We present the Human Organ Atlas (HOA), an open data repository making accessible multiscale three-dimensional imaging of human organs. The repository also provides software tools and training resources enabling worldwide access, sharing, and analysis of these datasets, facilitating further research and the continued expansion of the HOA. The images are generated using a synchrotron imaging technique, hierarchical phase-contrast tomography (HiP-CT), that uses the ESRF’s Extremely Brilliant Source, spanning whole-organ imaging at ~20 micrometers/voxel with local volumes of interest within intact organs imaged down to ~1 micrometer/voxel. This offers a comprehensive exploration of human anatomy, providing unparalleled insights into intricate structures and spatial relationships. The HOA offers researchers, clinicians, and educators a valuable resource for anatomical study, image analysis, medical education, and large-scale data mining.
Efficient algorithms are needed to segment vasculature in new three-dimensional (3D) medical imaging datasets at scale for a wide range of research and clinical applications. Manual segmentation of vessels in images is time-consuming and expensive. Computational approaches are more scalable but have limitations in accuracy. We organized a global machine learning competition, engaging 1,401 participants, to help develop new deep learning methods for 3D blood vessel segmentation. This paper presents a detailed analysis of the top-performing solutions using manually curated 3D Hierarchical Phase-Contrast Tomography datasets of the human kidney, focusing on the segmentation accuracy and morphological analysis, thereby establishing a benchmark for future studies in blood vessel segmentation within phase-contrast tomography imaging.
With increasing interest in studying biological systems across spatial scales—from centimeters down to nanometers—histology continues to be the gold standard for tissue imaging at cellular resolution, providing an essential bridge between macroscopic and nanoscopic analysis. However, its inherently destructive and two-dimensional nature limits its ability to capture the full three-dimensional complexity of tissue architecture. Here, we show that phase-contrast X-ray microscopy can enable three-dimensional virtual histology with subcellular resolution. This technique provides direct quantification of electron density without restrictive assumptions, allowing for direct characterization of cellular nuclei in a standard laboratory setting. By combining high spatial resolution and soft tissue contrast, with automated segmentation of cell nuclei, we demonstrated virtual Hematoxylin and Eosin (H&E) staining using machine learning-based style transfer, yielding volumetric datasets compatible with existing histopathological analysis tools. Furthermore, by integrating electron density and the sensitivity to nanometric features of the dark field contrast channel, we achieve stain-free, high-content imaging capable of distinguishing nuclei and extracellular matrix.
Biomedical systems span multiple spatial scales, encompassing tiny functional units to entire organs. Interpreting these systems through image segmentation requires the effective propagation and integration of information across different scales. However, most existing segmentation methods are optimised for single-scale imaging modalities, limiting their ability to capture and analyse small functional units throughout complete human organs. To facilitate multiscale biomedical image segmentation, we utilised Hierarchical Phase-Contrast Tomography (HiP-CT), an advanced imaging modality that can generate 3D multiscale datasets from high-resolution volumes of interest (VOIs) at ca. 1 /voxel to whole-organ scans at ca. 20 /voxel. Building on these hierarchical multiscale datasets, we developed a deep learning-based segmentation pipeline that is initially trained on manually annotated high-resolution HiP-CT data and then extended to lower-resolution whole-organ scans using pseudo-labels generated from high-resolution predictions and multiscale image registration. As a case study, we focused on glomeruli in human kidneys, benchmarking four 3D deep learning models for biomedical image segmentation on a manually annotated high-resolution dataset extracted from VOIs, at 2.58 to ca. 5 /voxel, of four human kidneys. Among them, nnUNet demonstrated the best performance, achieving an average test Dice score of 0.906, and was subsequently used as the baseline model for multiscale segmentation in the pipeline. Applying this pipeline to two low-resolution full-organ data at ca. 25 /voxel, the model identified 1,019,890 and 231,179 glomeruli in a 62-year-old donor without kidney diseases and a 94-year-old hypertensive donor, enabling comprehensive morphological analyses, including cortical spatial statistics and glomerular distributions, which aligned well with previous anatomical studies. Our results highlight the effectiveness of the proposed pipeline for segmenting small functional units in multiscale bioimaging datasets and suggest its broader applicability to other organ systems.