Skip to main content
. 2022 Mar 15;4(2):fcac040. doi: 10.1093/braincomms/fcac040

Figure 4.

Figure 4

Prediction of dementia in Parkinson disease within 10 years, by logistic regression. (A) Receiver operating characteristic curves with AUC for predicting in Parkinson disease (PDD) with the clinical model and the combined clinical-biomarker model. Asymptotic significance for having a better predictive capability than chance, P < 0.001 (both models). (B) List of factors (using three factors versus two factors, to the left and right, respectively) in the logistic regression model for prediction of dementia. (C) A risk chart with the individual risk of dementia within 10 years, shown for a patient in the middle value of each tertile of the risk factors shown in the predictive models in B. MCI, mild cognitive impairment; UPDRS, Unified Parkinson’s Disease Ranking Scale (part III was the motor subscale used for motor scores); CSF Aβ42, amyloid-β42 in CSF.