Intracardiac Thrombus Prediction in Anticoagulated Atrial Fibrillation Patients

A retrospective cohort of 149 atrial fibrillation patients undergoing catheter ablation procedures on continuous oral anticoagulation, used to develop a machine learning prediction model for intracardiac thrombus formations. The dataset includes clinical, echocardiographic, and p-wave morphology data from patients screened via transesophageal echocardiography. The model was developed by F. Ratajczak and published on figshare in 2026.

Use Cases

• Predicting thrombus formation risk based on clinical and echocardiographic parameters mentioned in the description
• Developing machine learning models for preprocedural decision-making in catheter ablation
• Comparing model performance against traditional CHA₂DS₂-VASc scoring
• Analyzing the relationship between p-wave characteristics and atrial substrate disease

Strengths

• Dataset is derived from a screened cohort of 1335 patients, providing a substantial initial population
• Machine learning models achieved a ROC AUC of 0.886 and 0.826, outperforming CHA₂DS₂-VASc (AUC 0.641) in the cohort
• Includes propensity-score-matched controls (93 TF+ and 93 TF- patients) for balanced analysis

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count for the final modeling subset (149 patients) is relatively small for machine learning applications
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

figshare

Collection Method

Retrospective cohort study of patients undergoing catheter ablation and transesophageal echocardiography.

Freshness

Last updated 2026-06-04 04:27:25; freshness should be verified