Medicine & Evolutionary Biology

Treating Cancer Like an Evolving Ecosystem: Why Timing Beats Dosage

For decades, the answer to a stubborn tumor has been a stronger dose. New models suggest the better answer may be to switch drugs sooner, before the cancer has time to adapt.

The paradox of maximum dose

Conventional cancer therapy is designed around a straightforward intuition: kill as many tumor cells as possible, as fast as possible. High-dose regimens often shrink a tumor quickly. But that same intensity is also the strongest evolutionary pressure on the surviving cells. The few cancer cells that happen to carry a mutation letting them survive the drug are now the only ones left to reproduce — and they pass that resistance on.

In population genetics terms, maximum dose is an efficient weed-killer that is also an efficient breeding program for the survivors. Resistance is not a bug in the treatment. It is the expected, almost mechanical, outcome of leaving one drug in place long enough.

Switch before the tumor switches

A new set of mathematical models, published by researchers at City St George's, University of London, tests a different premise: change the therapy on a tight schedule, before the tumor has had time to build resistance to any single one. The models treat the tumor as an evolving population and simulate sequences of drugs rather than single, escalating doses.

The result is counterintuitive but consistent. Carefully timed switches between treatments, initiated while the tumor is still responding, can hold the cancer below the threshold where resistant clones become dominant. In several model scenarios this approach outperforms the standard maximum-dose strategy on the metric that actually matters: long-term cure rate, not short-term shrinkage.

From "maximum tolerated dose" to "maximum evolvable dose"

The concept echoes an older idea called adaptive therapy, in which treatment is modulated to keep a tumor stable rather than erased. The new modeling sharpens the logic further. The key parameter is not how much drug you give, but how much evolutionary room you leave. A regimen that always leaves a window open for a resistant clone to emerge, no matter how high the dose, is losing by design.

Implementing this in the clinic is harder than writing a model. It requires real-time monitoring of tumor populations, drug combinations with compatible side effects, and trial designs that reward survival over radiological response. Those are steep requirements, but the principle is clear.

Knowledge takeaway: high-dose cancer therapy selects for resistant cells, so resistance is an expected evolutionary outcome rather than a treatment failure. Mathematical models show that tightly timed switches between drugs can outperform maximum dose on long-term cure rates. The reframing shifts the goal from "maximum tolerated dose" toward managing how much evolutionary room a tumor is allowed.