38 patients with glial brain tumors received 135 mg/m2 cisplatin intravenously every month for 5 courses. Signs and symptoms of peripheral neuropathy were evaluated clinically and electrophysiologically. This approach differs from methods previously reported in that it offers two major advantages: primary brain tumors do not cause paraneoplastic neuropathy; no neurotoxic drugs other than cisplatin were employed. Our study confirmed in this highly specific clinical model the existence of cisplatin peripheral neuropathy: we observed definite clinical signs or symptoms of neuropathy in 80% of patients receiving a cumulative cisplatin dosage of 675 mg/m2; there was a progressive dose-related decrease in SAP amplitudes, expression of a sensory axonopathy or neuronopathy.