Precision oncology matches a drug's molecular target to the mutation actually driving a patient's tumor. A targeted inhibitor only binds tightly to the mutant protein it was designed against; on a tumor lacking that mutation it has little to grab onto, so response falls back to a low baseline (similar to placebo/non-response rates seen in unselected patients).
P(response) = base(drug) if biomarker mismatch
P(response) = base(drug) + [match(drug) − base(drug)] × f_mutant if biomarker match
f_mutant = fraction of tumor cells expressing the driver mutation
base(drug), match(drug) = drug-specific response rates (mismatched vs matched)
- Mutation toggles — set which driver mutation(s) this tumor's genomic profile carries.
- Mutant cell fraction — tumors are heterogeneous; only some cells express the mutation, so only those cells present a docking site for a matched drug.
- Drug buttons — each targeted agent has one specific protein target; concordance between the drug's target and the tumor's mutation is what predicts benefit. Chemotherapy has no genomic target, so its response rate is flat regardless of biomarkers.
- 3D field — each sphere is a tumor cell; red spheres express the driver mutation, grey spheres are wild-type. Particles are drug molecules — they dock (turn green, cell lights up) only on cells presenting a matched target, and drift away (fade) on cells they cannot bind.
Real-world relevance: this concordance logic is exactly why oncology guidelines require genomic testing (NGS panels) before prescribing an EGFR/HER2/BRAF/KRAS inhibitor — giving a targeted drug to a biomarker-negative tumor wastes the window for effective treatment.