Abstract:Background: Identifying high-risk individuals with mild cognitive impairment (MCI) who are likely to progress to Alzheimer's disease (AD) is crucial for early intervention. Objective: This study aimed to develop and validate a novel clinical score for personalized estimation of MCI-to-AD conversion. Methods: The data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study were analyzed. Two-thirds of the MCI patients were randomly assigned to a training cohort (n = 478), and the remaining one-third formed the validation cohort (n = 239). Multivariable logistic regression was performed to identify factors associated with MCI-to-AD progression within 4 years. A prediction score was developed based on the regression coefficients derived from the logistic model and tested in the validation cohort. Results: A lipidomics-signature was obtained that showed a significant association with disease progression. The MCI conversion scoring system (ranged from 0 to 14 points), consisting of the lipidomics-signature and five other significant variables (Apolipoprotein ɛ4, Rey Auditory Verbal Learning Test immediate and delayed recall, Alzheimer's Disease Assessment Scale delayed recall test, Functional Activities Questionnaire, and cortical thickness of the AD signature), was constructed. Higher conversion scores were associated with a higher proportion of patients converting to AD. The scoring system demonstrated good discrimination and calibration in both the training cohort (AUC = 0.879, p of Hosmer-Lemeshow test = 0.597) and the validation cohort (AUC = 0.915, p of Hosmer-Lemeshow test = 0.991). The risk classification achieved excellent sensitivity (0.84) and specificity (0.75). Conclusions: The MCI-to-AD conversion score is a reliable tool for predicting the risk of disease progression in individuals with MCI.

A Novel Score to Predict Individual Risk for Future Alzheimer's Disease: A Longitudinal Study of the ADNI Cohort