Machine learning for predicting thrombotic recurrence in antiphospholipid syndrome

Area of Science:

• Medical research
• Machine learning applications
• Autoimmune disorders

Background:

• Thrombotic antiphospholipid syndrome (TAPS) presents a high risk of recurrent thromboembolic events.
• Current anticoagulation treatments face challenges in predicting event recurrence.
• Enhanced risk stratification is crucial for optimizing personalized TAPS management.

Purpose of the Study:

• To evaluate the predictive performance of the extreme gradient boosting (XGB) model for recurrent thrombotic events in TAPS.
• To compare XGB with other machine learning algorithms for TAPS risk prediction.
• To identify key clinical and biochemical predictors of TAPS recurrence.

Main Methods:

• Utilized demographic and clinical data for model development.
• Assessed model performance using metrics like accuracy, specificity, precision, and AUC.
• Compared XGB against various machine learning algorithms.

Main Results:

• XGB demonstrated superior performance, achieving the highest accuracy and AUC.
• Identified renal impairment, age, and lupus anticoagulant as significant predictors.
• XGB showed robust predictive capabilities for TAPS recurrence.

Conclusions:

• XGB shows significant potential for improving risk stratification in TAPS.
• The model can aid clinical decision-making and optimize anticoagulation strategies.
• Further validation in larger, prospective studies is recommended for clinical integration.