Progress in Understanding Creutzfeldt-Jakob Disease

The neurodegenerative diseases are experiencing a new lease on life. New treatments, such as anticholinesterases for Alzheimer disease, have become available, with several other promising drugs under assessment. Considerable progress has been achieved in the understanding of the underlying molecular and genetic basis of dementing diseases, but perhaps the most unexpected development has been the increasing contribution of imaging to this understanding. Medical progress increasingly requires a multidisciplinary approach. This paradigm is well illustrated by Creutzfeldt-Jakob disease (CJD), with imaging playing an increasingly important role in case definition and elucidating the cause of this disease. In this issue of the AJNR, Murata et al (page 1164), using diffusion-weighted imaging, provide important insights into the early diagnosis of CJD and how the disease may spread in the brain.

CJD is one of the transmissible spongiform encephalopathies, a rare but important group of diseases affecting humans and other animals, characterized by fatal progressive neurologic illness, unusual neuropathologic changes, and an unconventional transmissible causal agent. A number of subtypes of CJD are recognized, including sporadic CJD (sCJD, the most common form worldwide), familial CJD (fCJD, rare), iatrogenic CJD (iCJD, increasingly rare; from cadaveric hormone-related transmission or neurosurgical procedures), and the recently described variant CJD (vCJD). In some countries, CJD is a notifiable disease monitored by dedicated national CJD surveillance units.

An unusual feature of the agent causing CJD is its relative resistance to many routine sterilization procedures, including standard autoclaving. For many years, this transmissible agent resisted characterization. However, in his groundbreaking work for which he won a Nobel Prize, Stanley Prusiner proposed that the transmissible agent in CJD and related diseases was a protein (PrP^SC). This protein or prion catalyses the conversion of a normal native protein (PrP^C) into the isomeric PrP^SC form (the prion hypothesis [1]). The neuropathologic changes of CJD are spongiform change, neuronal loss, and astrocytic proliferation, associated with deposition of the PrP^SC protein throughout the brain.

sCJD most commonly affects patients who are between 60 and 75 years of age. It presents clinically with a rapidly progressive dementing illness, culminating in akinetic mutism. Neurologic features observed during the illness reflect widespread neuronal damage and include myoclonus, cerebellar ataxia, pyramidal and extrapyramidal signs, and cortical blindness. The role of investigations in the diagnosis of CJD has been comprehensively reviewed (2). Until recently, the clinical diagnosis has relied on identification of appropriate clinical features supported by characteristic EEG changes (periodic triphasic sharp wave complexes or periodic synchronous discharge or both, seen in two thirds of patients) and CSF protein electrophoresis findings (raised 14-3-3 protein, sensitivity and specificity of 85–90%).

Characteristic, usually symmetrical, hyperintensity (relative to cortical gray matter signal intensity) of the caudate head and putamen has been described in 67–79% of sCJD cases on MR images, with varying specificity (3, 4). Some investigators have proposed that these findings should be incorporated into the diagnostic criteria for sCJD. However, false-positive imaging results can occur, partly because the normal putamen is sometimes hyperintense relative to the cortical gray matter with some pulse sequences (particularly proton density-weighted images) and also because overlap of appearances exists with other conditions (although these are usually clinically distinct from sCJD). Other gray matter structures may be affected in sCJD, including the hippocampus, peri-aqueductal gray matter, and thalamus (although in sCJD, the signal intensity in the thalamus remains lower than that in the putamen).

These basal ganglionic changes have been shown on T2-weighted, proton density-weighted, fluid-attenuated inversion recovery, and diffusion-weighted images. Several of the described abnormalities, particularly the earlier features of cortical hyperintensity, are best seen on fluid-attenuated inversion recovery and diffusion-weighted images. The superiority of diffusion-weighted imaging over other sequences, including fluid-attenuated inversion recovery imaging, in detecting basal ganglionic changes and cortical changes is not a new concept in cases of sCJD and has been previously documented in two small case series (5, 6). T1-weighted imaging results are usually normal in cases of sCJD, and contrast enhancement does not occur.

MR imaging research in association with CJD has been hampered by the rarity of the disease. Murata et al (7) have overcome this by grouping imaging from different subtypes of CJD with differing degrees of diagnostic certainty. Clinical criteria (history, EEG, and CSF analysis) in cases of CJD are not 100% accurate (7), which is why so much interest has been generated in MR imaging as a means to further improve diagnostic accuracy noninvasively. It is unknown whether it is valid to group sporadic and genetic cases of CJD together when assessing MR imaging results. That most of their genetic cases had an atypical clinical course and absent EEG changes adds weight to the growing evidence that they are distinct diseases. This limits the validity of extrapolating their results. When imaging studies lack the accepted criterion standard of pathologic confirmation, the methodology of image assessment becomes more critical. When possible, the scales used for image assessment should be validated by simple inter- and intraobserver studies, preferably including comparison with control cases. The methods can then be more easily reproduced by other centers and the findings fairly and independently validated.

Despite these limitations, the article by Murata et al is an important one. It is the first substantial study to chart the progression of changes in CJD with serial diffusion-weighted imaging. In their study of 13 patients, they convincingly confirmed that diffusion-weighted imaging is considerably more sensitive to basal ganglionic changes than are other sequences, including fluid-attenuated inversion recovery imaging, particularly during the earlier stages of the illness. They also documented progressive changes, with spread of signal intensity changes from caudate head to putamen. They then present the interesting hypothesis that this may reflect synaptic spread through gray matter bridges between the caudate head and the putamen. Despite the rather heterogeneous CJD case mix, Murata et al made a number of other important observations. Cortical diffusion restriction on diffusion-weighted images has previously been thought to be a very transient phenomenon. However, they documented several cases in which this had persisted for several weeks, potentially distinguishing the changes from infarction. In addition, it seems that during the early stages of the clinical disease, asymmetrical signal intensity changes are more prevalent than previously thought, and this has important implications in the interpretation of MR images performed early in cases of rapidly progressive dementia.

MR imaging has played an even more important role in the diagnosis of the most recently discovered subtype of CJD, vCJD. This form of CJD is clinically and neuropathologically distinct from sCJD and has been causally linked to bovine spongiform encephalopathy (“mad cow disease”) in cattle. One hundred seventeen cases of probable or definite vCJD have been diagnosed to date through the UK National CJD Surveillance Unit (Western General Hospital, Edinburgh, Scotland). vCJD is characterized by younger age at time of onset (median, 29 years, although a confirmed case has been identified recently in a 74-year-old man), longer duration of disease (median, 14 months), and a clinical picture differing from that of sCJD (8). The presenting symptoms are usually nonspecific, with sensory symptoms (sensation of cold, paraesthesia, or pain) and psychiatric symptoms (withdrawal, depression, fleeting delusions) occurring commonly, and the nonspecific nature of these symptoms often leads to late diagnosis. Other neurologic features include cerebellar signs, abnormal eye movement, and involuntary movements (myoclonus, chorea, dystonia).

The histopathology of vCJD is also distinct from that of sCJD, with astrocytosis and neuronal loss being particularly prominent in the thalamus and with characteristic florid plaques seen in the cerebral cortex. In cases of vCJD, the EEG and CSF 14-3-3 have a much lower sensitivity than in cases of sCJD. Tonsil biopsy may be useful but is invasive and has yet to be fully validated in a large group of patients. Criteria for the in vivo diagnosis of probable vCJD have been defined, and MR imaging now plays a pivotal role.

In the past, the diagnosis of vCJD has often been delayed and made only post mortem, but recent studies have shown that in cases of vCJD, MR imaging shows a characteristic distribution of symmetrical hyperintensity in the pulvinar of the thalamus. The degree of hyperintensity is greater than that seen in the putamen or cortical gray matter and is known as the pulvinar sign of vCJD (9). The pulvinar hyperintensity is best appreciated on axial images on which the thalamus, basal ganglia, and cortex can all be compared on a single section (10). The sign was originally described on T2-weighted and proton density-weighted images and was 80% sensitive and 100% specific in the appropriate patient group (11), but the sensitivity has increased to greater than 90% with improvements in MR imaging technology, particularly with wider availability of fluid-attenuated inversion recovery imaging. Very limited data are available regarding diffusion-weighted imaging in cases of vCJD, with only four of the first 100 cases having included diffusion-weighted imaging. Therefore, the value of this sequence in association with vCJD remains unclear. Quantification studies of the pulvinar sign are currently in progress, and serial studies using diffusion-weighted imaging as performed by Murata et al may also provide insights into how this disease enters the brain from the digestive system.

Although the differential diagnosis of hyperintensity in the thalamus is extensive, the presence of high signal intensity limited to the pulvinar of the thalamus in conditions other than vCJD is extremely rare. To date, those conditions with true pulvinar hyperintensity have all remained clinically clearly distinct from vCJD. Hyperintensity is also seen in the dorsomedial nuclei of the thalamus in 60–80% of cases of vCJD (depending on pulse sequence performed). However, unlike sCJD characterisitcs, cerebral atrophy is not prominent.

Although CJD is a rare disease, it has generated great interest, particularly because of the discovery that it may be transmitted iatrogenically, and more recently, from animals to humans by food products, with subsequent enormous commercial and political consequences. To date, it has remained a uniformly fatal disease with no specific treatment. From an imaging perspective, it is also unusual, with the recent discovery of characteristic deep gray matter nuclei changes, which are surprisingly sensitive and specific for the disease. Many unanswered questions remain regarding the pathogenesis and imaging of CJD, and because of the rarity of the disease, it is unlikely that major progress will be made without coordinated multicenter collaboration. However, diffusion-weighted imaging offers an exciting new and useful way of following the clinical course and is beginning to help our understanding of the pathogenesis of this disease.

• Copyright © American Society of Neuroradiology