A Simulation-Based Validation Study of Incremental Predictive Value for Alzheimer's Disease Prediction Models
DOI:
https://doi.org/10.54097/4jttx317Keywords:
Alzheimer’s disease, Cox proportional hazards model, Monte Carlo.Abstract
Alzheimer's disease (AD) is the leading cause of dementia worldwide, and its pathological changes often occur several years before the onset of clinical symptoms. Therefore, early identification of individuals with high risk of progression has important clinical significance for subsequent timely intervention and disease management. However, the cumulative predictive value of a single biomarker and the combined use of multiple indicators under different modeling conditions and risk scenarios is still insufficiently quantified. In this study, a Monte Carlo simulation framework was constructed based on the ADEMP (Aims,Data-generating mechanism, Estimand, Methods, Performance measures) structure to systematically evaluate the incremental contribution of four plasma biomarkers (GFAP, pTau181, pTau217 and Aβ42 / 40) to AD risk prediction. We generated survival data covering a variety of scenarios, including demographic variables, cognitive assessment scores, and biomarker measurements. The Cox proportional hazards model was used in the analysis. The basic model was compared with the extended enhanced model with biomarkers, and the predictive performance was evaluated by C index, ΔC index, ΔR2, IDI, risk stratification analysis and sensitivity test. The results showed that GFAP and pTau217 were the most prominent in independent prediction, and the improvement of prediction effect was the most significant when the four markers were combined. This incremental effect remains stable under different sample sizes and censoring rates, indicating that the conclusion has strong robustness. The above findings provide strong support for integrating plasma biomarkers into traditional AD prediction models, and also build a reference methodological framework for future preclinical evaluation of the value of new biomarkers. There are some limitations in this study. It uses simulated data rather than real clinical cases, which cannot fully reflect the heterogeneity and complexity of patient data in the real world. Future research should conduct external validation of the above results, and further explore more flexible models such as adaptive ensemble estimation, in order to ultimately achieve truly personalized AD risk assessment.
Downloads
References
[1] Bun S, Ito D, Tezuka T, et al. Performance of plasma Aβ42/40, measured using a fully automated immunoassay, across a broad patient population in identifying amyloid status. Alzheimer‘s Research & Therapy. 2023;15(1):149. doi:10. 1186/s13195-023-01296-5 DOI: https://doi.org/10.1186/s13195-023-01296-5
[2] Austin PC, Pencinca MJ, Steyerberg EW. Predictive accuracy of novel risk factors and markers: A simulation study of the sensitivity of different performance measures for the Cox proportional hazards regression model. Statistical Methods in Medical Research. 2015;26(3):1053-1077. doi:10. 1177/0962280214567141 DOI: https://doi.org/10.1177/0962280214567141
[3] Honey MIJ, Van Maurik IS, Van Harten AC, et al. Individualized prediction of clinical progression to dementia using plasma biomarkers in non-demented elderly. Alzheimer‘s Research & Therapy. 2025;18(1):4. doi:10. 1186/s13195-025-01925-1 DOI: https://doi.org/10.1186/s13195-025-01925-1
[4] Morris TP, White IR, Crowther MJ. Using simulation studies to evaluate statistical methods. Statistics in Medicine. 2019;38(11):2074-2102. doi:10. 1002/sim. 8086 DOI: https://doi.org/10.1002/sim.8086
[5] Palmqvist S, Stomrud E, Cullen N, et al. An accurate fully automated panel of plasma biomarkers for Alzheimer’s disease. Alzheimer‘s & Dementia. 2022;19(4):1204-1215. doi:10. 1002/alz. 12751 DOI: https://doi.org/10.1002/alz.12751
[6] Salvadó G, Janelidze S, Bali D, et al. Plasma phosphorylated tau 217 to identify preclinical Alzheimer disease. JAMA Neurology. 2025;82(11):1122. doi:10. 1001/jamaneurol. 2025. 3217
[7] Wang X, Shi Z, Qiu Y, Sun D, Zhou H. Peripheral GFAP and NfL as early biomarkers for dementia: longitudinal insights from the UK Biobank. BMC Medicine. 2024;22(1):192. doi:10. 1186/s12916-024-03418-8 DOI: https://doi.org/10.1186/s12916-024-03418-8
[8] Sanusi B, Cai J, Hudgens MG. Nonparametric estimation of marked survival data in the presence of dependent censoring. Statistics in Medicine. 2023;42(12):1995-2008. doi:10. 1002/sim. 9710 DOI: https://doi.org/10.1002/sim.9710
[9] An S, Eo W, Lee S. Prognostic significance of modified Shine and Lal index in patients with non-small cell lung cancer undergoing surgical resection. Biomedicines. 2025;13(4):937. doi:10. 3390/biomedicines13040937 DOI: https://doi.org/10.3390/biomedicines13040937
[10] Andrade C. Survival analysis, Kaplan-Meier curves, and Cox regression: basic concepts. Indian Journal of Psychological Medicine. 2023;45(4):434-435. doi:10. 1177/02537176231176986 DOI: https://doi.org/10.1177/02537176231176986
[11] Shi Y, Zhang L, Do KA, Jenq RR, Peterson CB. survivalContour: visualizing predicted survival via colored contour plots. Bioinformatics Advances. 2024;4(1):vbae105. doi:10. 1093/bioadv/vbae105 DOI: https://doi.org/10.1093/bioadv/vbae105
[12] Muralidar S, Ambi SV, Sekaran S, Thirumalai D, Palaniappan B. Role of tau protein in Alzheimer‘s disease: the prime pathological player. International Journal of Biological Macromolecules. 2020;163:1599-1617. doi:10. 1016/j. ijbiomac. 2020. 07. 327 DOI: https://doi.org/10.1016/j.ijbiomac.2020.07.327
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

