Project Optimus

Launched by the FDA Oncology Center of Excellence in 2021, Project Optimus requires randomised dose comparison and exposure-response work before pivotal trials, reversing the maximum-tolerated-dose paradigm inherited from chemotherapy. It has changed early-phase design for targeted agents and ADCs (dose-optimisation cohorts, lower approved doses such as sotorasib 240 mg).

The reasoning is that the highest dose patients can survive for a few weeks in a phase 1 trial is rarely the dose that gives the best balance of benefit and tolerability over months of treatment. Chemotherapy has a steep dose-response curve, so more was usually better; targeted drugs and antibody-drug conjugates saturate their target well below the maximum tolerated dose, and the extra dose buys only toxicity, dose reductions and discontinuations. The FDA's 2024 guidance on dose optimisation asks sponsors to compare at least two doses in a randomised cohort, to characterise exposure-response for both efficacy and safety, to look at tolerability over the long term with patient-reported outcomes, and to do this before rather than after the pivotal trial.

In practice the programme has reshaped phase 1/2 protocols: a dose-escalation part using model-based designs such as BOIN, then a randomised dose-optimisation part of two or more doses that picks the dose taken into phase 3. REJOICE-Ovarian01 ran a phase 2 dose-optimisation part before randomising its phase 3 at the chosen dose. Outside the FDA's remit, the Tata Memorial trial of nivolumab at 20 mg, a small fraction of the standard dose, alongside metronomic chemotherapy in head and neck cancer is the leading example of a dose question answered by a randomised trial in the interest of affordability rather than approval.