How prion disease is diagnosed: more detail

Which criteria are in use?

Classification follows surveillance criteria, which combine the clinical syndrome with test results. The CDC publishes them as three levels.

Definite sporadic CJD requires brain tissue: standard neuropathology, immunocytochemistry, western blot showing protease-resistant PrP, or scrapie-associated fibrils.

Probable sporadic CJD can be reached by either of two routes. The first is a neuropsychiatric disorder together with a positive RT-QuIC assay in cerebrospinal fluid or other tissues. The second is rapidly progressive dementia with at least two of four clinical features, together with at least one positive result from EEG, the CSF 14-3-3 assay or MRI, and with no routine investigation pointing to another diagnosis. The four features are myoclonus, visual or cerebellar signs, pyramidal or extrapyramidal signs, and akinetic mutism.

Possible sporadic CJD is progressive dementia with at least two of those features, illness duration under two years, no alternative diagnosis on routine investigation, and none of the positive laboratory results that would make the case probable.

Two points follow from the structure of the criteria. One of the four clinical features, akinetic mutism, appears late in the illness, so a syndrome-based route to classification tends to lag behind the clinical question a family is asking. The other is that the EEG, 14-3-3 and MRI route requires the dementia syndrome, while the RT-QuIC route does not, which makes the assay the most useful single investigation in atypical presentations.

The 2021 Lancet Neurology review by Hermann and colleagues, written with authors from the surveillance centres, set out the amended criteria. Its figure describing them records that they were adapted from the National CJD Research and Surveillance Unit criteria, themselves based on the earlier WHO criteria, amended by RT-QuIC as an additional biomarker, with the imaging criteria refined. The authors describe RT-QuIC as a major advance for confident diagnosis before death, and state that updated criteria including it should improve early clinical confirmation, surveillance, assessment of seeding activity in different tissues, and trial monitoring.

MRI, EEG and the indirect CSF markers

MRI is the imaging test that carries diagnostic weight. The CDC criterion is met by high signal in the caudate or putamen, or in at least two cortical regions among temporal, parietal and occipital, on diffusion-weighted or FLAIR sequences. The sequence type affects what can be seen, so it is worth confirming that diffusion-weighted imaging was acquired and that a radiologist was told prion disease is in the differential.

EEG counts when it shows the typical periodic sharp wave complexes, at any stage of the illness. A normal or non-specific EEG argues against nothing.

CSF 14-3-3 counts as a positive result when the illness has lasted less than two years. Both 14-3-3 and total tau mark neuronal injury from any cause, and the 2021 review notes that conflicting studies have made some of these surrogate markers controversial. Raised total tau supports the picture, and neither marker distinguishes prion disease from other rapid neurodegeneration on its own.

RT-QuIC performance, and where it is weaker

RT-QuIC amplifies the misfolded prion protein itself, which is why it behaves differently from the injury markers.

Published figures differ with cohort and protocol. The United States national reference laboratory reports diagnostic sensitivity of 90.3 percent and specificity of 98.5 percent for CSF RT-QuIC, citing a 2020 Neurology paper, and gives an analytic turnaround of five working days with no urgent option. The 2022 review by Poleggi and colleagues reports around 96 percent sensitivity and close to 100 percent specificity in CSF and in olfactory mucosa. Taken together the sensitivity figures sit in the low 90s to mid 90s, with specificity near or above 98 percent.

Sensitivity is not uniform. The same review records lower sensitivity in the atypical sporadic forms, naming VPSPr, MM2-thalamic, MM2-cortical and VV1, and in some genetic forms including fatal familial insomnia and Gerstmann-Straussler-Scheinker disease. A negative RT-QuIC in a person whose presentation suggests one of those subtypes should not close the question, and a repeat lumbar puncture is a reasonable step when the clinical picture persists.

Olfactory mucosa brushing is the other sampled site with comparable reported accuracy. It is a specialist procedure, and brushings carry higher seeding activity than CSF, which has prompted attention to handling of instruments and nasal secretions.

Skin testing

A 2025 review in Molecular Neurodegeneration describes RT-QuIC on skin biopsy. In the prion study it summarises, biopsies were taken from four sites: near the ear, the upper arm, the lower back and the inner thigh. Single-site sensitivity ranged from 80.0 percent at the upper arm to 84.9 percent near the ear. Combining any two of the ear, inner thigh and lower back samples gave 92.1 percent, and pooling all sites gave 95.0 percent, against 75.5 percent for CSF alone in the same cohort. The review states that skin RT-QuIC was more accurate than CSF for prion disease diagnosis in that study.

The comparison group of non-prion patients numbered 23, which limits what can be said about specificity, and the review lists open questions about which sites are most useful and when prion protein appears in skin relative to brain. Skin RT-QuIC is not part of the surveillance criteria.

Genetic testing

In a symptomatic person, sequencing the prion protein gene is a diagnostic test and is usually arranged by the neurology team, particularly where there is a family history or an unusually young onset.

Predictive testing in an asymptomatic relative is a different undertaking, and it begins with genetic counselling, before any sample is taken. A 2026 practice framework from an Israeli expert panel sets out one worked approach. All asymptomatic relatives who want testing should have access to counselling, which may be given by a medical geneticist, a genetic counsellor or a neurologist familiar with the disease. Testing is not offered under the age of 18. People in acute distress, such as shortly after a relative’s diagnosis or death, are encouraged to wait at least six months. The process runs over four steps and at least one month from counselling to blood sampling, includes a structured psychological assessment with written confirmation of readiness, and is postponed where readiness is in doubt. Results, positive or negative, are given in person by the clinician who provided the counselling, with a written summary and contact details where the result is positive. The panel also recommends that health systems build a structured support programme for carriers and their families, and notes that Israel did not have one at the time of writing.

The same panel places preimplantation and prenatal genetic testing on a case-by-case footing. It records that exclusion testing, which selects embryos that did not inherit the at-risk allele from an affected grandparent while leaving the parent’s own status unknown, is available at tertiary centres there, that prenatal diagnosis is possible by chorionic villus sampling at about 11 to 12 weeks or amniocentesis at about 17 to 23 weeks, and that decisions about ending a pregnancy raise ethical questions the framework does not resolve.

Deciding not to be tested is a settled and respected choice, and counselling exists to support either decision, with no push in either direction.

What is coming

A 2026 Lancet Neurology review by Zerr, Hermann, Mok and Mead reports that seeded amplification assays have improved diagnostic accuracy and that blood-based diagnostic biomarkers are close to clinical application. It notes work in preclinical carriers of pathogenic variants with a focus on non-invasive sampling, biomarkers tested in blood, tear fluid and urine, and the possibility of markers that predict onset or track the clinical course. The authors call for a biomarker-based staging system to be evaluated in future clinical studies.

Research on diagnostics follows that work in detail, and the Atlas tests pages hold the records for each assay.

Practical points for a referring team

Specialist review sits with the national centre. In the United Kingdom, the National CJD Research and Surveillance Unit in Edinburgh runs the national reference RT-QuIC laboratory for CSF and offers neuropathology advice, with clinical advice on suspected cases from the National Prion Clinic. Its pathology service lists named contacts for neuropathological enquiries, and arrangements for specialist nursing support are given to referring clinicians.

Nothing here can classify an individual case. Interpretation of results, and any decision about predictive testing in a relative, belongs with the treating neurologist and a genetic counsellor.