Table 1_Integrating neuroimaging and plasma biomarkers to predict preclinical Alzheimer’s

Junjia Qi's dataset on figshare contains retrospective data from 320 patients with Aβ-positive preclinical Alzheimer's disease or mild cognitive impairment. The data was used to develop and validate multimodal machine learning models for predicting cognitive progression risk. The dataset was last updated in May 2026.

Use Cases

• Developing predictive models for cognitive progression based on neuroimaging and plasma biomarkers.
• Evaluating the clinical utility of multimodal biomarker models for Alzheimer's disease risk management.
• Training machine learning models using core variables like Mini-Mental State Examination scores and normalized hippocampal volume.
• Analyzing the influence of specific predictors, such as plasma phosphorylated tau181, using SHAP values.

Strengths

• Dataset includes 320 patient records, providing a substantial sample size.
• Model validation achieved an AUC of 0.818 in an independent validation set.
• Core predictive variables were identified through LASSO regression and multivariable analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect temporal or source bias inherent to the retrospective study design.

Provenance

Source

figshare

Collection Method

Retrospective study enrolling patients with Aβ-positive preclinical AD or AD-related mild cognitive impairment.

Freshness

Last updated 2026-05-05 13:26:55; freshness should be verified.