Risk, three years earlier.

What it gives you

A risk stratification for one person at one point in time, on a stated horizon. It is built to change screening interval and follow-up priority. It does not replace a diagnostic pathway and it does not replace histopathology.

What it does not

It does not diagnose cancer, and a raised score does not mean cancer is present. It means look sooner, under the protocol that already governs that population. Any deployment that treats the score as a finding is using it outside what it was designed for, and we say so in the contract.

On the three years

Three years is the far edge of the lead time we have observed, not the typical case. A single maximum is close to useless for planning a programme, so what we share under agreement is the distribution.

The number a programme should ask for

Screening populations have low prevalence. At low prevalence a model can have strong sensitivity and specificity and still produce more false positives than true ones, and every false positive is a person walking into an anxious diagnostic pathway.

So the number that decides whether this helps is positive predictive value at the prevalence of the population you are actually screening. Ask us for it early. We will not lead with a headline figure measured somewhere more flattering.

Where the value sits

Not in finding more cancer. In moving finite screening capacity toward the people most likely to benefit, and lengthening intervals safely for those who are not. That is a programme-level decision, so it needs a named clinical owner on your side before it needs a contract.