Abstract
OBJECTIVE: Current prediction of pregnancy outcomes in Recurrent Spontaneous Miscarriage (RSM) lacks multidimensional models integrating hemodynamic and immunological parameters. This study aimed to construct a predictive model for pregnancy failure in RSM based on uterine hemodynamic and immune microenvironment indicators, providing a foundation for precision clinical management. METHODS: A retrospective cohort study was conducted, enrolling 400 RSM patients (240 successful pregnancy cases, 160 failure cases) and 100 matched controls from January 2020 to July 2025 Core predictors were selected through univariate logistic regression, LASSO dimension reduction, and multivariate logistic regression. Model performance was evaluated using AUC, calibration, and decision curve analysis. RESULTS: Ten core predictors were identified, including hemodynamic parameters (uterine artery resistance index [RI], pulsatility index [PI], peak systolic velocity [PSV]) and immune microenvironment indicators (regulatory T cell [Treg] proportion, CD56⁺CD16⁺ NK cell proportion). The multivariate model demonstrated an AUC of 0.809 (95% CI: 0.790–0.862), sensitivity of 88.13%, specificity of 89.58%, and non-significant Hosmer-Lemeshow test (P = 0.599). Decision curve analysis confirmed clinical net benefit within the 10%-90% threshold probability range. CONCLUSION: This first hemodynamic-immunological integrated model accurately predicts RSM pregnancy outcomes, revealing synergistic mechanisms between placental perfusion disorders and immune tolerance breakdown, thereby informing targeted interventions.