Machine Learning Model for VTE Risk in Lung Transplant Recipients on ECMO

A Chinese study from the Second Affiliated Hospital of Zhejiang University retrospectively reviewed 189 patients who underwent elective lung transplantation between May 2023 and November 2024. The research developed a machine learning-based prediction model for venous thromboembolism risk in lung transplant recipients supported by ECMO. The Random Forest model demonstrated an AUC of 0.895 on the validation set.

Use Cases

• Developing clinical decision support tools based on the six identified predictive factors for VTE.
• Benchmarking machine learning models for binary medical outcomes based on reported performance metrics (accuracy, sensitivity, specificity).
• Studying risk factors for venous thromboembolism in the specific context of ECMO-supported lung transplantation.
• Validating model calibration and clinical utility using the described decision curve analysis methodology.

Strengths

• Model performance is quantitatively reported with multiple metrics, including an AUC of 0.895 and accuracy of 89.7%.
• The study used a defined cohort of 189 patients and employed a rigorous 5-fold cross-validation methodology.
• The model's clinical utility was assessed via decision curve analysis, as mentioned in the results.

Limitations

• The dataset is very small at 16.5 KB, suggesting limited raw data or a focus on model parameters rather than patient-level records.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for training other models.

Provenance

Source

Second Affiliated Hospital of Zhejiang University

Collection Method

Retrospective review of medical records

Time Range

May 2023 to November 2024

Freshness

Last updated 2026-06-04 05:47:25; freshness should be verified.

Geography

China