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

Abstract 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...

Full description

Saved in:
Bibliographic Details
Main Authors: 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, Simon Walker-Samuel
Format: Article
Language:English
Published: Nature Portfolio 2024-08-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-024-50911-y
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850067311731408896
author 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
Simon Walker-Samuel
author_facet 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
Simon Walker-Samuel
author_sort Emmeline E. Brown
collection DOAJ
description Abstract 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.
format Article
id doaj-art-0bd85a757e2e4546b32f953ae15e0ae5
institution DOAJ
issn 2041-1723
language English
publishDate 2024-08-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-0bd85a757e2e4546b32f953ae15e0ae52025-08-20T02:48:22ZengNature PortfolioNature Communications2041-17232024-08-0115111410.1038/s41467-024-50911-yPhysics-informed deep generative learning for quantitative assessment of the retinaEmmeline E. Brown0Andrew A. Guy1Natalie A. Holroyd2Paul W. Sweeney3Lucie Gourmet4Hannah Coleman5Claire Walsh6Athina E. Markaki7Rebecca Shipley8Ranjan Rajendram9Simon Walker-Samuel10Centre for Computational Medicine, University College LondonCentre for Computational Medicine, University College LondonCentre for Computational Medicine, University College LondonCancer Research UK Cambridge Institute, University of CambridgeCentre for Computational Medicine, University College LondonCentre for Computational Medicine, University College LondonCentre for Computational Medicine, University College LondonDepartment of Engineering, University of CambridgeCentre for Computational Medicine, University College LondonMoorfields Eye HospitalCentre for Computational Medicine, University College LondonAbstract 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.https://doi.org/10.1038/s41467-024-50911-y
spellingShingle 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
Simon Walker-Samuel
Physics-informed deep generative learning for quantitative assessment of the retina
Nature Communications
title Physics-informed deep generative learning for quantitative assessment of the retina
title_full Physics-informed deep generative learning for quantitative assessment of the retina
title_fullStr Physics-informed deep generative learning for quantitative assessment of the retina
title_full_unstemmed Physics-informed deep generative learning for quantitative assessment of the retina
title_short Physics-informed deep generative learning for quantitative assessment of the retina
title_sort physics informed deep generative learning for quantitative assessment of the retina
url https://doi.org/10.1038/s41467-024-50911-y
work_keys_str_mv AT emmelineebrown physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT andrewaguy physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT natalieaholroyd physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT paulwsweeney physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT luciegourmet physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT hannahcoleman physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT clairewalsh physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT athinaemarkaki physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT rebeccashipley physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT ranjanrajendram physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina
AT simonwalkersamuel physicsinformeddeepgenerativelearningforquantitativeassessmentoftheretina