The Glaucoma Intelligence Platform

One engine
for glaucoma.
Every modality.

Eyelomics turns visual fields, OCT, fundus imaging, and clinical data into a single, reproducible intelligence layer — severity, progression, and responder signals from the same AI engine.

01 / The problem

Glaucoma data is rich.
And wasted.

Every glaucoma visit generates visual fields, OCT, fundus images, IOP, and clinical history. Yet most platforms treat each in isolation — returning raw numbers with no unified severity signal, no trajectory, and no decision support.

Staging varies by clinician, by device, by practice. Clinical trials over-recruit across the severity spectrum, inflating sample size and diluting efficacy signals. The data is there. The intelligence layer isn't.

02 / The platform

Three capabilities,
one engine.

i. Severity

Disease severity, reproducible.

Modality-agnostic severity staging from any high-dimensional ophthalmic dataset. A 30-step granular cluster model maps to four clinical stages — Normal, Early, Moderate, Advanced — validated against gold-standard references.

VFOCTFundus
ii. Progression

Trajectory, phenotyped.

Patients classified into four progression phenotypes — No Progression, Stable, Slow, Rapid — using multi-modal data. Anchored to cluster transitions, not noisy point-to-point measurements, enabling earlier detection of meaningful change.

VF + OCT combined
iii. Response

Response, predicted.

AI-derived patient profiles identify who is likely to respond to a given intervention — enabling trial enrichment, smaller sample sizes, and precision treatment selection across clinical and pharma applications.

VFOCTClinical
03 / Modalities

Built across the
entire data stack.

Visual Fields
Active · Beachhead
52-point sensitivity maps. MD + PSD staging. 30-cluster granular model with stage-delta progression tracking.
OCT
Structural Layer
RNFL thickness, ganglion cell complex, and optic nerve head parameters — structural staging and thinning rates.
Fundus Imaging
Morphological Layer
Cup-to-disc ratio, optic disc hemorrhages, ISNT rule analysis, optic nerve head morphology staging.
Clinical Data
Risk Layer
IOP, central corneal thickness, demographics, family history, medication response — feeding responder prediction models.
Every modality. One severity profile per patient.
04 / Who we serve

Built for the
full ecosystem.

Clinicians & practices

  • Objective staging at point of care
  • Standardized across providers
  • Confident referral decisions
  • Stable longitudinal baseline

Device companies

  • Licensable AI engine for any platform
  • Regulatory-aligned CDS framework
  • Validation documentation included
  • Peer-reviewed scientific foundation

Pharma & trials

  • Severity-based enrollment criteria
  • Responder enrichment
  • Real-world evidence from existing data
  • Endpoints anchored to stage transitions
05 / Scientific foundation

A platform built on
peer-reviewed science.

Siamak Yousefi, PhD

Co-founder · Chief Scientific Officer
Associate Professor, Bascom Palmer Eye Institute, University of Miami
Among the most-published AI researchers in glaucoma and ophthalmic diagnostics. His peer-reviewed frameworks for AI-based severity staging across visual fields and OCT are recognized scientific reference points in the field.

Rehan Ahmed, MD

Co-founder · Chief Executive Officer
Practicing ophthalmologist · ophthalmic AI & clinical strategy
Bridges clinical practice and AI-driven diagnostics. Leads commercial strategy, partnerships, and go-to-market. Hosts a widely followed ophthalmology podcast with direct reach into the eyecare ecosystem.
Selected publications
An objective and easy-to-use glaucoma functional severity staging system based on artificial intelligence
Huang et al. · Journal of Glaucoma · 2022
An AI-enabled system for RNFL thickness damage severity staging
Yousefi et al. · Ophthalmology Science · 2023
06 / Contact

Let's build this together.

i.
Discovery call

30 minutes to explore synergies between Eyelomics and your network or program.

ii.
Live demo

Severity intelligence output demonstrated on de-identified multi-modal data.

iii.
Partnership

Co-development, licensing, or real-world evidence collaboration — we're flexible.