Retrospective study on affecting factors and predictive model for guide wire failure to pass through the foramen ovale in PFO patients during interventional closure

Patent foramen ovale (PFO) is a common congenital cardiac anomaly. Transcatheter closure is effective, but guidewire passage failure remains a challenge. We aimed to identify risk factors and develop a predictive model for procedural failure. This retrospective study included 83 PFO patients (October 2021–May 2024) undergoing interventional closure, divided into success (n = 52) and failure (n = 31) groups. Multivariate logistic regression analyzed risk factors, and a nomogram was developed. Model calibration was assessed using the Hosmer–Lemeshow test and calibration curves, while discriminative ability was evaluated via receiver operating characteristic analysis. Predictive accuracy was validated through 5-fold cross-validation, and clinical utility was examined using decision curve analysis and clinical impact curve. Left atrial aperture ≤1.8 mm was a significant risk factor for failure (odds ratio [OR] = 10.691, P < .05), while contrast transthoracic echocardiography-determined right-to-left shunt grade II (OR = 0.147) and III (OR = 0.056), and surgical team experience >3 years (OR = 0.077) were protective (all P < .05). The model was well calibrated (χ² = 4.002, P = .857), and the area under receiver operating characteristic curve was 0.913, indicating medium to strong predictive ability. The mean area under the curve for 5-fold cross-validation was 0.878, and decision curve analysis confirmed clinical utility at threshold probabilities of 0.01 to 0.85. The clinical impact curve demonstrated high consistency between predicted and actual outcomes when the threshold probability exceeded 60%. The model has good calibration, accuracy, distinguishing power, and clinical applicability, and provides a valuable tool for preoperative risk assessment of PFO occlusion.