A prognostic model of Alzheimer's disease relying on multiple longitudinal measures and time-to-event data

Alzheimers Dement. 2018 May;14(5):644-651. doi: 10.1016/j.jalz.2017.11.004. Epub 2018 Jan 4.


Introduction: Characterizing progression in Alzheimer's disease is critically important for early detection and targeted treatment. The objective was to develop a prognostic model, based on multivariate longitudinal markers, for predicting progression-free survival in patients with mild cognitive impairment.

Methods: The information contained in multiple longitudinal markers was extracted using multivariate functional principal components analysis and used as predictors in the Cox regression models. Cross-validation was used for selecting the best model based on Alzheimer's Disease Neuroimaging Initiative-1. External validation was conducted on Alzheimer's Disease Neuroimaging Initiative-2.

Results: Model comparison yielded a prognostic index computed as the weighted combination of historical information of five neurocognitive longitudinal markers that are routinely collected in observational studies. The comprehensive validity analysis provided solid evidence of the usefulness of the model for predicting Alzheimer's disease progression.

Discussion: The prognostic model was improved by incorporating multiple longitudinal markers. It is useful for monitoring disease and identifying patients for clinical trial recruitment.

Keywords: ADNI; External validation; Mild cognitive impairment; Multivariate functional component analysis; Prediction.

Publication types

  • Observational Study
  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Alzheimer Disease / diagnostic imaging*
  • Biomarkers* / cerebrospinal fluid
  • Brain / physiopathology
  • Cognitive Dysfunction / diagnostic imaging*
  • Disease Progression*
  • Early Diagnosis
  • Female
  • Humans
  • Longitudinal Studies
  • Male
  • Models, Statistical
  • Neuroimaging / methods
  • Neuropsychological Tests / statistics & numerical data
  • Prognosis


  • Biomarkers