A Nomogram For Predicting PFO-related Stroke Recurrence

Background: A number of clinical trials in recent years have demonstrated a link between patent foramen ovale(PFO)and stroke.There is a trend to differentiate PFO-related stroke from cryptogenic stroke.However,the risk factors for recurrence of such strokes,particularly biological markers,are still poorly reported and are not uniformly understood.Clinicians cannot sufficiently identify,treat,and follow-up the patients at high risk of recurrence.Therefore,this study aimed to explore risk factors for PFO-related stroke recurrence,establish a prediction model,and help improve the prognosis of patients with PFO-related stroke.Methods: This study included 392 patients with PFO-related stroke in a training set and164 patients with PFO-related stroke in a validation set.In the training set,Nomogram1 was created by forward stepwise Cox regression and Nomogram2 was created by least absolute shrinkage and selection operator(LASSO)regression.Nomogram performance was assessed using concordance index(C-index),receiver operating characteristic(ROC)curve,area under the curve(AUC),calibration curve,survival curve analysis,decision curve analysis(DCA)and tested in the validation set.Results: Nomogram1 was based on albumin(ALB),high-sensitivity C-reactive protein(hsCRP),homocysteine(Hcy)and Nomogram2 was based on Age,Hypertension,Diabetes,right-to-left shunt,ALB,prealbumin,hsCRP and Hcy.The C-index of Nomogram1 was 0.861,which was not significantly different from Nomogram2(0.893).Nomogram1 showed good 2-year and 5-year ROCs with AUCs at 0.863 and 0.777 respectively.The calibration curve confirmed good homogeneity between the predicted risk by Nommogram1 and actual risk.Survival curve analysis verified the risk stratification capability of Nomogram1.DCA demonstrated that Nomogram1 has reliable clinical utility.In the validation set,Nomogram1 generally outperformed the training set in all respects.Conclusions: Nomogram1 based on ALB,hsCRP,and Hcy levels provided a more clinically realistic prognostic prediction.This model could help patients with PFO-related stroke to facilitate personalized prognostic evaluations.