composition, size and fragility of the tissue. Use of the protocol enables scanning of intact organs with a diameter of 150 mm with a local voxel size of 1 μm. The protocol requires users with expertise in handling human or animal organs, laboratory operation and X-ray imaging.