You are in: eMedicine Specialties > Neurology > Electroencephalography and Evoked Potentials EEG in Brain TumorsArticle Last Updated: Sep 27, 2006AUTHOR AND EDITOR INFORMATIONAuthor: Edward B Bromfield, MD, Associate Professor of Neurology, Faculty Member, Division of Sleep Medicine, Harvard Medical School; Chief, Division of EEG, Epilepsy and Sleep Neurology, Consulting Neurologist, Brigham and Women's Hospital Edward B Bromfield is a member of the following medical societies: American Academy of Neurology, American Clinical Neurophysiology Society, American Epilepsy Society, American Neurological Association, and Massachusetts Medical Society Editors: Andrew S Blum, MD, PhD, Director, Comprehensive Epilepsy Program, Assistant Professor, Department of Clinical Neurosciences, Rhode Island Hospital; Francisco Talavera, PharmD, PhD, Senior Pharmacy Editor, eMedicine; Norberto Alvarez, MD, Assistant Professor, Department of Neurology, Harvard Medical School; Consulting Staff, Department of Neurology, Boston Children's Hospital; Matthew J Baker, MD, Consulting Staff, Collier Neurologic Specialists, Naples Community Hospital; Nicholas Y Lorenzo, MD, Chief Editor, eMedicine Neurology; Consulting Staff, Neurology Specialists and Consultants Author and Editor Disclosure Synonyms and related keywords: breach rhythm, brain neoplasms, electroencephalography, intermittent rhythmic delta activity, polymorphic delta activity INTRODUCTIONUntil the past 3 decades, localization of brain tumors and other focal lesions was difficult. Neuroimaging techniques consisted of skull radiography, which was usually negative, and pneumoencephalograms, which were invasive, painful, and often uninformative. In 1936, Walter, who introduced the term "delta waves," first identified the association between localized slow waves on EEG and tumors of the cerebral hemispheres. This established EEG as an important tool for localizing brain tumors. For the next 4 decades, electroencephalographers mounted an enormous effort to improve accuracy of localization and to seek clues to underlying pathological processes. Experience has shown EEG to be somewhat reliable in localizing lesions involving superficial accessible portions of the cerebral hemispheres, though it is of limited value in deep-seated lesions, especially posterior fossa tumors. The role of EEG in detecting focal cerebral disturbances has undergone a significant change since the development of CT scan and MRI. Today EEG is used primarily to complement these studies by evaluating functional changes in the patient's condition; it demonstrates aspects of brain physiology that are not reflected in structural neuroimaging. Functional neuroimaging techniques, such as positron emission tomography (PET), single-photon emission computed tomography (SPECT), and functional MRI (fMRI), can exhibit physiologic changes but not with the temporal resolution of EEG. Furthermore, EEG provides the only continuous measure of cerebral function over time. This article reviews the major EEG changes that occur with different brain tumors, as determined by location, histologic type, associated complications, and surgical and nonsurgical treatments. For excellent patient education resources, visit eMedicine's Cancer and Tumors Center and Procedures Center. Also, see eMedicine's patient education articles Brain Cancer and Electroencephalography (EEG). BACKGROUNDClassification Intracranial tumors have been classified by location and histology. Classification by location, modified from Merritt (1991), is as follows: I. Tumors of the skull
II. Tumors of the meninge
III. Tumors of the cranial nerve
IV. Neuroectodermal tumor
V. Other primary tumor
VI. Metastatic
The 1993 World Health Organization classification is based on histology. A simplified version is as follows: I. Tumors of neuroepithelial origin
H. Pineal parenchymal tumors II. Embryonal tumors III. Tumors of cranial and spinal nerves IV. Tumors of the meninge
V. Lymphomas and hematopoietic neoplasms VI. Germ cell tumors VII. Cysts and tumorlike lesion
VIII. Tumors of the sellar region
IX. Local extension from regional tumors X. Metastatic tumors XI. Unclassified tumors Epidemiology One estimate states that in the United States, approximately 35,000 new cases of primary brain tumor are diagnosed each year, corresponding to an incidence of 8-19 cases per 100,000 population. Metastatic brain tumors are 3-4-fold more common, estimated at 110,000-140,000 new cases per year in the United States. The following characteristics apply to primary brain tumors:
The various types of brain tumors occur with different frequencies in children and in adults.
Brain tumors presenting with seizures Seizures occur in 40-60% of patients with supratentorial tumors and are the presenting symptom in 10-40%. When seizures start in adulthood, neoplasms are the cause in 2-20% of patients. In general, the more slowly a tumor grows, the more likely it is to present with a seizure. For example, as many as 92% of oligodendrogliomas are associated with seizures, and in most patients they are the presenting symptom. Similar or higher percentages apply to gangliogliomas and dysembryoplastic neuroepithelial tumors. Location is also important in determining the likelihood of seizures. Perirolandic cortex is believed to be most epileptogenic, although this impression may reflect in part the ease of diagnosing seizures in the presence of prominent motor activity. Medial temporal structures are probably the next most epileptogenic, followed by other areas of frontal, temporal, and parietal lobes, with occipital lobe the least likely to produce seizures. Subcortical tumors, such as those in the thalamus and particularly the posterior fossa, rarely cause seizures. TYPES OF EEG ABNORMALITIES ASSOCIATED WITH BRAIN TUMORSEEG abnormalities in brain tumors depend on the stage at which the patient presents for evaluation. EEG changes observed with tumors result mainly from disturbances in bordering brain parenchyma, since tumor tissue is electrically silent (with the possible exception of tumors containing neuronal elements, such as gangliogliomas). For this reason, EEG localization often is misleading, although lateralization is generally reliable. The following are common findings:
Activation procedures are usually of limited value in patients with tumors, although hyperventilation occasionally can accentuate focal slowing. Asymmetries of photic driving can be useful at times, although they also can be misleading. Slow wave activity Focal delta activity is the classic electrographic sign of a local disturbance in cerebral function. A structural lesion is strongly suggested if the delta activity is continuously present; shows variability in waveform amplitude, duration, and morphology (polymorphic); and persists during changes in physiologic states, such as sleep or alerting procedures. When focal delta is found without a corresponding imaging abnormality, it is usually in the setting of acute seizures (especially postictally), nonhemorrhagic infarction, or trauma. Clinical and experimental observations indicate that PDA results primarily from lesions affecting cerebral white matter; involvement of superficial cortex is not essential, and lesions restricted to the cortical mantle generally do not produce significant delta activity. Functional deafferentation of cortex likely is critical. Delta activity that fails to persist into sleep or attenuates significantly with arousal or eye opening is less indicative of structural pathology, as is rhythmic or sinusoidal delta. The latter usually occurs intermittently and is termed IRDA. It is usually bilateral and of high amplitude and is typically maximal occipitally (OIRDA) in children and frontally (FIRDA) in adults. Unlike PDA, IRDA increases in drowsiness and attenuates with arousal. IRDA often is observed without structural pathology, as in metabolic encephalopathies, but it also can occur with diencephalic or other deep lesions; in this situation, an amplitude asymmetry can be present, with higher amplitude ipsilateral to the lesion. As in other clinical settings, theta activity is indicative of less severe localized or diffuse dysfunction than delta activity and is observed more commonly with functional disturbances than with structural disturbances. When unaccompanied by delta activity, theta is less likely to indicate a lesion that produces a focal neurological deficit or seizures. Localized loss and asymmetries of background activity Since tumor tissue probably does not generate electrical activity detectable with conventional recording techniques, electrical silence is the best localizing sign of a cerebral tumor. However, it is a rare finding, occurring only when the tumor involves significant cortical areas with minimal subcortical disruption. Incomplete loss of activity, especially faster normal rhythms, is observed more commonly and is diagnostically helpful. Alpha rhythm By the time the patient presents with focal or diffuse neurological symptoms and signs, disturbance of the alpha rhythm may be observed. Slowing of the alpha rhythm ipsilateral to a tumor is more common and significant than asymmetry of amplitude. However, disturbance of alpha rhythm depends on the site of the tumor. The more posterior the location, the more the alpha tends to be slowed, nonpersistent, or disturbed by admixed theta waves. The alpha rhythm also may fail to block to eye opening on the side of the neoplasm (ie, Bancaud phenomenon). Beta activity Abnormalities of beta activity usually are limited to voltage asymmetries. To be considered unequivocally abnormal, a persistent amplitude difference of one third or greater (expressed as a fraction of the higher voltage) should be present. Diminished beta activity results either from cortical dysfunction, as in parenchymal tumors, or from an increase in resistance of the medium-separating cortex from scalp-recording electrodes, as in meningiomas or subdural collections. Focally increased beta activity usually is associated with a skull defect. Interictal epileptiform dischargesIsolated discharges Spikes, sharp waves, or spike-wave complexes occurring with consistent localization are observed uncommonly early in the course of brain tumors. In one study predating the CT scan era, such discharges appeared with only 3% of gliomas and metastatic tumors at the time of craniotomy. However, they were more common either as early findings of slowly growing neoplasms associated with seizures or later after focal slowing had developed. Periodic lateralized epileptiform discharges Patients with tumors may exhibit periodic lateralized epileptiform discharges (PLEDs) at times, particularly after a series of seizures. In a study of 282 patients with typical PLEDs, tumor was present in 18%. Most patients with this EEG finding have had or will have seizures, if they are observed sufficiently closely and persistently; the pattern likely represents a transitional state between ictal and interictal epileptiform discharges. Aggressiveness of treatment depends in part on whether the discharges are resolving (ie, becoming less sharp, more localized, and further apart) or the opposite. Seizure patterns When electrographic seizures are recorded during a routine EEG, status epilepticus should be suspected. Clinical accompaniments may be subtle, as in aphasic or other forms of nonconvulsive status, particularly when the patient's baseline condition has been compromised by the tumor, its treatment, or complications. EEG CHANGES IN TUMORS BY LOCATIONSince EEG reflects activity of cortical neurons, hemispheric tumors affect EEG most consistently and prominently. In older studies, a normal EEG occurred in approximately 5% of hemispheric, 25% of deep or basal, and 25% of infratentorial tumors. The overall figure now may be 50% or higher, given the earlier diagnosis allowed by modern neuroimaging. Location is an important determinant of the likelihood and nature of EEG abnormalities. Frontal lobe tumors
Temporal lobe tumorsTemporal gliomas are generally the easiest to localize on EEG, since PDA occurs over the tumor site in more than 80% of patients. Tumors in other locations, such as the thalamus, also may produce temporal delta; however, focal delta is more reliably localizing when background rhythms also are attenuated. Anterior temporal
Posterior temporal
Parietal lobe tumors
Occipital lobe tumors
Deep hemispheric tumorsDeep hemispheric tumors include those than impinge on the lateral and third ventricle and surrounding structures, including the diencephalon, basal ganglia, and corpus callosum. Neuroimaging has led to earlier diagnosis of smaller tumors that may be associated with normal EEGs. When abnormalities are observed, the following apply:
Intraventricular and sellar tumorsLateral ventricle (ependymoma, meningioma)
Third ventricle (colloid cyst, epidermoid, craniopharyngioma, hypothalamic hamartoma)
Sellar region
Infratentorial tumorsBrain stem and cerebellum
Cerebellopontine angle (acoustic neuroma)
TUMOR TYPE AND EEGEEG patterns are not specific for tumor pathology, but some general correlations exist.
Meningiomas Being extraaxial, meningiomas compress the brain but cause little destruction of brain tissue. Therefore, meningiomas of the anterior or middle cranial fossa, unless large, infrequently alter EEGs. Convexity meningiomas are more likely to cause EEG changes. With rolandic or parasagittal meningiomas, common EEG changes include the following:
Gliomas Slowly growing gliomas such as oligodendrogliomas and fibrillary astrocytomas (excluding tumors of deep structures) often can be distinguished from the more rapidly growing anaplastic astrocytoma and glioblastoma multiforme.
Metastases Metastatic tumors to the brain occur commonly with carcinomas of lung, kidney, and breast and with melanomas and chorionic carcinomas. When metastases are present bilaterally, slowing often appears diffuse, although it is often asymmetric; slowing from multiple bilateral lesions is often difficult to distinguish from a toxic-metabolic disturbance. Meningeal carcinomatosis usually causes changes that correlate with the clinical situation; when deposits are widespread and cause an encephalopathy, slowing is usually diffuse. Isolated metastases usually cause less prominent abnormalities than gliomas of similar size and location. Slow waves show higher frequency, lower amplitude, and less persistence than with high-grade gliomas, and normal background rhythms are more likely to be preserved. EEG CHANGES OVER TIMEBecause of improvements in neuroimaging and neurosurgery and recognition of the benefits of early resection, serial EEG studies now are rarely performed prior to surgery. Older studies suggest that EEG evolution during tumor growth is characterized mainly by increased slowing—lower frequency, higher amplitude, more persistence, and wider distribution—with rate of change depending mainly on rate of tumor growth. In addition, epileptiform discharges are more likely to occur as the tumor grows. Occasionally, successful treatment with steroids or chemotherapy can cause reduction in slowing and epileptiform activity. Following resection, dramatic changes may occur in the EEG; these usually stabilize over periods of weeks to months. Since screening for tumor recurrence now depends on neuroimaging, serial EEGs usually are reserved for patients with clinical changes that are not explained by imaging, particularly when seizures are suspected. EEG changes after neurosurgery EEG changes after neurosurgery usually exhibit the following features:
Use of EEG to predict postoperative seizures Because of the tendency of focally increased high-frequency activity after craniotomy to sharpen the contour of background waves, identification of interictal epileptiform discharges is difficult. A distribution other than the breach rhythm, asymmetric up-slope and down-slope, extremely sharp peak, and prominent after-coming slow wave suggest epileptogenicity. However, in a preliminary study, degree of slowing (predominantly delta rather than theta) was associated more closely with seizures than with amount of sharp activity. Similar lack of predictive value of interictal discharges has been noted in a recent series of meningiomas and in a mixed group of mainly low-grade temporal lobe tumors. Another recent study failed to show a correlation between spikes or slow waves and seizure recurrence after withdrawal of antiepileptic drugs in children with brain tumors several years after presentation. Complications of brain tumors and their treatment The possibility of perioperative stroke has been mentioned. Sizable ischemic infarcts typically exhibit increased slowing in the region of the stroke, with loss of fast activity if cortex is involved. Hemorrhagic stroke or hemorrhage into the tumor bed also is accompanied by increased slowing, which may be bilateral if deep structures are affected. If chemotherapy or radiation is effective, slowing can diminish even without surgery. Conversely, late effects of radiation can result in increased slowing, as well as new epileptiform activity and clinical seizures in the case of radiation necrosis. EEG probably does not help in distinguishing recurrent tumor from radiation necrosis. In radiation-induced or chemotherapy-induced encephalopathies, including methotrexate leukoencephalopathy, slowing of the EEG usually parallels the clinical situation. The clinician and electroencephalographer also must remember that patients with brain tumors can develop additional diseases, particularly when immunosuppressed, such as progressive multifocal leukoencephalopathy or herpes simplex encephalitis. In these patients, EEG changes such as focal slowing or periodic discharges reflect the new condition. CONCLUSIONSThe character and distribution of EEG changes produced by tumors depend primarily on lesion size, rate of growth, distance from the cortical surface, and specific structures involved. In general, the following are true:
Despite advances in neuroimaging, EEG still offers a unique view of physiologic changes over time in patients with brain tumors. MULTIMEDIA
REFERENCES
Article Last Updated: Sep 27, 2006 | ||||||||||||||