Model-informed drug discovery and development (MID3) uses mathematical modeling to inform decision-making in drug development programs. Applying MID3 early in development can reduce late-stage risk by determining feasibility of drugging a given target, prioritizing between targets, or defining optimal drug properties for a target product profile. However, the lack of pharmacokinetic (PK) and pharmacodynamic (PD) data available at early stages can make modeling a challenge. For this reason, we developed Early Feasibility Assessment (EFA): a method for making effective dose predictions using mechanistic PK/PD models built from general biophysical principles and parameterized by data that is readily available early in drug discovery. EFA centers on defining a notion of “dose feasibility,” that a drug may be administered with a reasonable dosing regimen and conceivably achieve a therapeutic impact.
Here we demonstrate the ability of EFA to predict clinical effective doses for nine approved biotherapeutics, including monoclonal antibodies (mAbs) with both soluble and membrane-bound targets in mono- and bispecific formats. These model predictions are accurate to within three-fold of the clinically approved dose, which is sufficient for informing preclinical study design and early decision making. Such predictions can be used to assess whether a novel therapeutic candidate can feasibly attain efficacious levels of target engagement or inhibition, prior to performing any animal or human studies.