Importance: Sporadic Creutzfeldt-Jakob disease (sCJD) is a rapidly lethal disease. Rapid, accurate diagnosis is imperative for epidemiological surveillance and public health activities to exclude treatable differentials and facilitate supportive care. In 2017, the International CJD Surveillance Network diagnostic criteria were revised to incorporate cortical ribboning on magnetic resonance imaging and the real-time quaking-induced conversion (RT-QuIC) assay, developments that require multicenter evaluation.
Objective: To evaluate the accuracy of revised diagnostic criteria through the retrospective diagnosis of autopsy-confirmed cases (referred to as in-life diagnosis).
Design, setting, and participants: This diagnostic study used a 3-year clinicopathological series using all cases of autopsy-confirmed sCJD and a noncase group with alternative neuropathological diagnoses from national surveillance centers in the United Kingdom, France, Germany, and Italy. Data were collected from January 2017 to December 2019 and analyzed from January 2020 to November 2021.
Main outcomes and measures: Sensitivity and specificity of revised diagnostic criteria and diagnostic investigations. Secondary analyses assessing sCJD subgroups by genotype, pathological classification, disease duration, and age.
Results: A total of 501 sCJD cases and 146 noncases were included. Noncase diagnoses included neurodegenerative diseases, autoimmune encephalitis, and cerebral insults such as anoxia. Participants in the sCJD cases cohort were younger (mean [SD] age, 68.8 [9.8] years vs 72.8 [10.9] years; P < .001) and had longer median (IQR) disease duration (118 [74.8-222.3] days vs 85 [51.5-205.5] days; P = .002); sex ratios were equivalent (253 [50.5%] male cases vs 74 [50.7%] male noncases). Sensitivity of revised criteria in in-life diagnosis (450 of 488 [92.2%] diagnoses; 95% CI, 89.5%-94.4%) was increased compared with prior criteria (378 of 488 [77.5%] diagnoses; 95% CI, 73.5%-81.1%; P < .001), while specificity (101 of 125 [80.8%] diagnoses; 95% CI, 72.8%-87.3%) was unchanged (102 of 125 [81.6%] diagnoses; 95% CI, 73.7%-88.0%; P > .99). Among 223 cases and 52 noncases with the full panel of investigations performed, sensitivity of revised criteria (97.8%; 95% CI, 94.9%-99.3%) was increased compared with prior criteria (76.2%; 95% CI, 70.1%-81.7%; P < .001) while specificity was unchanged (67.3%; 95% CI, 52.9%-79.7% vs 69.2%; 95% CI, 54.9%-81.3%; P > .99). In 455 cases and 111 noncases, cortical ribboning was 67.9% sensitive (95% CI, 63.4%-72.2%) and 86.5% specific (95% CI, 78.7%-92.2%). In 274 cases and 77 noncases, RT-QuIC was 91.6% sensitive (95% CI, 87.7%-94.6%) and 100% specific (95% CI, 96.2%-100%). Investigation sensitivity varied with genetic and pathological features, disease duration, and age.
Conclusions and relevance: This diagnostic study demonstrated significantly improved sensitivity of revised sCJD diagnostic criteria with unaltered specificity. The revision has enhanced diagnostic accuracy for clinical care and surveillance.