Osteopenia Risk Prediction Model for Perimenopausal Women Based on Clinical Biomarkers

1,108 female participants aged 45 and above were used to develop a machine learning model for osteopenia risk. The random forest model identified 17 predictors, including menopause status, age, and specific biochemical markers, achieving an AUC of 0.933 on a validation set. The dataset, created by Xiaoling Zhuo and last updated in May 2026, is a retrospective single-center study with results published on figshare.

Use Cases

• Training a diagnostic model for osteopenia based on clinical and biochemical markers like PINP and β-CTX.
• Analyzing feature importance for osteoporosis risk factors using SHAP analysis as described in the study.
• Developing a web-based clinical calculator for early detection of bone density loss in perimenopausal women.
• Comparing the performance of machine learning algorithms (RF, LASSO, GBDT) on a clinical prediction task.

Strengths

• Model performance is quantitatively reported, with a random forest model achieving an AUC of 0.978 in training and 0.933 in validation.
• The study identifies 17 specific predictors for osteopenia, including menopause, age, PINP, β-CTX, height, and eGFR.
• Data is structured for machine learning with a clear 70/30 train-test split applied to 1,108 participant records.

Limitations

• Row count for the underlying dataset is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is from a single-center retrospective study, which may limit generalizability.

Provenance

Source

figshare, author Xiaoling Zhuo.

Collection Method

Retrospective single-center study of female participants aged 45+ with femoral neck BMD data.

Freshness

Last updated 2026-05-29 06:11:27; freshness should be verified.