Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Natalie Aroha Holroyd has a total of 142 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Organ-specific features of human kidney lymphatics are disrupted in chronic transplant rejection

    by Daniyal J. Jafree, Benjamin J. Stewart, Karen L. Price, Maria Kolatsi-Joannou, Camille Laroche, Barian Mohidin, Benjamin Davis, Hannah Mitchell, Lauren G. Russell, Lucía Marinas del Rey, et al.

    Lymphatic vessels maintain tissue fluid homeostasis and modulate inflammation, yet their spatial organization and molecular identity in the healthy human kidney, and how these change during chronic transplant rejection, remain poorly defined. Here, we show that lymphatic capillaries initiate adjacent to cortical kidney tubules and lack smooth muscle coverage. These vessels exhibit an organ-specific molecular signature, enriched for CCL14 , DNASE1L3 , and MDK , with limited expression of canonical immune-trafficking markers found in other organ lymphatics, such as LYVE1 and CXCL8 . In allografts with chronic mixed rejection, lymphatics become disorganized and infiltrate the medulla, with their endothelial junctions remodeling from a button-like to a continuous, zipper-like, architecture. Lymphatics in rejecting kidneys localize around and interconnect tertiary lymphoid structures at different maturation stages, with altered intralymphatic and perilymphatic CD4 + T cell distribution. The infiltrating T cells express IFN-γ, which upregulates coinhibitory ligands in lymphatic endothelial cells, including PVR and LGALS9. Simultaneously, lymphatics acquire HLA class II expression and exhibit C4d deposition, consistent with alloantibody binding and complement activation. Together, these findings define the spatial and molecular features of human kidney lymphatics, revealing tolerogenic reprogramming accompanied by structural perturbations during chronic transplant rejection.


  • Open Access

    Real time evaluation of the liver microcirculation by whole organ machine perfusion within an MRI system

    by Zainab L Rai, Natalie A Holroyd, Morenike Magbagbeola, Emre Doganay, Katie Doyle, Lucy Caselton, Agositino Stili, Danail Stoyanov, Simon Walker-Samuel, Brian R Davidson

    Objectives: Machine perfusion of organs outside of the body is a growing area of research with significant applications in the fields of organ preservation and transplantation, but more widely it offers a new approach to study disease processes and to evaluate new therapeutics and devices. Magnetic Resonance Imaging (MRI) allows for non-invasive assessment of organ structure and function, enabling quantitative measurement of tissue perfusion and microstructure. In this study, we demonstrate that MR imaging sequences can be obtained from machine-perfused porcine livers using a modified perfusion rig for MR compatibility and highlight the quantitative measures that can be obtained through this methodology. Materials and Methods: 7 porcine livers were retrieved fresh from the abattoir using a previously published protocol and following transport in cold preservative underwent perfusion with oxygenated autologous blood inside a 3T clinical MRI scanner using a custom modified perfusion rig. Multiple MR imaging sequences were acquired: T2-weighted imaging, Diffusion Weighted Imaging and Dynamic Contrast Enhanced imaging following injection of Gadolinium dye into the portal vein and hepatic artery. Histological analysis was performed to assess preservation injury to the liver. Control samples for histology were obtained from livers with similar preservation periods but preserved in standard cold storage on ice (Static Cold Storage). Results: Concurrent MR imaging and machine perfusion were successfully performed, allowing dynamic measurement of tissue perfusion to be obtained in ex vivo livers, including calculation of gadolinium contrast enhancement curves and Apparent Diffusion Coefficient maps. Segmentation of vessels down to a radius of 0.45mm allowed detailed morphological analysis of the vascular network, including extraction of clinically relevant parameters such as vessel tortuosity. Histological evaluation showed better preservation of the hepatic acinar structure in perfused than non-perfused livers. Conclusions: Our results demonstrate that MR imaging of machine-perfused organs enables high-resolution quantitative evaluation of whole-organ vascular morphology and flow dynamics. This platform provides opportunities to study vascular pathology in diseased human organs and evaluate novel therapeutic interventions, with particular relevance for drug-delivery strategies.

  • Open Access

    Author Correction: Physics-informed deep generative learning for quantitative assessment of the retina

    by Emmeline E. Brown, Andrew A. Guy, Natalie A. Holroyd, Paul W. Sweeney, Lucie Gourmet, Hannah Coleman, Claire Walsh, Athina E. Markaki, Rebecca Shipley, Ranjan Rajendram, et al.


  • Open Access

    tUbeNet: a generalizable deep learning tool for 3D vessel segmentation

    by Natalie A Holroyd, Zhongwang Li, Claire Walsh, Emmeline Brown, Rebecca J Shipley, Simon Walker-Samuel

    Deep learning has become an invaluable tool for bioimage analysis but, while open-source cell annotation software such as Cellpose is widely used, an equivalent tool for three-dimensional (3D) vascular annotation does not exist. With the vascular system being directly impacted by a broad range of diseases, there is significant medical interest in quantitative analysis for vascular imaging. We present a new deep learning model, coupled with a human-in-the-loop training approach, for segmentation of vasculature that is generalizable across tissues, modalities, scales, and pathologies. To create a generalizable model, a 3D convolutional neural network was trained using curated data from modalities including optical imaging, computational tomography, and photoacoustic imaging. Through this varied training set, the model was forced to learn common features of vessels’ cross-modality and scale. Following this, the pre-trained ‘foundation’ model was fine-tuned to different applications with a minimal amount of manually labelled ground truth data. It was found that the foundation model could be specialized to a new datasets using as little as 0.3% of the volume of said dataset for fine-tuning. The fine-tuned model was able to segment 3D vasculature with a high level of accuracy (DICE coefficient between 0.81 and 0.98) across a range of applications. These results show a general model trained on a highly varied data catalogue can be specialized to new applications with minimal human input. This model and training approach enables users to produce accurate segmentations of 3D vascular networks without the need to label large amounts of training data.


  • Open Access

    Physics-informed deep generative learning for quantitative assessment of the retina

    by Emmeline E. Brown, Andrew A. Guy, Natalie A. Holroyd, Paul W. Sweeney, Lucie Gourmet, Hannah Coleman, Claire Walsh, Athina E. Markaki, Rebecca Shipley, Ranjan Rajendram, et al.

    Disruption of retinal vasculature is linked to various diseases, including diabetic retinopathy and macular degeneration, leading to vision loss. We present here a novel algorithmic approach that generates highly realistic digital models of human retinal blood vessels, based on established biophysical principles, including fully-connected arterial and venous trees with a single inlet and outlet. This approach, using physics-informed generative adversarial networks (PI-GAN), enables the segmentation and reconstruction of blood vessel networks with no human input and which out-performs human labelling. Segmentation of DRIVE and STARE retina photograph datasets provided near state-of-the-art vessel segmentation, with training on only a small (n = 100) simulated dataset. Our findings highlight the potential of PI-GAN for accurate retinal vasculature characterization, with implications for improving early disease detection, monitoring disease progression, and improving patient care.


  • Open Access

    Reconstructing microvascular network skeletons from 3D images: What is the ground truth?

    by Claire L. Walsh, Maxime Berg, Hannah West, Natalie A. Holroyd, Simon Walker-Samuel, Rebecca J. Shipley


  • Open Access

    Reconstructing microvascular network skeletons from 3D images: what is the ground truth?

    by Claire Walsh, Maxime Berg, Hannah West, Natalie A. Holroyd, Simon Walker-Samuel, Rebecca J. Shipley

    Structural changes to microvascular networks are increasingly highlighted as markers of pathogenesis in a wide range of disease, e.g. Alzheimer’s disease, vascular dementia and tumour growth. This has motivated the development of dedicated 3D imaging techniques, alongside the creation of computational modelling frameworks capable of using 3D reconstructed networks to simulate functional behaviours such as blood flow or transport processes. Extraction of 3D networks from imaging data broadly consists of two image processing steps: segmentation followed by skeletonisation. Much research effort has been devoted to segmentation field, and there are standard and widely-applied methodologies for creating and assessing gold standards or ground truths produced by manual annotation or automated algorithms. The Skeletonisation field, however, lacks widely applied, simple to compute metrics for the validation or optimisation of the numerous algorithms that exist to extract skeletons from binary images. This is particularly problematic as 3D imaging datasets increase in size and visual inspection becomes an insufficient validation approach. In this work, we first demonstrate the extent of the problem by applying 4 widely-used skeletonisation algorithms to 3 different imaging datasets. In doing so we show significant variability between reconstructed skeletons of the same segmented imaging dataset. Moreover, we show that such a structural variability propagates to simulated metrics such as blood flow. To mitigate this variability we introduce a new, fast and easy to compute super-metric that compares the volume, connectivity, medialness, correct bifurcation point identification and homology of the reconstructed skeletons to the original segmented data. We then show that such a metric can be used to select the best performing skeletonisation algorithm for a given dataset, as well as to optimize its parameters. Finally, we demonstrate that the super-metric can also be used to quickly identify how a particular skeletonisation algorithm could be improved, becoming a powerful tool in understanding the complex implication of small structural changes in a network.

  • Open Access

    Emergent mechanical control of vascular morphogenesis

    by Jordan Whisler, Somayeh Shahreza, Karin Schlegelmilch, Nil Ege, Yousef Javanmardi, Andrea Malandrino, Ayushi Agrawal, Alessandro Fantin, Bianca Serwinski, Hesham Azizgolshani, et al.

    Vascularization is driven by morphogen signals and mechanical cues that coordinately regulate cellular force generation, migration, and shape change to sculpt the developing vascular network. However, it remains unclear whether developing vasculature actively regulates its own mechanical properties to achieve effective vascularization. We engineered tissue constructs containing endothelial cells and fibroblasts to investigate the mechanics of vascularization. Tissue stiffness increases during vascular morphogenesis resulting from emergent interactions between endothelial cells, fibroblasts, and ECM and correlates with enhanced vascular function. Contractile cellular forces are key to emergent tissue stiffening and synergize with ECM mechanical properties to modulate the mechanics of vascularization. Emergent tissue stiffening and vascular function rely on mechanotransduction signaling within fibroblasts, mediated by YAP1. Mouse embryos lacking YAP1 in fibroblasts exhibit both reduced tissue stiffness and develop lethal vascular defects. Translating our findings through biology-inspired vascular tissue engineering approaches will have substantial implications in regenerative medicine.


  • Open Access

    tUbe net: a generalisable deep learning tool for 3D vessel segmentation

    by Natalie A. Holroyd, Zhongwang Li, Claire Walsh, Emmeline Browng, Rebecca J. Shipley, Simon Walker-Samuel

    Deep learning has become an invaluable tool for bioimage analysis but, while open-source cell annotation software such as cellpose are widely used, an equivalent tool for three-dimensional (3D) vascular annotation does not exist. With the vascular system being directly impacted by a broad range of diseases, there is significant medical interest in quantitative analysis for vascular imaging. We present a new deep learning model, coupled with a human-in-the-loop training approach, for segmentation of vasculature that is generalisable across tissues, modalities, scales and pathologies. To create a generalisable model, a 3D convolutional neural network was trained using curated data from modalities including optical imaging, computational tomography and photoacoustic imaging. Through this varied training set, the model was forced to learn common features of vessels cross-modality and scale. Following this, the pre-trained ‘foundation’ model was fine-tuned to different applications with a minimal amount of manually labelled ground truth data. It was found that the foundation model could be specialised to a new datasets using as little as 0.3% of the volume of said dataset for fine-tuning. The fine-tuned model was able to segment 3D vasculature with a high level of accuracy (DICE coefficient between 0.89 and 0.98) across a range of applications. These results show a general model trained on a highly varied data catalogue can be specialised to new applications with minimal human input. This model and training approach enables users to produce accurate segmentations of 3D vascular networks without the need to label large amounts of training data.


  • Open Access

    Physics-informed deep generative learning for quantitative assessment of the retina

    by Emmeline Brown, Andrew Guy, Natalie Holroyd, Paul Sweeney, Lucie Gourmet, Hannah Coleman, Claire Walsh, Athina Markaki, Rebecca Shipley, Ranjan Rajendram, et al.

    Disruption of retinal vasculature is linked to various diseases, including diabetic retinopathy and macular degeneration, leading to vision loss. We present here a novel algorithmic approach that generates highly realistic digital models of human retinal blood vessels based on established biophysical principles, including fully-connected arterial and venous trees with a single inlet and outlet. This approach, using physics-informed generative adversarial networks (PI-GAN), enables the segmentation and reconstruction of blood vessel networks that requires no human input and out-performs human labelling. Our findings highlight the potential of PI-GAN for accurate retinal vasculature characterization, with implications for improving early disease detection, monitoring disease progression, and improving patient care.