Intended for healthcare professionals


Magnetic resonance imaging is preferred in diagnosing suspected cerebral infections

BMJ 2000; 320 doi: (Published 15 January 2000) Cite this as: BMJ 2000;320:187
  1. N D Marchbank, consultant radiologist,
  2. D C Howlett, consultant radiologist,
  3. D F Sallomi, consultant radiologist,
  4. D V Hughes, consultant radiologist
  1. Eastbourne District General Hospital, Eastbourne, East Sussex BN21 2UD

    EDITOR—In their lesson of the week Fitzpatrick and Gan state that contrast enhanced computed tomography is currently the preferred investigation in diagnosing cerebral abscess.1 They describe the role of magnetic resonance imaging in such diagnosis as unclear, citing a 1988 comparative study of two cases as proof.2 We use magnetic resonance imaging as the preferred imaging technique in patients with suspected cerebral infections.3 It has improved resolution with multiplanar capability, and it is far more sensitive at detecting subtle white matter abnormalities as may be found in patients with herpes simplex encephalitis. We accept that computed tomography may be more widely available in the United Kingdom, though we do not necessarily concur that it is the best investigation.

    Fitzpatrick and Gan point out that cerebral abscess is a rare condition with appreciable morbidity and mortality, and early diagnosis is mandatory.1 Unfortunately, many patients with cerebral abscess present with clinical and radiological features identical with stroke. Stroke is common, and most centres do not use intravenous contrast medium in patients who have had a suspected stroke. If the history is of a sudden onset of neurological loss without clinical features suggestive of a cerebral abscess (such as fever and leucocytosis) we use unenhanced computed tomography. Acute cerebral infarction can cause a mass effect, and unless the possibility of alternative disease is raised by the referring clinician then these patients may well be misdiagnosed as having had cerebral infarction. If the pattern of oedema is atypical for cerebral infarction, or if the patient's condition deteriorates and has features suggestive of an abscess, then further imaging with either gadolinium enhanced magnetic resonance imaging or contrast enhanced computed tomography is required.

    The authors state that without expert neuroradiological advice “it is too easy to interpret findings as representing cerebral infarction.” We agree that expert advice is often helpful. We emphasise, however, that cerebral abscess and infarction can appear identical on unenhanced computed tomography. Therefore clinical suspicion of infection or an atypical history of stroke must be communicated to the radiology department so that appropriate imaging can be performed. We believe that this should ideally be gadolinium enhanced magnetic resonance imaging, or contrast enhanced computed tomography if magnetic resonance imaging is unavailable.


    1. 1.
    2. 2.
    3. 3.