The scientific and regulatory communities support the use of pharmaco-statistical models of efficacy and safety during drug development, termed "Model-Informed Drug Development (MIDD)." Under the US Food and Drug Administration (FDA) Reauthorization Act of 2017, 1 the FDA will convene several workshops to identify best practices for MIDD. This summary highlights viewpoints from the first workshop, co hosted by the FDA and International Society of Pharmacometrics (ISoP) on February 1, 2018, entitled "Model-Informed Drug Development in Oncology."
ADVANCING MIDD IN ONCOLOGY DRUG DEVELOPMENTThe traditional oncology drug-development paradi g m , r ooted in the identification of a maximum tolerated dose, has resulted in rapid de velopment of treatments but with significant toxicities and minimum optimization. Advancing MIDD approaches presents an opportunity to transform the oncology drug-de velopment paradigm and improve patient outcomes, noted Issam Zineh (FDA) ( Figure 1). 2 To fu l l y l everage MIDD in oncology drug development, Janet Woodcock (FDA) noted that integration of the basic sciences, pharmacology, and cancer biology is needed to resolve knowledge gaps and define best practices. 3 To address these issues, the FDA and ISoP convened a workshop to highlight MIDD a p p r o a ches from non clinical to clinical drug development as well as regulatory review and postapproval (all available worksh o p m a terials and related citations are found on the FDA website 4 ). Workshop object i v e s included: (i) discussing best practices for integrating pharmacokinetics (PK), p h a r m acodynamics (PD), efficacy, and safety data into models; (ii) evaluating diseas e s p ecific and mechanism-specific early e n d p oints to predict long-term efficacy ; and (iii) discussing the regulatory implications of model-informed decisions.
MIDD DURING PRE CLINICAL DEVELOPMENTMIDD a p p r o aches can provide insights into the likelihood of clinical trial success for drug products, doses, regimens, and combinations. Armin Sepp (GlaxoSmithKline) presen t e d in silico modeling that optimized the binding of a bispecific antibody (DuetMab) to cells expressing both CD4 and CD70. Given the structural and functional complexity of bispecific antibodies, this modeling approach can help define the design, target-antibody kinetics, stability, and PK of bispecific antibodies to improve the likelihood of clinical trial success.The myriad of possible doses, combinations, and schedules when using multiple cancer treatments precludes the use of empirical clinical trials. However, MIDD approaches, such as quantitative systems pharmacology and ex p o s ure-response (ER) analyses can streamline combination drug development by nom i n a t i ng dose combinations more likely to be safe and effective. Sergey Aksenov (Astra Zeneca) presented a quantitative systems p h a r m a cology model incorporating PK, biological drug targets, and physiological compon e n t s to estimate immune interactions, T-cell tumor infiltration, and tumor response. A proof-of-conc...
