Purpose <p>Poly(ADP-ribose) polymerase inhibitors (PARPi) can induce synthetic lethality in cancer cells with defective homologous recombination (HR) repair. Combination therapies with DNA-damaging chemotherapeutics can overcome PARPi resistance and improve response in HR-proficient tumors, but myelotoxicities have limited their clinical success. Therefore, we developed a pharmacokinetic (PK)-pharmacodynamic (PD)/toxicodynamic (TD) model to explore the impact of dosing regimen modifications on anti-tumor efficacy and myelotoxicity for PARPi combination therapy with DNA-damaging chemotherapeutics.</p> Methods <p>A mathematical framework was created linking drug PK exposure to downstream tumor PD and bone marrow TD effects driven by drug-induced DNA damage and cell death. Human PK parameters for a representative PARP inhibitor, olaparib, and DNA-damaging chemotherapeutic, temozolomide, were combined with <i>in vitro</i>-derived PD parameters and literature-derived TD parameters to enable model-based prediction of anti-tumor efficacy and myelotoxicity. To illustrate how the model framework can be parameterized to support extrapolation to the clinical setting, virtual patients (VPs) were created using distributions across a subset of physiological parameters to approximate clinical datasets describing response of breast cancer patients to olaparib and temozolomide therapy. Representation of tumor drug resistance was captured through incorporation of mixtures of HR-deficient and HR-proficient cells.</p> Results <p>Simulations with the human parameterization of the integrated PK-PD/TD model framework were used for model-based exploration of olaparib monotherapy and combination therapy regimens. While olaparib monotherapy can provide a strong initial size reduction for HR-deficient tumors, model simulations illustrated how the outgrowth of HR-proficient cells can drive rapid disease progression. In contrast, combination regimens with DNA-damaging chemotherapeutics such as temozolomide can prolong this time to progression by enhancing killing of HR-deficient cells while also restricting the outgrowth of HR-proficient cells. To facilitate identification of efficacious combination dosing regimens with acceptable toxicity profiles, we developed a combined optimization metric integrating simulated tumor size changes, time to progression, and toxicity for conducting model-based dose optimization.</p> Conclusion <p>An integrated PK-PD/TD model framework was developed to support model-based exploration of potential factors contributing to the limited clinical success that has been observed to date with combinations of PARP inhibitors and DNA-damaging chemotherapeutics. Model simulations indicated that while many of the combinations could lead to rapid tumor size reductions, the corresponding rapid myelosuppression would become dose-limiting. However, simulated dosing regimens involving longer dose intervals or lower dose levels of the DNA-damaging chemotherapeutic might provide acceptable myelosuppression while still driving clinical benefit through an extension in the time to progression.</p>