Purpose: The starting dose of sunitinib in children with gastrointestinal stromal tumors (GIST) was extrapolated based on data in adults with GIST or solid tumors and children with solid tumors. Methods: Integrated population pharmacokinetics (PK), PK/pharmacodynamics (PD), and exposure–response analyses using nonlinear mixed-effects modeling approaches were performed to extrapolate PK and PD of sunitinib in children with GIST at projected dose(s) with plasma drug exposures comparable to 50-mg/day in adults with GIST. The analysis datasets included PK/PD data in adults with GIST and adults and children with solid tumors. The effect of covariates on PK and safety/efficacy endpoints were explored. Results: Two-compartment models with lag time were successfully used to describe the PK of sunitinib and its active metabolite SU012662. PK/PD models were successfully built to describe key continuous safety and efficacy endpoints. The effect of age on sunitinib apparent clearance (CL/F) and body surface area on SU012662 CL/F was statistically significant (P ≤ 0.001): children who were younger or of smaller body size had lower CL/F; however, age and body size did not appear to negatively affect safety or efficacy response to plasma drug exposure. Conclusion: Based on PK, safety, and efficacy trial simulations, a sunitinib starting dose of ~ 25 mg/m2/day was predicted to provide comparable plasma drug exposures in children with GIST as in adults with GIST treated with 50 mg/day. However, in the absence of a tumor type effect of sunitinib on CL/F in children, the projected equivalent dose for this population would be ~ 20 mg/m2/day. [ABSTRACT FROM AUTHOR]