Toward greater insights on pharmacokinetics and exposure–response relationships for therapeutic biologics in oncology drug development
Methodologically rigorous approaches are needed to robustly leverage understanding of ER for regulatory purposes. The randomized, concentration‐controlled clinical trial as a method to minimize the impact of pharmacokinetic (PK) variability on response assessment has been logistically challenging. Meanwhile, significant effort has been invested in methods development to address potential interaction between drug exposure and prognostic factors that may complicate assessment of a drug's ER relationship. For example, ER and “case‐control” approaches have been combined and employed in oncology to reduce the likelihood that imbalanced baseline prognostic factors would spuriously impact interpretation of ER analyses.4 The potential for prognostic imbalances to impact the exposure–survival relationship must be routinely considered. Without accounting for the effects of baseline differences in patient risk factors for poorer survival, a falsely steep ER relationship is generally expected if data from the investigational arm alone (under a single regimen) were used to quantify the ER relationship (details discussed in the Drug Exposure – Disease Interactions section, below). One may be tempted to use the apparent positive relationship between the control‐corrected effect size and the exposure level across different exposure levels as justification to explore higher exposure for better efficacy in subgroups with lower exposures. Such a decision, however, is only valid when the ER relationship is similar for all subgroups with different distributions of risk factors.5
When the ER relationship is correlated with baseline risk factors (e.g., sicker patients have a flatter or no ER relationship), the apparent positive ER relationship across different matched subgroups is the outcome of both exposure and risk factor distributions. Under an extreme case where no ER relationship is assumed for sicker patients and exposure is high enough to achieve the maximum benefit for the less sick patients within each subgroup, an apparent ER relationship can be observed across different subgroups with different exposure levels. Such an apparent relationship, however, is entirely due to the different proportions of sicker patients in different subgroups, and increasing drug exposure in any subgroup will not further improve the efficacy (Figure1). In the oncology context, it is not uncommon to observe different ER relationships in subgroups with different degrees of disease severity (e.g., based on Eastern Cooperative Oncology Group (ECOG) status).7 The common explanation for this observation is that sicker patients tend to have greater tumor burden and increased clearance of the biologic product, leading to a lower drug exposure level. This explanation, however, does not take into consideration that cancer is a systemic disease. In many forms of cancer, tumors not only disrupt normal function at the site of anatomical location, but actively produce inflammatory cytokines and growth factors that contribute to the “sickness” of the patient.