The CASPER-Score, integrating AI-based ECG analysis with clinical predictors, demonstrated good discrimination (AUC = 0.86) for identifying relevant coronary artery stenosis in OHCA patients.
Cohort (n=204)
No
Does the CASPER-Score accurately identify relevant coronary artery stenosis in patients after out-of-hospital cardiac arrest?
The novel CASPER-Score, which integrates AI-based ECG analysis with clinical predictors, provides good discrimination for identifying relevant coronary artery stenosis in OHCA patients, potentially aiding early angiography decisions.
Effect estimate: AUC 0.86
Introduction Out-of-hospital cardiac arrest (OHCA) is a leading cause of mortality worldwide, frequently associated with acute coronary syndromes. While immediate coronary angiography is recommended in patients with ST-segment elevation myocardial infarction, it remains controversial in other cases. Further on diagnosing ST- segment elevation can be challenging. Artificial intelligence (AI) has shown promise in electrocardiogram (ECG) interpretation, but its value for predicting relevant coronary artery lesions and the need for intervention after OHCA is unknown. Methods We conducted a single-center, retrospective, cohort study of OHCA patients admitted to the intensive care unit between 01/2019 and 12/2021 who underwent coronary angiography within seven days. Admission ECGs were analyzed using an AI-powered interpretation tool (PMcardio®). Clinical, echocardiographic, angiographic, laboratory, and mortality data were collected. The primary outcome was relevant coronary artery stenosis, defined as ≥50% diameter stenosis of the left main coronary artery or ≥70% of all other vessels. The secondary outcome was coronary revascularization. Multivariable logistic regression analysis was performed. Results Among 204 OHCA patients undergoing coronary angiography within 7 days, relevant coronary artery stenosis was present in 157 patients (77%), and 127 patients (62.3%) underwent coronary revascularization. Age, male sex, initial shockable rhythm, and AI-based ECG-derived occlusion myocardial infarction (OMI) detection were independently associated with relevant coronary artery lesions. These variables formed the CASPER-Score (range 0-8 points), demonstrating good discrimination (AUC = 0.86). The model predicting revascularization demonstrated moderate discrimination (AUC= 0.74). Conclusion The CASPER-Score may improve early risk stratification and support decision-making regarding coronary angiography in OHCA patients.
̈beck et al. (Wed,) conducted a cohort in Out-of-hospital cardiac arrest (OHCA) (n=204). CASPER-Score was evaluated on Relevant coronary artery stenosis, defined as ≥50% diameter stenosis of the left main coronary artery or ≥70% of all other vessels (AUC 0.86). The CASPER-Score, integrating AI-based ECG analysis with clinical predictors, demonstrated good discrimination (AUC = 0.86) for identifying relevant coronary artery stenosis in OHCA patients.