01 / THE TEAM
Expertise that connects.
Founders spanning machine learning, retinal disease, and clinical innovation. Different disciplines, united by questions that matter to patients.
A Duke spinout · Precision ophthalmology
See earlier.Understand more.
Turning retinal imaging into predictive insight—for more precise research, better-informed trials, and a future of earlier, individualized care.
Clinically grounded. Research driven. Built for collaboration.
The eye is only
the beginning.
the beginning.
Ophthalmic intelligence / A new perspective
From retinal images to research insight
Beyond what a scan can show.Deep learningOphthalmic imagingClinical translation
Conceptual visualizationDuke-rooted scienceAcademic origins. Translational vision.
OCT-first intelligenceBuilding on ophthalmic imaging.
Peer-reviewed researchA published scientific foundation.
Partner-mindedScience designed to move together.
Our perspective
A clearer view of disease.
A different future for care.
HealthLadder.AI brings together clinical insight, machine learning, and ophthalmic research to understand not only where disease is—but where it may be going.
02 / THE SCIENCE
Insight beyond the image.
A research foundation in predictive algorithms and imaging biomarkers, with an initial focus on progression from intermediate AMD to geographic atrophy.
03 / THE CLINICAL LENS
Built around the real question.
Who may progress? When might it happen? How can better evidence inform clinical trials and the development of more individualized treatment?
The intelligence layer
From image
to foresight.
Our development approach connects ophthalmic imaging, predictive modeling, and clinical interpretation. The goal: useful insight that can be evaluated in real research workflows.
01
Start with the scan.
Use OCT imaging to study retinal structure, with multimodal research opportunities across fundus images and clinical measures.
02
Learn the patterns.
Explore signals associated with disease progression through multi-scan modeling and attention-based interpretation.
03
Ask a better clinical question.
Evaluate progression risk, trial-enrichment strategies, and the potential for earlier intervention research.
A familiar image. A deeper question.
Retinal imaging provides the starting point for investigating how structural patterns relate to future disease progression.
Conceptual illustration. Not patient data, a product interface, or live AI output.
Imaging-informed developmentClinical relevance firstEvidence-led validation
One scientific foundation. Shared opportunity.
Better questions.
More meaningful progress.
Connecting the people who understand disease with those developing the next generation of care.
Understand progression.
Investigate earlier possibilities.
We are advancing research that connects retinal imaging to clinically relevant questions about progression, timing, and patient selection.
Examine the evidenceProgression-risk research
Investigate which eyes with intermediate AMD may develop geographic atrophy.
Clinically meaningful timing
Study lead time and the distinction between early disease detection and true prediction.
Evidence that travels
Prioritize independent testing, device-aware evaluation, and subgroup analysis.
Move from broad cohorts
to more precise research.
Explore predictive biomarkers to inform trial-enrichment strategies and investigate treatment response, dose, and interval questions in therapeutic development.
Explore collaboration modelsTrial-enrichment strategies
Evaluate whether imaging-derived risk can identify participants relevant to a study’s progression endpoint.
Earlier-intervention research
Align prediction horizons with a therapeutic program’s biological rationale and clinical objectives.
Biomarker co-development
Connect structural and functional outcomes to the development of more individualized therapies.
Scientific depth.
Multiple paths to strategic value.
A clinician-informed AI approach with opportunities spanning sponsored validation, algorithm licensing, and development partnerships—plus selective investment and acquisition discussions.
Explore the strategic opportunityA focused starting point
Lead with a defined question in AMD progression rather than an undifferentiated AI platform.
Milestone-based development
Link expansion to validation results, data permissions, deployment feasibility, and clinical relevance.
Complementary capabilities
Bring together an AI team, a research and IP foundation, and clinical translation expertise.
Clinical research priorities
Lead focus: Intermediate AMD → GAAMD biomarkers & fibrosisTreatment-response researchLong-term vision: Oculomics
Research priorities and future directions—not a list of approved or commercially available indications.
Our scientific foundation
Ambitious by design.
Grounded in evidence.
Our approach starts with published science—and advances through careful, use-case-specific evaluation.
A retrospective study investigated whether volumetric OCT could predict progression from intermediate AMD to geographic atrophy over approximately one year.1
0.94
AUROC in the development cohort
Five-fold cross-validation
95% confidence interval: 0.92–0.95
95% confidence interval: 0.92–0.95
417
Patients in the published study
Across one development cohort
and two independent validation cohorts
and two independent validation cohorts
1year
The principal prediction horizon
Progression assessed within 13 months
of the index scan
of the index scan
These are retrospective research results, not prospective clinical-outcome claims or guaranteed product performance. AUROC measures discrimination; it is not the percentage of patients correctly classified.
Read the publication (opens in a new tab)Understand the study—and the next validation questions
| Cohort | Patients | Role in the study |
|---|---|---|
| AREDS2 A2A | 316 | Development and cross-validation |
| Routine-care cohort 2 | 53 | Independent validation |
| Routine-care cohort 3 | 48 | Independent validation |
The validation cohorts used Heidelberg Spectralis images; the development cohort used Bioptigen images. Validation AUROCs differed across cohorts: 0.94 and 0.84, respectively.1
What comes next: prespecified, blinded studies that test performance in the intended population, distinguish baseline atrophy markers from future progression, and evaluate sensitivity across clinically relevant lead times.
Why it matters: a promising retrospective model is a foundation for further evaluation—not evidence that a particular treatment, dosing strategy, or deployment improves patient outcomes.
1. Dow ER, Jeong HK, Katz EA, et al. Prediction of near-term geographic atrophy from SD-OCT. JAMA Ophthalmology. 2023;141(11):1052–1061. doi:10.1001/jamaophthalmol.2023.4659 (opens in a new tab)
Independent evaluationTransparent study limitationsClinical relevance before scale
The people behind the perspective
Deep science.
Human purpose.
A founding team connecting artificial intelligence, ophthalmic research, and the work of translating ideas into meaningful clinical progress.
Co-founder / AI & machine learning
Larry Carin
PhD
Connecting advanced machine learning with the challenges of building useful, clinically relevant AI.
Co-founder / Ophthalmic clinical science
Eleonora “Nora” Lad
MD, PhD
Bringing retinal disease, imaging biomarkers, and therapeutic-development questions into focus.
Co-founder / Translation & strategy
Nandan Lad
MD, PhD
Uniting clinical innovation, company building, and strategic collaboration to move research forward.
HealthLadder.AI is an independent Duke spinout. Academic affiliations describe the team’s scientific roots and do not imply institutional endorsement.
For partners & investors
A shared vision.
Multiple ways forward.
Explore the fit between your capabilities and ours—from focused validation and co-development to licensing, strategic investment, and acquisition opportunities.
01 / PHARMA & BIOTECH
Develop with
greater precision.
Explore trial enrichment, predictive biomarkers, and research into earlier intervention and individualized treatment.
02 / IMAGING & TECHNOLOGY
Extend the value
of every scan.
Investigate device-specific validation, research workflow integration, and routes from imaging to predictive insight.
03 / CAPITAL & STRATEGY
Build the
next chapter.
Evaluate investment, licensing, and strategic transactions around a complementary team, research foundation, and clinical vision.
Progress with clarity. A productive partnership starts with a defined clinical question, documented rights, measurable milestones, and a shared decision about what to do next.
A clear perspective
Good partnerships
start with
good questions.
How we think about the science, the opportunity, and the path ahead.
What is HealthLadder’s initial clinical focus?
Our initial focus is predictive research in age-related macular degeneration, particularly progression from intermediate AMD to geographic atrophy. Broader research directions include imaging biomarkers, fibrosis, and treatment response. Oculomics—the study of ocular features as signals of wider health—is a longer-term direction, not a claim of established systemic-disease detection.
Are these tools available for routine patient care?
This website describes research and development activities, not an offer of an approved diagnostic or treatment product. Published findings do not establish authorization for routine clinical use. Any deployment requires an appropriate intended use, study approvals, and applicable regulatory review. The visualizations here are illustrative, not clinical outputs.
What could a collaboration look like?
Potential structures include sponsored evaluation, algorithm licensing, biomarker co-development, and research integration. The right model depends on the clinical question, evidence requirements, data permissions, and each party’s responsibilities. Investment and acquisition opportunities can be evaluated separately where there is a strong strategic fit.
Are Duke data or intellectual property automatically included?
No. University research, institutional datasets, and company capabilities are distinct. Data access, licensing scope, publication rights, and any transfer of intellectual property must be established in the relevant agreements and institutional approvals. A relationship with HealthLadder does not itself confer access to Duke data or technology.
The next step in intelligent healthcare
A clearer future.
Built together.
Bring scientific possibility and clinical purpose into the same field of view.
Explore strategic opportunities