The ocular findings fund the network. The network proves the biomarkers. Neither half stands alone — and nobody else is building both.
Decisions already made — do not reopen without new data (23 Aug 2026): pilot stays observational, no research blood or urine; CKD in the undiagnosed population is the first systemic proof-of-platform; labels are records-based and the model locks before pivotal validation; Texas first, 47 sites the first state-scale target; enrolment up to 5,000 under the current amendment strategy; protocol stays broad across ocular + cardiovascular + neuro/CNS + renal + metabolic; two fuel tanks — the retina practice as bridge capital plus the network's own revenue; no new equity unless it buys an unreachable inflection; no NIH dependency; $300M is the credible strategic target, $1B an earned option.
Critical path — the six workstreams that gate everything
Before site #2 signs — current highlights are now false
▸
Release the reference-set RFQ
Gates the reading QS → payers AND sponsors
▸
Ascertainment gate on first 100–200
The largest unmeasured assumption, converted to arithmetic
▸
This gate costs nothing — it runs on data the pilot produces anyway. Its output is also two of the five KPIs below.
Five Site-01 KPIs
Each one replaces a load-bearing assumption in the financial model
▸
1Scans/month — from platform invoices. Breakeven vs the retainer is ~27/mo; the model assumes 150.
2OD net position — platform fees minus $400. Net-payer means Retina IQ must earn its price; ask Dr. K directly whether it does.
3MA members/year — count one full week of exams: total × %65+ × %MA, then annualize. The model assumes 297; it drives 61% of modeled revenue.
4Ascertainment yield — positives per 1,000 enrolled, from the gate. Sets the true enrollment target for S1 and model lock.
5Re-contact response rate — % responding to the health-status check. The multiplier on the one differentiator that does not erode; never measured anywhere.
Deploy the Azure thin build
07_Engineering zip · ~$50–120/mo · HIPAA BAA is automatic
▸
If the account is created under a Microsoft 365 tenant you already own for spectocularhealth.com email, use that tenant — one identity system. Otherwise the signup creates a fresh tenant, which is fine for the pilot.
Commercial discovery in parallel — payer and life sciences
Payer is no longer required to be the first validation · new in v2.0
▸
The enrollment arithmetic
Current working hypotheses — each is replaced by observed data from the ascertainment gate and the Site 01 KPIs, not by re-modelling.
~160
positives → S1 go/no-go (3,200–4,600 enrolled) set by observed yield + records fulfilment
300–500
positives → model LOCK (~6–8K enrolled) set by S1 effect size + label quality
~11K/yr
capacity of 47 TX sites at 5/site/week set by actual site throughput
Before lock
payer + sponsor discovery starts ahead of biomarker completion set by buyer willingness to pay
The ladder
$10Mnow → mid-2027the machine exists
2–5 sites live on the IRB workflow · ascertainment ≥40 positives/1,000 confirmed
Site economics measured; payer and pharma discovery underway
Reading quality system operational with a first inter-reader statistic
Early execution value — not yet a valuation proof point
$100Mlate 2027 → 2028signal + someone pays
47 TX sites · S1 positive in the undiagnosed stratum · ~6–8K enrolled
Algorithm locked, versioned, frozen before pivotal validation
First meaningful payer or life-sciences contract
Scientific risk falls and external willingness to pay is demonstrated
$300M2029regional platform
400–1,000 sites · pivotal underway · second disease showing signal
The credible strategic target if the platform repeats
$1B2030–31national platform
National distribution · regulatory authorisation in hand
$100M+ high-quality revenue or a credible trajectory to it
An earned upside option — platform valuation, not a forecast
The second-disease rule
One successful CKD model proves a potentially valuable CKD product. A second independently successful cardiovascular, neurologic/CNS, metabolic or other systemic model is what proves the pipeline — and the pipeline is the company.
When you feel wayward
SpectOcular is not:
A raw-data brokerage. Retrospective images alone are not the moat.
A standalone reading service. Retina IQ is network utility, quality infrastructure and an ocular-product lane — not the thesis.
A payer company or a pharma CRO. Those customers finance and monetise the network; the longitudinal retinal intelligence platform is the asset.
A single-diagnostic company. CKD authorisation proves the pipeline; it does not exhaust it.
Optimised around a modelled payer price before the first payer contract exists.
Scaling site count faster than quality, ascertainment, re-contact and unit economics can be measured.
Kill criteria — pre-declared, still binding
ODs will not pay platform fees at three sites → Retina IQ is not the stand-alone economic engine; payer/life-sciences revenue must carry expansion.
S1 flat and a second independent POC flat → the systemic biomarker pipeline has not demonstrated repeatable signal.
Passive payer economics <$20/member and completed outreach/evidence economics <$75 → the $1B payer-funded option is impaired; optimise for the smaller outcome.
Consent, FMV, privacy, IRB or site-contract structure fails counsel/regulatory review → stop scaling until remediated.
Records fulfilment or re-contact too weak to produce a longitudinal cohort → fix the cohort architecture before adding sites.
Next 90 days — priority order
Complete the Sterling amendment process; use only the currently approved consent until approval lands.
Agreement v2.0 + urgent-notification SOP in force before Site #2. done 24 Aug
Release the reference-set RFQ and obtain the first reader-quality statistic. deferred by decision until 20+ sites
Run ascertainment + re-contact on the first 100–200 and measure the five Site-01 KPIs.
Deploy only the Azure infrastructure those workflows actually need.
Onboard Sites #2–5 only after the Site 01 workflow is reproducible.
Prepare payer discovery materials around completed evidence encounters — not $1,600 bounties.
Prepare a separate life-sciences one-pager; begin sponsor/CRO discovery once reading quality and early cohort metrics are credible.
Pre-specify S1 before examining the signal; preserve independent validation and comparator discipline.
