VP-16 100 mg/m2 was given intravenously to 10 patients undergoing surgical resection of intracerebral tumors, and the drug was assayed in resected tumor using high pressure liquid chromatography. VP-16 concentrations varied from undetectable (less than .1 microgram/g) to 5.9 micrograms/g (mean, 1.4 microgram/g). VP-16 concentrations in tumors were lower than concurrent plasma concentrations. In addition, intracerebral tumors had a lower concentration of VP-16 than did extracerebral tumors (mean VP-16 concentration, 3.9 micrograms/g) from 7 patients receiving VP-16 50-100 mg/m2 intravenously. Plasma pharmacokinetics of VP-16 were different in our patients with intracerebral tumors than in previously studied patients with extracerebral tumors and it is unclear what role this may played in variability of tumor VP-16 concentrations. VP-16 concentrations were similar in glioblastomas and brain metastases. Specimens from patients with small cell undifferentiated carcinoma of the lung had the highest VP-16 concentrations. A patient who had both viable and necrotic tumor resected during an occipital lobectomy had a higher drug concentration in the necrotic than in the viable area of tumor. In addition, VP-16 concentration decreased as a function of distance into brain from the tumor. Based on our data, VP-16 might be expected to have less activity against intracerebral than against extracerebral human tumors.