Abstract 4138217: A Machine Learning-Based Novel Risk Score Model for Spontaneous Coronary Artery Dissection

Background: Spontaneous coronary artery dissection (SCAD) remains a poorly understood differential diagnosis of acute coronary syndrome (ACS). We aimed to develop a machine learning-based risk score model to predict in-hospital mortality in patients with SCAD. Methods: The National Inpatient Sample (NIS) 2016-2020 database was used to identify all adult patients (≥18 years of age) with SCAD, identified using ICD-10 code I25.42. Relevant clinical characteristics and outcomes were extracted. The primary endpoint was in-hospital all-cause mortality. The dataset was randomly split into 3 subsets: training, validation, and testing, with a ratio of 0.7, 0.2, and 0.1, respectively. The risk score was generated using R software’s Autoscore package, which is a machine learning-based automatic clinical score generator, and performance was evaluated using the area under receiver-operative characteristics curve (AUC) with 95% confidence intervals (95%CI). Results: Among the 8,260 SCAD patients identified [age 60.9±14.4 years, female 4,708 (57%)], there were 480 (5.8%) primary outcome events. Out of the top 20 covariates seen in the Parsimony plot, we used six for our risk score (0–100): age, comorbidity burden, cardiogenic shock, balloon pump, extracorporeal membrane oxygenation, and race (Figure 1A). The AUC for the derivation and validation cohorts was 0.862 (0.817–0.907) and 0.833 (0.788–0.879), respectively (Figures 1B and 1C). Conclusion: We developed and internally validated a machine learning-based risk score model to predict outcomes in patients with SCAD with a high AUC of 0.833 (0.788–0.879). Risk score models guide decision-making, and future studies are needed for external validation of our findings.