Low cerebrospinal fluid sulfatide predicts progression of white matter lesions: The LADIS study.

BACKGROUND/AIMS: Demyelination and axonal degeneration are the hallmarks of established white matter lesions (WML). The neurochemistry of ongoing WML is only partially known. We explored cerebrospinal fluid (CSF) substances as markers of brain tissue damage in relation to progression of WML rated on magnetic resonance imaging. METHODS: CSF from elderly individuals with WML was analyzed for amyloid markers, total τ, hyperphosphorylated τ, neurofilament protein light subunit, sulfatide and CSF/serum-albumin ratio. After 3 years, a follow-up magnetic resonance imaging was performed. Progression of WML was rated using the Rotterdam Progression Scale (RPS). RESULTS: 37 subjects (age 73.6 ± 4.6 years) were included. Subjects with more pronounced progression (RPS > 2; n = 15) had lower mean sulfatide concentration at baseline as compared to subjects with no or minimal progression (RPS 0-2; n = 22) according to univariate analyses (p = 0.009). Sulfatide was the only biomarker that predicted the RPS score according to regression analysis, explaining 18.9% of the total variance (r = 0.38, p = 0.015). CONCLUSION: The correlation of CSF sulfatide levels and RPS scores may reflect a remyelination response to the demyelination process associated with WML. Furthermore, the results strengthen the notion that WML pathology is different from that of Alzheimer's disease.

Monte Carlo feature selection and rule-based models to predict Alzheimer's disease in mild cognitive impairment.

The objective of the present study was to evaluate a Monte Carlo feature selection (MCFS) and rough set Rosetta pipeline for generating rule-based models as a tool for comprehensive risk estimates for future Alzheimer's disease (AD) in individual patients with mild cognitive impairment (MCI). Risk estimates were generated on the basis of age, gender, Mini-Mental State Examination scores, apolipoprotein E (APOE) genotype and the cerebrospinal fluid (CSF) biomarkers total tau (T-tau), phospho-tau(181) (P-tau) and the 42 amino acid form of amyloid ß (Aß42) in two sets of longitudinally followed MCI patients (n = 217 in total). The predictive model was created in Rosetta, evaluated with the standard tenfold cross-validation approach and tested on an external set. Features were ranked and selected by the MCFS algorithm. Using the combined pipeline of MCFS and Rosetta, it was possible to predict AD among patients with MCI with an area under the receiver operating characteristics curve of 0.92. Risk estimates were produced for the individual patients and showed good correlation with actual diagnosis in cross validation, and on an external dataset from a new study. Analysis of the importance of attributes showed that the biochemical CSF markers contributed the most to the predictions, and that added value was gained by combining several biochemical markers. Despite a correlation with the biochemical markers, the genetic marker APOE ε4 did not contribute to the predictive power of the model.