What does this teach about the drug?
38% of Poor metabolizer patients (n 99) are responding and under the churn line, against 46% overall. Enrich the next cohort and the label language for this segment; the biology is dna · drug exposure.
19% carry toxicity ≥ 50 today (n 27). Dose-timing and meal-timing interventions are the cheapest test: in the data, taking the dose after a meal and low-residue meals on dosing days are the primitives with the largest GI effect. A dose-timing sub-study in this segment is a small, fast trial.
Dosing / Timing Recommendation is the protocol that turns an accepted recommendation into a retained patient most often (n 15). Bundling it into the companion from the first dosing day for every patient in the Poor metabolizer segment is the strongest strategy candidate in the data.
ONX-7 keeps 46% of patients responding and under the churn line. The ones it loses leave through side effects that are predictable from biology and daily behaviour, not from the molecule alone, and the same data shows which protocol brings each biology back. The strategic asset is therefore not the drug by itself but the drug plus its steering layer: diagnostics at baseline, a companion that carries the right content, and a model that keeps learning from every patient.
Rules behind these memos: Learning. Emerging patterns: Watch.