Abstract
The diagnosis of neurodegenerative diseases is made complex by the heterogenous phenotype of the patients and the regular occurrence of concomitant pathology. Studying clinicopathological correlations in autopsy series is a central approach to improve pathological prediction in clinical practice. However, such method requires a wealth of information, and the use of standard spreadsheet software is hardly suitable. To overcome this constraint, we designed a customizable and freely available neuropathology form with 456 data entry fields driven by an open-source DataBase Management Systems (DBMS) using Structured Query Language (SQL). This approach allowed us to optimize the compilation of clinical and pathological data from our brain collection (264 autopsied patients, 22,885 data points). Information was then easily retrieved using general and specific queries, facilitating the analysis of demographics, clinicopathological correlations, and incidental and concomitant proteinopathies. Tau, amyloid-β and α-synuclein incidental pathology was observed in respectively 78.1%, 42.8%, and 10.7% of all the patients. These proportions increased with age, reaching 100% for Tau pathology after 80. Concomitant proteinopathy was observed in 46.4% of the patients diagnosed with neurodegenerative diseases and prion disease. We observed a particularly high rate of co-pathology in patients with Dementia with Lewy bodies (81.3% of associated Tau and amyloid-β pathology) and Creutzfeldt-Jakob disease (68.4% of associated Tau pathology). Finally, we used specific queries to identify old cases that could meet newly defined neuropathological criteria and revised the diagnosis of a 90-year-old patient to LATE Stage 2. Increasing our understanding of clinicopathological correlations in neurodegenerative diseases is crucial given the implications in clinical diagnosis, biomarker identification and targeted therapies assessment. The precise characterization of clinical and pathological data of autopsy series remains a central approach but the large amount of generated data should encourage a more systematic use of DBMS.