The science
Oxygen. Steady diffusion with first-order uptake, 0 = D∇²c − k c (D = 2000 µm²/s; k = 0.003 /s normal tissue, 0.18 /s tumour). Perfused vessel cells hold c = 1. The penetration depth is √(D/k) ≈ 105 µm in the tumour. Cells below c = 0.25 are hypoxic; after 12 h below c = 0.06 a tumour cell dies (necrosis, irreversible, no further uptake).
VEGF. 0 = D_V∇²V − λV + q(1 − c/c_h) for hypoxic live cells (D_V = 60 µm²/s, half-life about 31 min, decay length √(D_V/λ) = 400 µm). The anti-VEGF slider multiplies λ by (1 + 9 s), a simple binding sink.
Sprouts. Each tip (12 µm/h) turns toward ∇V with a receptor-saturating response Δθ = χ S sin(φ), S = min(1, 40|∇V|/(V+K)), plus random wander (Anderson & Chaplain 1998 biased random walk). Branching probability is proportional to V/(V+K). A tip that touches a foreign vessel closes a loop (anastomosis); only then do the sprout and its parent chain carry blood and supply oxygen. Blind-ended sprouts stay unperfused.
Krogh limit. For tissue with zero-order uptake M around a capillary of radius rv, c(r) = c0 + M/(4D)[r² − rv² − 2R² ln(r/rv)]; tissue farther than RK from a vessel starves, which is why a growing tumour must recruit vessels.
Simplifications. 2-D slice of tissue; fixed tumour size (no growth or division); VEGF is quasi-steady (it equilibrates within hours, a tick is 30 min); one vessel type, no blood flow, pericytes, Notch/Dll4 tip selection or matrix; constants are order-of-magnitude literature values, so outcomes are qualitative. Educational model only, not medical advice.
Independent checks. (1) O2 slab solver vs cosh(φ(1−x/L))/cosh φ; (2) VEGF decay length vs √(D/λ); (3) radial Krogh finite differences vs the closed form; (4) VEGF mass balance (secretion = decay); (5) finite fields, maximum principle 0 ≤ c ≤ 1 and seed determinism.