MRI-Based Patellofemoral Morphometrics for Knee Pathology Diagnosis in Adolescents

A dataset of 168 adolescents with anterior knee pain, containing 13 patellofemoral morphometric parameters measured from knee MRI scans. The data was collected in a retrospective single-center pilot study between January 2018 and December 2024 by author Dusan Spasic. It was used to model three binary MRI endpoints and compare logistic regression with machine-learning approaches.

Use Cases

• Training diagnostic models for patellar retinacular lesions based on patellofemoral morphometric parameters.
• Comparing logistic regression with gradient-boosting methods like CatBoost and LightGBM for medical prediction tasks.
• Evaluating the diagnostic utility of routine MRI measurements for composite chondromalacia in adolescents.
• Assessing model performance for a composite endpoint of ACL injury or patellar bone bruise using morphometric data.

Strengths

• Includes data from 168 adolescent patients (97 girls, 71 boys) with a mean age of 15.5 years.
• Contains 13 distinct patellofemoral morphometric parameters measured by two radiologists.
• Model performance for patellar retinacular lesions is quantified with specific metrics, including an AUC of 0.85 for a CatBoost model.
• Data collection spans a 6-year period from January 2018 to December 2024.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The study is a retrospective single-center pilot, which may limit generalizability.

Provenance

Source

figshare

Collection Method

Retrospective single-center pilot diagnostic modeling study using knee MRI scans.

Time Range

January 2018 to December 2024

Freshness

Last updated 2026-05-22 05:58:09; freshness should be verified.