Tumour Modelling Lab

Tumour modelling on radiographs

Mark a starting region and watch the simulated density change over time.

Reaction–diffusion model

Axial brain MRI

526 × 555 px · Demo image

Axial contrast-enhanced brain MRI

Click the image to move the seed.

Week 12

Modelled tumour burden

Relative to the starting field

Untreated With treatment

Week 12 · Untreated

—Relative burden
—Area with c ≥ 0.20
How the model works
∂c/∂t = D∇²c + ρc(1 − c) − k(t)c

The point seed begins as a Gaussian density field. An imported mask starts at a density of 0.78 inside the marked region. The model spreads and grows that field on a rectangular two-dimensional grid with a zero-flux boundary. Treatment reduces density during the selected weeks.

The image provides the background and the locations you select. The model uses a single spread coefficient throughout the field. The burden curve sums modelled density over the plane. The contour traces the selected threshold and the area counts cells at or above it.

The solver uses an exact update for local growth and treatment with implicit finite differences for spread. Model units give the longer image edge a length of 144 units. Calibration sets millimetres from a known distance or DICOM pixel spacing.