All- retinoic acid (RA) is a front-line treatment of acute promyelocytic leukemia (APL). Due to its activity in regulating the cell cycle, it has also been evaluated for the treatment of other cancers. However, the efficacy of RA has been limited byRA inducing its own metabolism during therapy, resulting in a decrease of RA exposure during continuous dosing. Frequent relapse occurs in patients receivingRA monotherapy. In an attempt to combat therapy resistance, inhibitors of RA metabolism have been developed. Of these, ketoconazole and liarozole have shown some benefits, but their usage is limited by side effects and low potency toward the cytochrome P450 26A1 isoform (CYP26A1), the mainRA hydroxylase. We determined the pharmacokinetic basis of therapy resistance to RA and tested whether the complex disposition kinetics ofRA could be predicted in healthy subjects and in cancer patients in the presence and absence of inhibitors of RA metabolism using physiologically based pharmacokinetic (PBPK) modeling. A PBPK model ofRA disposition was developed and verified in healthy individuals and in cancer patients. The population-based PBPK model of RA disposition incorporated saturable metabolic clearance ofRA, induction of CYP26A1 by RA, and the absorption and distribution kinetics ofRA. It accurately predicted the changes in RA exposure after continuous dosing and when coadministered with ketoconazole and liarozole. The developed model will be useful in interpretation ofRA disposition and efficacy, design of novel dosing strategies, and development of next-generation RA metabolism inhibitors.