Identification of and a Prediction Model for Cognitive Trajectories in Patients With Mild Cognitive Impairment Due to Alzheimer’s Disease

Abstract

Background: Few studies have investigated cognitive trajectories or developed a prediction model for amyloid beta-positive (Aβ+) mild cognitive impairment (MCI) patients. We aimed to identify distinct cognitive trajectories in Aβ+ MCI patients based on longitudinal Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-cog) 13 scores. Furthermore, we aimed to develop and visualize a prediction model to predict trajectory groups using the demographic, genetic, and clinical biomarkers of Aβ+ MCI patients.Methods: We performed a retrospective analysis of the data in 238 Aβ+ MCI patients from the Alzheimer’s Disease Neuroimaging Initiative who underwent at least three rounds of annual neuropsychological testing to identify cognitive trajectories. A group-based trajectory model (GBTM) was used to classify distinct groups based on ADAS-cog 13 scores. The prediction model was estimated using multinomial logistic regression and visualized using a bar-based method for risk prediction. Results: Three distinct classes, namely slow decliners (18.5%), intermediate decliners (42.9%), and fast decliners (38.7%), were suggested. Intermediate decliners were associated with higher age (≥70 years) (odds ratio [OR] 2.72, 95% confidence interval [CI] 1.78-6.28), higher AV45 standardized uptake value ratios (SUVRs)*10 (OR 1.69, 95% CI 1.22-2.34), and lower fluorodeoxyglucose (FDG) SUVR*10 (OR 0.65, 95% CI 0.46-0.93) than slow decliners. Fast decliners were associated with higher age (≥70 years) (OR 3.76, 95% CI 1.40-10.10), greater likelihood of being an apolipoprotein E 4 carrier (OR 4.2, 95% CI 1.53-11.58), higher AV45 positron emission tomography SUVR*10 (OR 2.14, 95% CI 1.50-3.05), and lower FDG SUVR*10 (OR 0.31, 95% CI 0.20-0.48) than slow decliners. The predicted probability of being classified to a trajectory group according to the risk scores of predictors was visualized.Conclusions: Our GBTM analysis yielded novel insights into the heterogeneous cognitive trajectories among Aβ+ MCI patients, which further facilitates the effective stratification of Aβ+ MCI patients in Aβ-targeted clinical trials.