Optimizing patient selection for patent foramen ovale closure after stroke: The CRISP-PFO score.

Abstract

Although patent foramen ovale (PFO) transcatheter closure lowers recurrent cryptogenic stroke (CS), identifying patients who benefit most remains challenging. We developed, validated and assessed the diagnostic performance of a new scoring system to predict high-risk PFO patients who may benefit from PFO transcatheter closure in secondary prevention. The CRISP-PFO score was derived from 936 PFO patients with a RoPE score ≥7, enrolled in a single-center prospective registry from March 2003 to March 2025. The score was derived from 624 patients and validated in 312 patients. Significant baseline variables (p < 0.1) from univariable logistic regression were entered into a multivariable model. Model performance was assessed via receiving operating curves and calibration slope. Patients were stratified into low-, intermediate-, and high-risk groups. In the derivation cohort, multivariate analysis identified four independent predictors of CS: permanent right-to-left shunt (2 points), ischemic brain lesions (2 points), an antero-posterior left atrial (LA) diameter ≥43 mm (1 point), and atrial septal aneurysm (1 point). Based on these factors, patients were stratified into low (0-2 points), intermediate (3-4 points), and high (5-6 points) risk categories, with event rates increasing across these groups (p<0.001). The CRISP-PFO score demonstrated strong diagnostic performance, with an AUC of 0.80-0.81, C-statistics of 0.78-0.79, and Hosmer-Lemeshow p-values of 0.44-0.47 in the derivation and validation cohorts, respectively. The CRISP-PFO score is a novel and straightforward tool that accurately identifies PFO patients at high risk for CS who may benefit from transcatheter PFO closure.