Tumour growth prediction
The question isn't how big.
It's how fast.
For a slow-growing brain tumour, one scan tells you almost nothing on its own. What changes the plan is the trajectory, is it growing, how quickly, and when will it cross the line where you'd act.
Growth tracking
Turn scattered scans
into a curve
Using a multiomics approach, with imaging data (radiomics), you can outline the tumour on every scan a patient has had, fit a growth model across them, and those loose dots become a projected line. Amalgamate this with genetics and molecular data, and we are able to build a robust prediction model. The fast-growers stand out sooner, and the stable ones can be left alone instead of pushed into surgery or another round of scans.
- A real 3D volume on each MRI, not a diameter
- A growth rate with a confidence band, from the patient's own history
- A rough time-to-threshold, to set sensible scan intervals
The condition we started with
Vestibular schwannoma
our worked examplePlenty of brain tumours are slow and get watched rather than treated straight away. Surgery or radiosurgery is held back for the ones that grow or start pressing on something. That makes measuring the same thing the same way, scan after scan, the part that matters most.
We started with the vestibular schwannoma: a benign tumour on the hearing and balance nerve, and a textbook watch-and-wait problem. Measuring it by volume and projecting its growth flags the fast ones earlier and spares everyone else needless intervention.
Research & development work, not a medical device or a substitute for clinical judgement.
Request a demo
Twenty minutes is usually enough to work out whether we can help. Tell us about your service.