PVBM: A Python Vasculature Biomarker Toolbox Based on Retinal Blood Vessel Segmentation

Fhima, Jonathan; Eijgen, Jan Van; Stalmans, Ingeborg; Men, Yevgeniy; Freiman, Moti; Behar, Joachim A.

doi:10.1007/978-3-031-25066-8_15

Cited by 5 publications

(8 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DFIs provided by the University Hospitals of Leuven (UZ) were manually segmented using Lirot.ai (Fhima et al 2022a). These segmentations were used to train LUNet.…”

Section: Methodsmentioning

confidence: 99%

“…The INSPIRE-AVR, UZLF and UNAF datasets were manually segmented by the retinal experts of the UZ Leuven Hospital using the Lirot.ai app developed by Fhima et al (2022a) and following the protocol described in Fhima et al (2022b).…”

Section: Reference Segmentationsmentioning

confidence: 99%

“…Finally, LUNet was evaluated on the local test set with respect to the ability to correctly estimate vasculature biomarkers (VBMs). VBMs were computed using the open source PVBM toolbox (https://pvbm.readthedocs.io/ ) (Fhima et al 2022b). The VBMs include: the area (AREA), the tortuosity index (TI), the median arc-chord tortuosity (TOR), the length (LEN), the median branching angles (BA), the number of blood vessels that intersect with the optic disc (START), the number of endpoints (END), the number of intersection points (INTER), the fractals dimension (D0, D1 and D2).…”

Section: Performance Measuresmentioning

confidence: 99%

See 2 more Smart Citations

LUNet: deep learning for the segmentation of arterioles and venules in high resolution fundus images

Fhima,

Van Eijgen,

Billen Moulin-Romsée

et al. 2024

Physiol. Meas.

Self Cite

View full text Add to dashboard Cite

Objective: This study aims to automate the segmentation of retinal arterioles and venules (A/V) from digital fundus images (DFI), as changes in the spatial distribution of retinal microvasculature are indicative of cardiovascular diseases, positioning the eyes as windows to cardiovascular health.Approach: We utilized active learning to create a new DFI dataset with 240 crowd-sourced manual A/V segmentations performed by 15 medical students and reviewed by an ophthalmologist. We then developed LUNet, a novel deep learning architecture optimized for high-resolution A/V segmentation. The LUNet model features a double dilated convolutional block to widen the receptive field and reduce parameter count, alongside a high-resolution tail to refine segmentation details. A custom loss function was designed to prioritize the continuity of blood vessel segmentation.Main Results: LUNet significantly outperformed three benchmark A/V segmentation algorithms both on a local test set and on four external test sets that simulated variations in ethnicity, comorbidities and annotators.Significance: The release of the new datasets and the LUNet model (URL upon publication) provides a valuable resource for the advancement of retinal microvasculature analysis. The improvements in A/V segmentation accuracy highlight LUNet’s potential as a robust tool for diagnosing and understanding cardiovascular diseases through retinal imaging.

show abstract

“…DFIs provided by the University Hospitals of Leuven (UZ) were manually segmented using Lirot.ai (Fhima et al 2022a). These segmentations were used to train LUNet.…”

Section: Methodsmentioning

confidence: 99%

Section: Reference Segmentationsmentioning

confidence: 99%

Section: Performance Measuresmentioning

confidence: 99%

See 1 more Smart Citation

LUNet: deep learning for the segmentation of arterioles and venules in high resolution fundus images

Fhima,

Van Eijgen,

Billen Moulin-Romsée

et al. 2024

Physiol. Meas.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Three images will be taken from the retina of the affected eye. Retinal vessel diameters will be identified and analyzed by use of a semi-automated software (Vesselmap 2, Visualis, Imedos Systems UG) and automated softwares (LUNet and PVBM) ( Fhima et al, 2023 ).…”

Section: Interventional Methodsmentioning

confidence: 99%

High-intensity interval training in patients with glaucoma (HIT-GLAUCOMA): protocol for a multicenter randomized controlled exercise trial

Van Eijgen,

Schuhmann,

Fingerroos

et al. 2024

Front. Physiol.

Self Cite

View full text Add to dashboard Cite

BackgroundGlaucoma stands as a prominent global cause of irreversible blindness and the primary treatment approach involves reducing intraocular pressure (IOP). However, around one-third of patients exhibit disease progression despite effective IOP reduction. Microvascular endothelial function, chronic inflammation, and oxidative stress are known to affect retinal neuronal networks and have been associated with disease severity and progression. Exercise training has the potential to counteract these mechanisms as add-on treatment to usual care.AimsThe HIT-GLAUCOMA study will investigate the effects of a 6-month high-intensity interval training (HIIT) on intermediate endpoints such as local retinal microvascular and systemic large artery function, inflammation, and oxidative stress as well as clinical endpoints such as visual field indices, optic nerve rim assessment, retinal nerve fiber layer thickness, IOP, number of eye drops, vision-related quality of life and ocular surface disease symptomatology.MethodsThe study is a multi-center randomized controlled clinical trial in patients with both normal tension and high-tension primary open angle glaucoma. Across two study centers, 128 patients will be enrolled and randomized on a 1:1 basis into an exercise intervention group and a usual care control group. The primary microvascular endpoints are retinal arteriolar and venular flicker light-induced dilation at 6 months. The primary endpoint in the systemic circulation is brachial artery flow-mediated dilation at 6 months.Anticipated resultsWe hypothesize that exercise therapy will improve retinal microvascular function and thus ocular blood flow in patients with glaucoma. As clinical outcomes, we will investigate the effect of exercise on visual field indices, optic nerve rim assessment, retinal nerve fiber layer thickness, IOP, number of eye drops, vision-related quality of life and ocular surface disease symptomatology.DiscussionHIT-GLAUCOMA is a blueprint trial design to study the effect of exercise training on neurodegenerative and cardiovascular diseases. Importantly, patients are also expected to benefit from improvements in general health and cardiovascular co-morbidities. If proven effective, exercise may offer a new add-on treatment strategy to slow glaucoma progression.Clinical Trial Registration NumberThe trial is registered at Clinicaltrials.gov under the identifier NCT06058598 and is currently in the recruitment stage.

show abstract

“…While there have been several studies analysing retinal vascular phenotypes [26][27][28][29][30][31], most of them focused on measuring just one or few retinal phenotypes, often in small image sets, and some required expert input rather than being fully automated [32][33][34]. Furthermore, the software used for vascular phenotyping is usually not openly accessible, with one very recent exception [35]. Together this precludes the establishment of a comprehensive and reproducible characterisation of large retinal image collections.…”

Section: Introductionmentioning

confidence: 99%

Phenotypic and Genetic Characteristics of Retinal Vascular Parameters and their Association with Diseases

Bergmann,

Vela,

Beyeler

et al. 2023

Preprint

View full text Add to dashboard Cite

Fundus images allow for non-invasive assessment of the retinal vasculature whose features provide important information on health. Using a fully automated image processing pipeline, we extracted 17 different morphological vascular phenotypes, including median vessels diameter, diameter variability, main temporal angles, vascular density, central retinal equivalents, the number of bifurcations, and tortuosity, from over 130k fundus images of close to 72k UK Biobank subjects. We performed genome-wide association studies (GWAS) of these phenotypes. From this, we estimated their individual heritabilities, ranging between 5 and 25%, and genetic cross-phenotype correlations, which mostly mirrored the corresponding phenotypic correlations, but tended to be slightly larger. Projecting our genetic association signals onto genes and pathways revealed remarkably low overlap suggesting largely decoupled mechanisms modulating the different phenotypes. Our disease phenotype associations confirmed some previously known findings and revealed many novel connections. Notably, diameter variability, especially for the veins, seems to have new and interesting associations with diseases, including heart attack, pulmonary embolism, and age of death. Mendelian Randomization analysis suggests causal influence of blood pressure and body mass index on retinal vessel morphology, among other results. Our analyses provide evidence that large-scale analysis of image-derived vascular phenotypes has sufficient power for obtaining functional, as well as some initial causal insights into the processes modulating the retinal vasculature.

show abstract

PVBM: A Python Vasculature Biomarker Toolbox Based on Retinal Blood Vessel Segmentation

Cited by 5 publications

References 62 publications

LUNet: deep learning for the segmentation of arterioles and venules in high resolution fundus images

LUNet: deep learning for the segmentation of arterioles and venules in high resolution fundus images

High-intensity interval training in patients with glaucoma (HIT-GLAUCOMA): protocol for a multicenter randomized controlled exercise trial

Phenotypic and Genetic Characteristics of Retinal Vascular Parameters and their Association with Diseases

Contact Info

Product

Resources

About