Prediction of Bailout Mechanical Circulatory Support Use for Hemodynamic Collapse During PCI: The ORACLE Score

Background: Bailout mechanical circulatory support (MCS) may be used in the setting of intraprocedural hemodynamic collapse during percutaneous coronary intervention (PCI). We developed and externally validated a simple risk prediction model for bailout MCS use for hemodynamic collapse during PCI. Methods: Consecutive patients undergoing PCI at a tertiary hospital between 2018 and 2025 were included in the derivation cohort (n=6,931), after excluding cases of salvage PCI and procedures performed with pre-existing or prophylactic MCS. The primary endpoint was bailout MCS use for hemodynamic collapse. A multivariable Firth logistic regression model was developed, internally validated, and translated into a simplified integer-based risk score (ORACLE). External validation was performed in 214,877 PCIs from the Blue Cross Blue Shield of Michigan Cardiovascular Consortium (BMC2) registry. Results: The incidence of bailout MCS was 1.4% in the derivation cohort. Six variables (non-elective presentation, systolic blood pressure, chronic heart failure, prior PCI, rotational/orbital atherectomy, and left main PCI) independently predicted the primary endpoint. The regression model demonstrated good discrimination both in the derivation (area under the curve [AUC] 0.817; 95% confidence interval [CI] 0.774–0.860) and external validation (AUC 0.761; 95% CI 0.753–0.770) cohorts. The simplified ORACLE score preserved most of the discriminatory performance of the full model (AUC 0.791 in the derivation cohort and 0.718 in the external validation cohort) and effectively stratified patients, with observed bailout MCS rates ranging from 0.4% in the low-risk category to 8.7% in the high-risk group, with similar findings in the external validation cohort. Discrimination was consistent across clinically relevant subgroups. Negative predictive value was >99% in both the derivation and external validation cohorts. Conclusions: : ORACLE is a simple score to estimate the risk of bailout MCS use for hemodynamic collapse during PCI, which may support preprocedural risk assessment.