Decoding Fatigue in Systemic Lupus Erythematosus with Artificial Intelligence: Hematologic or Inflammatory Factors?

This study investigated the predictors of fatigue in 144 patients with systemic lupus erythematosus (SLE) using a random forest model. Hemoglobin level was identified as the most significant predictor of fatigue, with a threshold of 12.5 g/dL indicating increased fatigue risk. Other factors such as SLAM score, erythrocyte sedimentation rate, SLEDAI, and SLICC index also contributed to fatigue, though to a lesser extent. The model's predictive accuracy was reflected in an area under the ROC curve of 76.9%. The study highlights the importance of hematologic status, disease activity, and systemic inflammation in SLE-related fatigue.