OnCo
ideasIdea

Forecast the next resistance mutation like the weather

Flu vaccines are chosen by predicting which virus strains will dominate next season. The same forecasting maths could predict which resistance mutation a patient's tumour will develop next.

Evolutionary forecasting methods from influenza and bacterial resistance estimate the fitness of circulating variants from their frequency trajectories. Applied to longitudinal ctDNA and to population-level databases of resistance under each drug, they could give per-patient probabilities of specific next-step mechanisms. Forecasts would be scored prospectively, as in weather forecasting, to build calibrated models.

Hypothesis
A forecasting model trained on longitudinal ctDNA from patients on a given targeted drug predicts the dominant resistance mechanism at progression with calibrated accuracy well above the population base rate.
Rationale
Resistance mechanisms under a given drug are strongly constrained (a handful of routes dominate for osimertinib, alectinib, or sotorasib), and their early emergence is visible in serial plasma. Predictability is what makes pre-emptive combination possible.
What would test it
Train on serial plasma from completed trials, publish locked forecasts for a prospective cohort, and score them against observed progression biopsies; a Brier score better than base rate confirms.
Maturity
speculative
Who has to act
data
Cost to try
Small (under $1M)
Years to first evidence
3
Bottlenecks it attacks

Connected

8top

Pages like this

not linked directly; found by shared links