A sensitive high-performance liquid chromatography method was used to measure mitoxantrone in autopsy tissue samples of 11 patients who had received the drug iv 10-272 days antemortem. Mitoxantrone was readily detectable in tissues from all patients. Tissue concentrations were proportional to lifetime cumulative dose of mitoxantrone, and decreased very slowly with time. The thyroid and the liver had the highest mitoxantrone concentrations, followed by the heart. These high cardiac concentrations of mitoxantrone could be partially responsible for the occasional case of cardiotoxicity seen with mitoxantrone. The brain had the lowest mitoxantrone concentrations. Organ mitoxantrone concentrations did not conform to a flow-limited model. Tumor mitoxantrone concentrations varied quite markedly from one site to another in the same patient. Tumors generally had lower mitoxantrone concentrations than did surrounding normal tissues. Mitoxantrone concentrations were consistently highest in intrahepatic tumors and lowest in intracerebral tumors. It is unclear whether the low concentrations in brain tumors were due to a partially intact blood-brain barrier or to the fact that most brain tumors had been irradiated prior to mitoxantrone administration. Further studies are warranted to more fully explore the relationship between human tissue mitoxantrone concentrations and efficacy and toxicity.