Magnetic resonance imaging is not a sensitive test for Creutzfeldt-Jakob disease

EDITOR,—Terence Featherstone asks whether the prevalence of signal abnormalities in the basal ganglia seen on magnetic resonance imaging in Creutzfeldt-Jakob disease is being studied.1 We have been conducting systematic surveillance of Creutzfeldt-Jakob disease in Britain since 1990 and have collected data, including reports of magnetic resonance imaging, on 256 definite and probable cases (as defined by Masters et al2). We are also currently involved in a study to ascertain radiological and pathological correlations, using postmortem magnetic resonance imaging of the brain before histological processing.

Ninety six people are known to have undergone magnetic resonance imaging of the brain during their illness. Of these, eight had iatrogenic and four familial disease. Only four patients (all with histologically confirmed disease; three sporadic cases and one related to treatment with growth hormone) were reported to have high signal abnormalities of the basal ganglia on imaging. The remainder of the scans were either normal (41) or showed atrophy (29), scattered areas of high signal (6), or non-specific changes. Of the 61 patients referred to our unit with suspected Creutzfeldt-Jakob disease but with another histological diagnosis, six are known to have had magnetic resonance imaging. None of the scans were reported as showing the high signal abnormalities of the basal ganglia.

Only one of the four patients with high signal abnormalities of the basal ganglia presented with an extrapyramidal syndrome. This patient and two of the others had extensive and characteristic histological features of Creutzfeldt-Jakob disease in the basal ganglia. The final (sporadic) case was diagnosed on frontal biopsy, and results of postmortem histological examination are awaited. Our three sporadic cases were unusual in two other respects: the characteristic electroencephalographic pattern seen in Creutzfeldt-Jakob disease was absent, and the duration of their illness was relatively prolonged (10, 14, and 17 months compared with the median duration of sporadic Creutzfeldt-Jakob disease of 5 months). Cases of long duration without the typical electroencephalographic pattern can present diagnostic difficulty, and we agree that for these patients the appearance on magnetic resonance imaging might be useful diagnostically.

In the correct clinical context high signal abnormalities in the basal ganglia on magnetic resonance imaging may be a relatively specific (although not sensitive) sign of Creutzfeldt-Jakob disease; in younger people rare conditions, such as Wilson's disease3 and mitochondrial cytopathies,4 may cause similar appearances, but the clinical presentation of these conditions distinguishes them from Creutzfeldt-Jakob disease. In older patients high signal in the periventricular white matter is common and senile changes may affect the basal ganglia, but the changes are usually asymmetric and patchy. The characteristic electroencephalographic pattern of generalised repetitive triphasic periodic complexes remains the most sensitive non-invasive diagnostic test of Creutzfeldt-Jakob disease.