An early diagnostic prediction model for acute coronary syndrome incorporating eight clinical variables demonstrated high discrimination, with an AUC of 0.921 in the modeling set.
Cohort (n=480)
Does an 8-variable clinical prediction model accurately diagnose acute coronary syndrome in emergency department patients with chest pain?
A novel 8-variable risk scoring system demonstrated excellent discrimination (AUC >0.90) for the early diagnosis and risk stratification of acute coronary syndrome in patients presenting with chest pain.
Effect estimate: AUC 0.921 (95% CI 0.890-0.952)
Objective To develop and validate an early diagnostic prediction model for acute coronary syndrome (ACS) in patients with chest pain; thus providing scientific evidence for clinical decision-making. Material and methods A retrospective cohort study design was employed, including 480 chest pain patients who presented to the emergency department from January 2020 to January 2025. The patients were randomly divided into a modeling set (336 cases) and a validation set (144 cases) at a 7:3 ratio. Data collected included demographic characteristics, clinical symptoms and signs, medical history, laboratory tests, electrocardiogram, and imaging examinations. Univariate and multivariate logistic regression analyses were used to screen independent predictors and establish the prediction model. Model performance was evaluated through receiver operating characteristic (ROC) curves, Hosmer-Lemeshow test, and Bootstrap resampling, and a simplified risk scoring system was established. Results Multivariate logistic regression analysis showed that elevated cTnI (OR=17.231), ST-segment changes (OR=8.451), typical chest pain (OR=4.047), age ≥60 years (OR=2.441), smoking history (OR=2.103), sweating (OR=1.931), male sex (OR=1.799), and pain duration >30 min (OR=1.689) were independent predictors of ACS (all p<0.05). The area under the curve (AUC) of the model in the modeling set and validation set were 0.921 (95 %CI: 0.890–0.952) and 0.908 (95 %CI: 0.857–0.959), respectively, with sensitivities of 86.4 % and 82.9 %, and specificities of 89.7 % and 87.4 %, respectively. The Hosmer-Lemeshow test indicated good model calibration (modeling set p=0.609, validation set p=0.776). The established risk scoring system (0–20 points) classified patients into four risk stratifications: low risk (0–4 points, ACS incidence 3.8 %), moderate risk (5–8 points, 23.1 %), high risk (9–12 points, 62.2 %), and very high risk (13–20 points, 91.9 %), with an optimal cutoff value of 8 points (Youden index 0.716). Conclusion The ACS early diagnostic prediction model established in this study incorporated eight readily accessible clinical variables and demonstrated good discrimination and calibration. The risk scoring system based on this model is simple and practical. This scoring system can effectively perform risk stratification and provide a valuable clinical tool for early diagnosis and risk assessment of chest pain patients.
He et al. (Wed,) conducted a cohort in Acute coronary syndrome in patients with chest pain (n=480). Early diagnostic prediction model was evaluated on Model discrimination (AUC) for acute coronary syndrome (AUC 0.921, 95% CI 0.890-0.952). An early diagnostic prediction model for acute coronary syndrome incorporating eight clinical variables demonstrated high discrimination, with an AUC of 0.921 in the modeling set.