Report on previous research and research programme - Microvascular health endpoints from retinal Doppler holography

The retina microvascular bed is routinely accessible non-invasively, enabling acquisitions with beat-resolved hemodynamics in arteries and veins. Doppler holography already makes these measurements practical: a short eye exam yields arterial and venous velocity waveforms, enabled by real-time digital hologram rendering. The remaining translational barrier is standardisation: converting this exam into auditable, comparable, trial-ready endpoints and a microvascular health report whose interpretation is stable across operators, sites, devices, and software versions. This programme is infrastructure-first. It establishes a shared reference configuration across partner centres and a deterministic end-to-end pipeline - reconstruction, quality control, endpoint extraction, and reporting—with strict traceability (versioned releases, processing fingerprints, provenance-rich metadata). Comparability is enforced by design at admission rather than repaired retrospectively: universal calibration, public benchmark packs (“golden runs”), and tolerance-banded conformance scorecards that every hardware or software change must pass before entering multi-centre studies. Outputs follow a frozen reporting contract (one-page report plus machine-readable exports) backed by a locked endpoint registry. Endpoints are tiered by construction. Tier 1 is trial-first and transportable: velocity-based arterial and venous waveform descriptors and artery-to-vein transmission signatures reported only when quality gates pass and always with coverage and visit-level uncertainty. Tier 2 adds mechanistic depth—such as shear-related indices, Womersley-derived quantities, and impedance- or compliance-like descriptors—only under explicit feasibility, conditioning, and validity gates, with calibrated uncertainty and mandatory fallback to Tier 1 to prevent fragile modelling from contaminating trial-facing claims. The proving ground is glaucoma and ocular hypertension. International deployments are operational, and early Tier~1 metrics shifts are already observable. Near-term (2026--2028) deliverables are a public benchmark pack and conformance suite, frozen trial-grade releases, quantified repeatability and inter-site transportability across eight sites, and validation of Tier 1 primary endpoints under a pre-specified analysis plan, including sensitivity-to-change under intraocular pressure lowering. Beyond this first clinical axis, the same standardised framework enables reserve testing and choroidal phenotyping in ophthalmology, and opens a path to universal microvascular health markers relevant to systemic disease (diabetes, cardiometabolic conditions, hypertension, neurovascular dysfunction), where fast, non-invasive microvascular readouts are needed for screening, risk stratification, and pharmacodynamic monitoring in large-scale trials. In short, the programme turns a seconds-long retinal scan into an endpoint standard: benchmarked, auditable, and comparable by construction across centres and over time.