Surgery of intrinsic cerebral neoplasms in eloquent areas under local anesthesia.

28 patients with a mean age of 43.6 years were operated on for a cerebral neoplasm situated in close proximity to an eloquent area (24 speech area, 4 motor cortex) from 1996 to 1999. Preoperatively, all patients had undergone a detailed neuropsychological examination. In 10 patients aphasic disturbances could be detected. All patients underwent preoperative PET studies (methionine and (15)O-labeled water with activation during speech or finger tapping). These were performed and co-registered with MRI data to demonstrate the topographical relationship between motor or language function and the tumor borders. Anesthesia was induced with i.v. administration of propofol (150-250 mg/h). Craniotomy was performed under local infiltration anesthesia. After opening of the dura, sedation was stopped and operation was continued with the patient being alert and co-operative. With close clinical observation during electrical cortex stimulation, a speech arrest could be triggered or avoided. The motor cortex was identified by recording the phase reversal of the contralateral SEP of the median nerve and by direct cortical stimulation. As soon as aphasic or motor disturbances appeared, the tumor removal was continued with the goal of avoiding these specific regions. In 27 patients, preexisting neuropsychological and neurological deficits did not worsen. Only one patient was left postoperatively with a major permanent aphasic deficit that was present preoperatively to a minor degree. The use of local anesthesia in craniotomy for surgery of intrinsic cerebral neoplasms in eloquent areas allows for a continuous and repetitive monitoring of speech and motor function during the removal of even those tumors that were previously considered inoperable.