Results of initial clinical trials of brain tumor chemotherapy after osmotic blood-brain barrier disruption are promising. In general, the procedure is well tolerated. The major complication has been seizures. In this report, data are presented which indicate that the etiology of these seizures is related to the use of contrast agent (meglumine iothalamate) to monitor barrier modification. A series of 19 patients underwent a total of 85 barrier modification procedures. Documentation of barrier disruption was monitored by contrast-enhanced computed tomographic (CT) scanning, radionuclide brain scanning, or a combination of both techniques. In 56 procedures (19 patients) monitored by enhanced CT, seizures occurred a total of 10 times in eight patients. Twenty-three barrier modification procedures (in nine of these 19 patients) documented by nuclear brain scans alone, however, resulted in only one focal motor seizure in each of two patients. In eight of the 19 patients who had seizures after barrier disruption and enhanced CT scan, four subsequently had repeat procedures monitored by radionuclide scan alone. In only one of these patients was further seizure activity noted; a single focal motor seizure was observed. Clearly, the radionuclide brain scan does not have the sensitivity and spatial resolution of enhanced CT, but at present it appears safer to monitor barrier modification by this method and to follow tumor growth between barrier modifications by enhanced CT. Four illustrative cases showing methods, problems, and promising results are presented.