A dynamic prediction model for predicting the time at which patients with MCI progress to AD based on time-dependent covariates
Abstract Background Alzheimer’s Disease (AD) is an irreversible neurodegenerative disorder that imposes a significant burden on families and society. Timely intervention during the transitional stages from Mild Cognitive Impairment (MCI) to AD can help mitigate this issue. The MCI-to-AD conversion t...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-07-01
|
| Series: | BMC Medical Informatics and Decision Making |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12911-025-03040-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background Alzheimer’s Disease (AD) is an irreversible neurodegenerative disorder that imposes a significant burden on families and society. Timely intervention during the transitional stages from Mild Cognitive Impairment (MCI) to AD can help mitigate this issue. The MCI-to-AD conversion time would be helpful if it could be predicted. Most studies rely on Cox models, which possess certain limitations and do not intuitively forecast the duration until patients with MCI progress to AD. Thus we construct a new dynamic prediction model based on the conditional restricted mean survival time (cRMST) from a time-scale perspective to explore the factors influencing progression to AD in patients with MCI and predict the average time required MCI patients to progress to AD at different time points in the future. Methods We construct a new two-stage dynamic prediction model (tRMST model) based on the conditional restricted mean survival time (cRMST) in combination with landmark method to apply in the analysis of the ADNI database. Results The results of the ADNI analysis showed that four variables (Education, MMSE, ADAS-Cog13 and P-tau) have dynamic effects over time. The C-index and the mean prediction error of the cross validation are better than the static RMST model. Conclusion This study presents a time-scale dynamic prediction model that effectively leverages longitudinal data to identify the dynamic effects of the factors’ impact on the outcome over time, thereby assisting physicians in personalizing treatment for patients. |
|---|---|
| ISSN: | 1472-6947 |