習慣流産における脈絡膜下血腫の予測因子としての免疫学的および凝固学的予測因子：予測モデリングアプローチ
Xue Wei1, Cai Liu1, Weijing Wang1
1Department of Reproductive Medicine, Lanzhou University Second Hospital, Lanzhou, China.
Problem:
Subchorionic hematoma (SCH) is a common complication in early pregnancy, particularly in women with recurrent pregnancy loss (RPL), and is associated with adverse pregnancy outcomes. However, reliable predictive tools for SCH in this high-risk group are lacking. The aim of this study was to identify coagulation and immune-related predictors associated with SCH in women with RPL and to construct a predictive model.
Method Of Study:
A retrospective analysis was performed on 1002 patients with RPL in the Reproductive Medicine Center of the Second Hospital of Lanzhou University. Clinical data, coagulation indicators, thromboelastography (TEG), and immunity indicators were collected. Predictors were selected via LASSO and logistic regression, and a LightGBM model was developed and validated.
Results:
Eleven predictors were included in the final model: age, BMI, previous pregnancy losses, TEG-lysis index at 30 min, TEG- maximum amplitude, serum immunoglobulins IgG, complement C3, antiβ2 glycoprotein 1 antibody, rheumatoid factor IgM and interleukin-2/interleukin-6, interferon-γ/interleukin-6. The model demonstrated moderate and stable discrimination with AUCs of 0.745, 0.653, and 0.647 in the training, internal, and external validation sets, respectively.
Conclusions:
The model effectively identifies SCH risk in RPL patients and informs individualized monitoring and early intervention strategies in clinical practice.
