The oculomics paradigm: A comprehensive review

The landscape of optometric practice is currently undergoing a foundational shift, transitioning from a discipline primarily focused on refractive error and localized ocular pathology to a central hub for systemic health surveillance. This evolution is encapsulated in the term oculomics, a field formally institutionalized in 2020 to describe the high-throughput, quantitative evaluation of ocular imaging data sets to derive biomarkers indicative of systemic disease.¹ While the concept of the eye as a window into the body has historical roots spanning over a century, the modern instantiation of oculomics is powered by a triad of technological catalysts: the widespread clinical adoption of high-resolution noninvasive imaging (hardware), the availability of massive, population-scale data sets (big data), and the maturation of deep learning algorithms (software) capable of identifying complex phenotypic patterns imperceptible to human clinicians.²

The clinical significance of this field is underscored by the unique anatomical status of the retina. It provides the only site in the living human body where both the microvasculature and the central nervous system (CNS) tissue can be directly visualized and quantified without invasive procedures.² The retinal blood vessels share structural, physiological, and regulatory similarities with the microvasculature of the heart, brain, and kidneys, while the retinal nerve fibers are direct extensions of CNS axons.³ Consequently, an alteration in retinal architecture often serves as a canary in the coal mine for systemic conditions, manifesting structural and functional changes long before traditional systemic markers or clinical symptoms appear.⁴

Technological foundations and the evolution of oculomics imaging

The rapid expansion of oculomics—evidenced by the publication of over 400 research papers between 2020 and 2025—is primarily driven by the transition from qualitative observation to quantitative, artificial intelligence (AI)-enhanced analysis.⁵ Early oculometric investigations were limited to manual measurements of retinal vascular parameters, such as arteriolar-to-venular ratios and tortuosity.⁶ However, the modern epoch leverages multimodal imaging platforms that provide a multidimensional neurovascular unit evaluation framework.⁷

High-resolution ocular imaging modalities

The efficacy of oculomics relies on the precision of its inputs. The integration of structural optical coherence tomography (OCT) with functional OCT angiography (OCTA) has revolutionized the ability to quantify layer-specific changes.⁵ Swept-source OCT and spectral-domain OCT allow for the measurement of the retinal nerve fiber layer (RNFL) and the ganglion cell-inner plexiform layer (GCIPL), which are critical for monitoring neurodegenerative processes.⁷ Meanwhile, OCTA provides high-resolution visualization of the superficial capillary plexus, deep capillary plexus, and choriocapillaris, enabling the detection of microvascular rarefaction that parallels systemic vascular decay.⁸

Beyond these established modalities, emerging technologies such as snapshot hyperspectral imaging (HSI) are providing novel insights into disease pathophysiology.¹¹ HSI enables the simultaneous collection of spectroscopic and spatial data, identifying typically elusive biomarkers by analyzing the characteristic absorption spectra of specific molecules.¹¹ This technology is particularly relevant for identifying aggregated amyloid-β in the retina, a primary marker for Alzheimer disease (AD), by detecting its unique spectral signature.¹¹

The integration of AI and machine learning

The true power of oculomics lies in the application of AI, particularly deep learning and convolutional neural networks (CNNs), to these imaging data sets.¹² AI algorithms are uniquely suited for oculomics because they can process complex patterns beyond human recognition, such as subtle variations in vessel bifurcation or the texture of the retinal pigment epithelium.¹² These models can accurately predict demographic factors—including age, sex, and smoking status—directly from fundus images, often achieving near accuracy in sex determination.¹²

A significant trend in 2024 and 2025 was the shift from task-specific AI models to large-scale foundation models. These models, such as RETFound, are trained on millions of unlabeled retinal images using self-supervised learning, creating a generalized representation of the eye that can be fine-tuned for diverse diagnostic tasks.² This approach reduces the burden of expert annotation and improves the generalizability of the AI across different patient populations and imaging devices.¹³ The Global RETFound initiative, a consortium involving over 100 study groups across 65 countries, is currently working to produce the most geographically and ethnically diverse ophthalmic imaging data set to date, comprising 100 million eye images.²

Neuro-oculomics: The eye as a proxy for the brain

The anatomical and physiological connection between the retina and the brain provides the foundation for neuro-oculomics.² Because the retinal nerve fibers are part of the CNS, they are susceptible to the same neurodegenerative processes that affect the brain, such as axonal loss, protein aggregation, and inflammation.⁷ Oculomics offers a noninvasive, cost-effective alternative to expensive PET scans or invasive lumbar punctures for the early detection and monitoring of neurodegenerative diseases.¹⁴

AD and cognitive decline

AD is characterized by the accumulation of amyloid-β plaques and phosphorylated tau tangles, which lead to progressive neuronal death. Oculomics research has identified several key biomarkers that correlate with AD pathology. Structural changes, measured via OCT, include significant thinning of the RNFL and the GCIPL, which often reflects cerebral atrophy.⁷ Vascular changes observed via OCTA include decreased retinal perfusion density, reduced vessel density, and an enlarged foveal avascular zone, suggesting that AD has a significant microvascular component.⁸

In 2024, the Granular Neuron-Level Explainer (LAVA) was introduced as a novel explainable AI framework for AD assessment.¹⁵ Unlike previous "black box" models, LAVA can determine the continuum of AD severity by identifying specific retinal imaging features that correspond to different stages of the disease.¹⁵ Furthermore, hyperspectral imaging has successfully identified a spectral signature for amyloid-β in humans with early AD whose disease was confirmed via brain PET scans, providing a path toward a repeatable, noninvasive screening tool.¹¹

Parkinson disease and dopaminergic integrity

In Parkinson disease (PD), the loss of dopaminergic neurons in the brain is paralleled by the loss of dopaminergic cells in the retina.⁷ This loss manifests as reduced contrast sensitivity and structural thinning of the inner retinal layers.⁷ AI models, specifically CNNs such as ResNet-18, have demonstrated an accuracy of0.76 in distinguishing patients with PD from healthy controls based on fundus photography.⁷

A breakthrough biomarker in this area is the retinal age gap, calculated by deep learning models that estimate an individual's biological age from retinal patterns.¹⁸ Research indicates that for each incremental one-year increase in the retinal age gap, there is an independent elevation in the risk of incident PD.¹⁸ This suggests that reti-aging is not merely a marker of senescence but a predictive tool for neurodegenerative risk.¹⁸

Multiple sclerosis and ocular biometry

The application of oculomics to multiple sclerosis (MS) has expanded beyond the monitoring of optic neuritis. Ophthalmic biomarkers such as tear composition, saccadic eye movements, and RNFL thickness are emerging as sensitive tools for evaluating disease progression and therapeutic response.¹⁹ MS is a chronic autoimmune demyelinating disease, and the eye's embryological similarity to the brain makes it ideal for assessing neurovascular changes in the CNS.¹⁹ Machine learning algorithms integrated with eye-tracking technology are now capable of discovering new parameters of pathology, potentially identifying MS before the onset of clinical symptoms.¹⁹

Systemic microvascular surveillance: Cardiovascular and renal health

The retinal microvasculature is an ideal noninvasive surrogate for the systemic microvasculature. Conditions such as hypertension, diabetes, and chronic kidney disease (CKD) leave distinct signatures on the retinal vessels, which can be quantified through oculomics to provide a holistic view of a patient’s cardiovascular-kidney-metabolic health.⁹

Cardiovascular risk stratification

Oculomics can predict major adverse cardiovascular events by analyzing the caliber, geometry, and tortuosity of retinal vessels.¹² Research from the Multi-Ethnic Study of Atherosclerosis has shown that narrower retinal arteriolar diameters and wider venular diameters are associated with impaired vascular function and an increased risk of cardiovascular disease (CVD).⁹ AI models can now estimate coronary artery calcium scores directly from fundus images, offering a radiation-free method for heart disease screening.⁴

In 2025, real-world studies applying radiomic features to multimodal retinal images achieved near-perfect accuracy in predicting cardiovascular risk in type 1 diabetic populations.⁵ These models integrate parameters from OCT and OCTA, such as capillary density and retinal layer thickness, to provide a more nuanced risk stratification than traditional clinical markers alone.⁵

Renal-ocular cross-talk and CKD

The anatomical similarities between the retinal capillary network and the renal glomerular and tubular systems create a "renal-ocular" axis.⁹ Biomarkers of kidney tubule injury, such as kidney injury molecule-1 (KIM-1) and soluble urokinase plasminogen activator receptor, have been found to correlate with retinal microvascular changes in individuals who do not yet have overt CKD, diabetes, or CVD.⁹ Specifically, higher levels of KIM-1 are associated with narrower retinal arterioles, reflecting systemic microvascular rarefaction.⁹

This finding is clinically significant because it suggests that tubular perturbations can manifest in the eye well before traditional renal markers, such as EGFR or albumin-to-creatinine ratio, detect overt disease.⁹ Oculomics thus enables the detection of subclinical tubular-vascular cross-talk, allowing for intervention before irreversible organ damage occurs.⁹

Metabolic syndrome and hepatobiliary health

Oculomics also provides a window into metabolic health and systemic inflammation. The RetiPhenoAge model, a cutting-edge biological aging marker, prognosticates morbidity and mortality by discerning retinal patterns linked to fluctuations in blood biomarkers for liver function, inflammation, and energy metabolism.¹⁸ A larger retinal age gap has been associated with an increased risk of developing metabolic syndrome, including abdominal obesity and hyperglycemia.¹⁸

Furthermore, AI-based analysis of slit-lamp and fundus images has demonstrated the potential to detect hepatobiliary disorders, such as liver cancer and cirrhosis.⁴ Ocular signs such as scleral icterus or corneal copper deposition (Kayser-Fleischer rings) have long been used to diagnose liver disease, but AI can quantify more subtle vascular and structural alterations that are missed by human observers.²²

Clinical implementation: The "healthcare from the eye" framework

As the field of oculomics matures, there is a pressing need to integrate these scientific advances into routine clinical care. In January 2026 a comprehensive framework titled "Healthcare From the Eye" was proposed by the leadauthor, Robert N. Weinrebto guide the ethical and scalable integration of oculomics into coordinated health care delivery models.²³

Four foundation principles of oculomics integration

The 2026 framework identifies 4 essential domains for the successful adoption of oculomics-enabled programs, as follows²³:

• Access to affordable diagnostics: Scalability requires the deployment of high-quality, often handheld or automated, imaging devices in both primary care and optometric settings.²³
• Responsible data governance and AI development: Ensuring data privacy, ethical AI use, and the development of transparent, explainable algorithms is critical for clinician and patient trust.⁵
• Patient-care coordination and integrated workflows: Oculomics results must be seamlessly integrated into electronic health records and follow-up referral networks to ensure that eye-derived insights lead to systemic health interventions.⁵
• Sustainable economic and operational models: For oculomics to thrive, there must be a shift from reactive, episodic management to proactive, longitudinal care, supported by sustainable reimbursement pathways.²³

Practice management, billing, and coding in 2025-2026

The professionalization of oculomics is accompanied by significant changes in the regulatory and billing landscape. In 2025, several critical updates impacted optometric billing, reflecting the expanding role of doctors of optometry in systemic health care.²⁴

A major milestone was the creation of Current Procedural Terminology (CPT) code 92137 to report computerized ophthalmic diagnostic imaging of the posterior segment (retina) that includes OCTA.²⁵ This code requires both traditional OCT and OCTA to be performed and interpreted on the same day.²⁵ Additionally, the descriptors for codes 92132 through 92134 were updated to "ocular coherence tomography" to better reflect the modern technologies used in the field.²⁴

Optometrists must prioritize clear documentation of medical necessity to avoid audits and denials.²⁴ In 2025, insurance companies increased scrutiny on the frequency of eye imaging, requiring that tests be justified by evidence of disease progression or systemic risk assessment.²⁹ High-performing diagnosis codes for medical necessity include E11.3X (diabetic retinopathy) and H40.11X (glaucoma), while routine exam codes (Z01.00) are generally noncovered for medical imaging.²⁹

Ethical considerations and the future of oculomics

The ability to detect chronic, often incurable, systemic diseases through a routine eye exam raises profound ethical questions. The Wilson and Jungner criteria for population-based screening programs stipulate that a test should be cost-effective, noninvasive, and implemented only if an effective treatment is available.³⁰ While early detection of AD or PD allows for lifestyle modifications and enrollment in clinical trials for new disease-modifying therapies (eg, lecanemab [Leqembi; Eisai Co and Biogen, Inc], bepranemab [UCB]), clinicians must balance the benefit of early diagnosis against the potential psychological harm to patients and their families.³¹

The future of the field will likely see the integration of generative AI and large language models to enhance clinical decision-making.³² These "intelligent AI agents" could synthesize multimodal data—including genetic information, social determinants of health, and retinal images—to provide personalized health forecasts.³² Furthermore, the development of federated learning allows AI models to be trained across multiple institutions without sharing raw patient data, protecting privacy while maintaining the benefits of large-scale data sets.³³

Summary and strategic recommendations for the forward-thinking OD

Oculomics is not merely a technological advancement but a fundamental shift in the identity of the optometric profession.¹ As the primary point of contact for eye care, optometrists are uniquely positioned to serve as the front line for systemic disease screening.¹ To prepare for this future, practitioners should consider the following:

• Optimizing imaging protocols: Ensure that imaging equipment can produce high-quality, repeatable scans suitable for AI analysis. The value of data grows exponentially when it is linked to longitudinal patient records.⁵
• Expanding the scope of the medical exam: Move beyond vision to incorporate systemic risk factors—such as blood pressure, body mass index, and metabolic history—into the interpretation of retinal findings.¹²
• Strengthening interdisciplinary networks: Build formal referral relationships with cardiology, nephrology, and neurology providers. Clear communication pathways are essential to translate an eye scan into a life-saving systemic intervention.¹
• Adopting modern billing practices: Stay informed on coding changes (eg, CPT 92137) and quality reporting requirements (merit-based incentive payment systems [MIPS]) to ensure that the practice is reimbursed for its expanding role in health care.²⁴

By embracing oculomics, the optometric community can transition from being vision specialists to being integral members of a proactive, personalized, and multidisciplinary health care team.¹ The eye is no longer just for seeing; it is a gateway to the entire human body, offering a noninvasive path toward early detection and the prevention of chronic disease on a global scale.¹²

References

1. Breeman D. As the field of oculomics gains traction, role of ODs comes into focus. Vision Monday. December 12, 2025. Accessed April 9, 2026. https://www.visionmonday.com/scene-and-heard/todays-read/article/as-the-field-of-oculomics-gains-traction-role-of-ods-comes-into-focus/

2. Professor Pearse Keane on using eye health to detect dementia. Brain Sciences. Accessed April 9, 2026. https://www.ucl.ac.uk/brain-sciences/research/dementia-ucl-priority/professor-pearse-keane-using-eye-health-detect-dementia

3. Wang T, Feng Z, Yang J, Liu C, Zhao H, Huang S. Stroke-induced retinal dysfunction: mechanisms, diagnostics, and future directions. Aging Dis. Published online September 29, 2025. doi:10.14336/AD.2025.0451

4. Wang J, Wang YX, Zeng D, et al. Artificial intelligence-enhanced retinal imaging as a biomarker for systemic diseases. Theranostics. 2025;15(8):3223-3233. doi:10.7150/thno.100786

5. Lewis K. Top oculomics breakthroughs in 2025: what they mean for optometry. Optometry411. October 28, 2025. Accessed April 9, 2026. https://optometry.industry411.com/top-oculomics-breakthroughs-in-2025-what-they-mean-for-optometry/

6. Chandra M, Singh J, Dubey G, Ansari A, Kunwar A, Jha A. Oculomics: an artificial intelligence-driven framework for retinal imaging in systemic disease assessment. Euro J Pharma Med Res. 2026;13(2):575-583. doi:10.5281/zenodo.18604531

7. Rahman F, Rahman A, Talha M, Irshad NUN, Imran SB. AI-powered oculomics for early diagnosis of neurodegenerative diseases through retinal microvasculature analysis. Ann Med Surg (Lond). 2025;88(1):1036–1037. doi:10.1097/MS9.0000000000004391

8. Yu X, Zhang H, Tan H, et al. Retinal microvascular and structural changes in chronic coronary artery disease and ischemic stroke: a comparative OCTA and OCT study. Research Square. Published online December 8,2025.doi:10.21203/rs.3.rs-8014609/v1

9. Cohen-Hagai K. Tubular biomarkers and retinal microvasculature: illuminating early cardiovascular-kidney-metabolic pathways beyond the glomerulus. Kidney360. 2025;6(12):2054–2056. doi:10.34067/KID.0000001002

10. Karpecki PM. A new, great indicator. Review of Optometry. October 15, 2025.Accessed April 9, 2026. https://www.reviewofoptometry.com/article/a-new-great-indicator

11. Saeed A, Hadoux X, van Wijngaarden P. Hyperspectral retinal imaging biomarkers of ocular and systemic diseases. Eye (Lond). 2025;39(4):667-672. doi:10.1038/s41433-024-03135-9

12. Conem E, Gill A, Habaya M, et al. Systemic disease screening using deep learning analysis of retinal images (oculomics): a review. OxJournal. September 19, 2025. Accessed April 9, 2026. https://www.oxjournal.org/systemic-disease-screening-using-deep-learning-analysis-of-retinal-images/

13. Zhou Y, Chia MA, Wagner SK, et al. A foundation model for generalizable disease detection from retinal images. Nature. 2023;622(7981):156–163. doi:10.1038/s41586-023-06555-x

14. Oculomics: a non-invasive window into systemic diseases. ReachMD. June 13, 2025. Accessed April 9, 2026. https://reachmd.com/news/oculomics-a-non-invasive-window-into-systemic-diseases/2475136/

15. Tan TF, Dai P, Zhang X, et al. Explainable artificial intelligence in ophthalmology. Curr Opin Ophthalmol. 2023;34(5):422-430. doi:10.1097/ICU.0000000000000983

16. Zhang J, Lin S, Cheng T, et al. RETFound-enhanced community-based fundus disease screening: real-world evidence and decision curve analysis. NPJ Digit Med. 2024;7(1):108. doi:10.1038/s41746-024-01109-5

17. Zou K, Goh JHL, Zhou Y, et al. FusionFM: fusing eye-specific foundational models for optimized ophthalmic diagnosis. arXiv. Published online August 15, 2025. Accessed April 15, 2026.
    doi:10.48550/arXiv.2508.11721

18. Zhang D, Li N, Li F. Advances in ocular aging: combining deep learning, imaging, and liquid biopsy biomarkers. Front Med (Lausanne). 2025;12:1591936. doi:10.3389/fmed.2025.1591936

19. Suh A, Hampel G, Vinjamuri A, et al. Oculomics analysis in multiple sclerosis: current ophthalmic clinical and imaging biomarkers. Eye (Lond). 2024;38(14):2701–2710. doi:10.1038/s41433-024-03132-y

20. Ahmadi A, Gorji H, Shashikant S. Associations of biomarkers of kidney tubule health with retinal microvascular signs: the multi-ethnic study of atherosclerosis. Kidney360. 2025;6(12):2157–2165. doi:10.34067/KID.0000000970

21. Elsayed SA, Mounir A, Mostafa EM, Saif DS, Mounir O. The correlation between retinal microvascular changes by optical coherence tomography angiography and nailfold capillaroscopic findings in patients with systemic sclerosis. J Rheum Dis. 2025;32(3):198-210. doi:10.4078/jrd.2024.0124

22. Parmar UPS, Morya AK, Gupta PC, Arora A, Verma N. Role of artificial intelligence-based ocular biomarkers in hepatobiliary diseases: a scoping review. World J Hepatol. 2025;17(8):109801. doi:10.4254/wjh.v17.i8.109801

23. Weinreb RN, Keane PA, Cooley A, et al. A framework for healthcare from the eye: oculomics as a powerful window to systemic health. Ophthalmology. Published online January 21, 2026.doi:10.1016/j.ophtha.2026.01.018

24. Johnson MP. Coding and reimbursement: 2026 update. Rev Ophthalmol. January 13, 2026. Accessed April 9, 2026. https://www.reviewofophthalmology.com/article/coding-and-reimbursement-2026-update

25. Access recent quarter 2025 coding updates curated by AAOE experts. American Academy of Ophthalmology. December 4, 2025. Accessed April 9, 2026. https://www.aao.org/practice-management/news-detail/access-fall-2025-coding-updates-curated-by-aaoe

26. Optometry billing in 2025: what you must know. medisys. June 18, 2025. Accessed April 9, 2026. https://www.medisysdata.com/blog/optometry-billing-in-2025-what-you-must-know/

27. Coding and Reimbursement Committee. 2025 code changes: what doctors of optometry need to know. American Optometric Association. November 13, 2024. Accessed April 9, 2026. https://www.aoa.org/news/practice-management/billing-and-coding/2025-code-changes-what-doctors-of-optometry-need-to-know

28. Optometry billing & coding: 2025 year in review + 2026 updates. RevolutionEHR. January 8, 2025. Accessed April 9, 2026. https://www.revolutionehr.com/blogs/optometry-billing-coding-review

29. Optometry billing and coding cheat sheet for hospital (2026). MBW RCM. Accessed April 9, 2026. https://www.mbwrcm.com/the-revenue-cycle-blog/optometry-billing-and-coding-cheat-sheet

30. Cheng H, Sarnat JA, Walker DI, et al. Oculomics meets exposomics: a roadmap for applying multi-modal ocular biomarkers in precision environmental health research. Exposome. 2025;5(1):osaf013. doi:10.1093/exposome/osaf013

31. UCB presents key data from Alzheimer's and Parkinson's disease research programs at AD/PD 2025. News release. UCB. April 1, 2025. Accessed April 9, 2026. https://www.ucb.com/newsroom/press-releases/article/ucb-presents-key-data-from-alzheimer-s-and-parkinson-s-disease-research-programs-at-adpd-2025

32. Shi D, Zhang W, Chen X, et al. EyeFound: a multimodal generalist foundation model for ophthalmic imaging. arXiv. Updated May 22, 2024. Accessed April 15, 2026. doi:10.48550/arXiv.2405.11338

33. Rehman MU, Masip D. Artificial intelligence in retinal imaging for early Alzheimer's disease detection: a systematic review. Research Square. Published online June 5,2025. doi:10.21203/rs.3.rs-6761315/v1