Knee Morphology and Medial Meniscus Posterior Root Tear Risk Factors for 608 Patients

MinYuan Zhang's dataset contains clinical and MRI-based morphological parameters for 608 patients who underwent arthroscopic knee surgery. It includes 281 patients with confirmed medial meniscus posterior root tears and 327 control patients without. The data was used to train and evaluate 10 machine learning models, with results published on figshare in May 2026.

Use Cases

• Developing predictive models for medial meniscus posterior root tear risk based on patient age and tibial morphology.
• Assessing the non-linear contribution of medial tibial slope and medial tibial plateau depth to injury risk.
• Training explainable AI models for clinical decision support using SHAP analysis.
• Investigating the relationship between knee joint morphology and degenerative conditions.

Strengths

• Includes data from 608 patients, providing a substantial sample size for analysis.
• Features arthroscopically confirmed diagnoses for 281 cases, offering a reliable ground truth.
• Compares 10 different machine learning models, including CatBoost, XGBoost, and Neural Networks.
• Employs SHAP analysis for model explainability, identifying key thresholds for risk factors.

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 selection bias inherent to the retrospective cohort study design from a single surgical center.

Provenance

Source

MinYuan Zhang

Collection Method

Retrospective analysis of patients who underwent arthroscopic surgery for knee joint injuries.

Freshness

Last updated 2026-05-14 05:31:02; freshness should be verified.