Quantification of retinal microvascular imaging features from fundus photos in ocular and systemic disease: a framework for standardization

TY - JOUR

T1 - Quantification of retinal microvascular imaging features from fundus photos in ocular and systemic disease

T2 - a framework for standardization

AU - van der Heide, Frank CT.

AU - Zamoum, Youba

AU - Ounissi, Mehdi

AU - Berendschot, Tos Tjm

AU - Cheung, Carol Y.

AU - Koronyo-Hamaoui, Maya

AU - Schmetterer, Leopold

AU - Chua, Jacqueline

AU - Peto, Tunde

AU - Lengyel, Imre

AU - Csincsik, Lajos

AU - Creuzot, Catherine

AU - Arnould, Louis

AU - Stalmans, Ingeborg

AU - Trucco, Emanuele

AU - MacGillivray, Tom J

AU - Khawaja, Anthony P.

AU - Zhu, Lisa Zhuoting

AU - Keane, Pearse A.

AU - Wong, Tien-Yin

AU - Milea, Dan

N1 - © 2026 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

PY - 2026/3/25

Y1 - 2026/3/25

N2 - Microvascular dysfunction is increasingly recognized as an important contributor to ocular and systemic diseases, including diabetic retinopathy, cardiovascular disease, and Alzheimer's disease and related dementias. Elucidating how early microvascular injury contributes to disease pathobiology, and development of sensitive, quantifiable features of microvascular dysfunction, represents a critical frontier for improving risk stratification, enabling earlier diagnosis and guiding targeted interventions. In this context, retinal imaging has emerged as a powerful modality, enabling non-invasive, high-resolution visualization of the microvasculature with color and (ultra)widefield fundus imaging. Rapid technological advances have led to an array of, open-source tools for quantitative extraction of retinal microvascular features. However, progress in this field is hampered by poor comparability between tools, limiting reproducibility and cross-study integration. Differences between tools originate from variations in image processing steps, including vessel segmentation, arteriole–venule classification, optic disc detection, region-of-interest definition, image quality assessment, and the algorithms for metric calculation. In this review we comprehensively compare existing analytical analysis tools for static vessel analyses and delineate how these methodological differences influence the vascular measure quantification. To advance robust, scalable biomarker development, we propose a methodological framework to standardize and harmonize retinal microvascular quantification. This framework can guide future studies in addressing key gaps in literature and in developing imaging biomarkers for clinical use. Critical open questions include whether and when retinal imaging features change during ocular and systemic disease, whether microvascular dysfunction is reversible, how microvascular dysfunction contributes to neurodegeneration, and how central and peripheral retinal microvascular dysfunction differ.

AB - Microvascular dysfunction is increasingly recognized as an important contributor to ocular and systemic diseases, including diabetic retinopathy, cardiovascular disease, and Alzheimer's disease and related dementias. Elucidating how early microvascular injury contributes to disease pathobiology, and development of sensitive, quantifiable features of microvascular dysfunction, represents a critical frontier for improving risk stratification, enabling earlier diagnosis and guiding targeted interventions. In this context, retinal imaging has emerged as a powerful modality, enabling non-invasive, high-resolution visualization of the microvasculature with color and (ultra)widefield fundus imaging. Rapid technological advances have led to an array of, open-source tools for quantitative extraction of retinal microvascular features. However, progress in this field is hampered by poor comparability between tools, limiting reproducibility and cross-study integration. Differences between tools originate from variations in image processing steps, including vessel segmentation, arteriole–venule classification, optic disc detection, region-of-interest definition, image quality assessment, and the algorithms for metric calculation. In this review we comprehensively compare existing analytical analysis tools for static vessel analyses and delineate how these methodological differences influence the vascular measure quantification. To advance robust, scalable biomarker development, we propose a methodological framework to standardize and harmonize retinal microvascular quantification. This framework can guide future studies in addressing key gaps in literature and in developing imaging biomarkers for clinical use. Critical open questions include whether and when retinal imaging features change during ocular and systemic disease, whether microvascular dysfunction is reversible, how microvascular dysfunction contributes to neurodegeneration, and how central and peripheral retinal microvascular dysfunction differ.

KW - Fundus photography

KW - (ultra)widefield

KW - Color fundus

KW - Microvascular dysfunction

KW - Standardization

KW - Cohort

U2 - 10.1016/j.preteyeres.2026.101461

DO - 10.1016/j.preteyeres.2026.101461

M3 - Article

C2 - 41871655

SN - 1350-9462

VL - 112

JO - Progress in Retinal and Eye Research

JF - Progress in Retinal and Eye Research

M1 - 101461

ER -