AUTHOR=Chen Meixiang , Li Pengfei , Huang Yuekang , Li Shuang , Ruan Zheng , Qin Changyu , Huang Jianyu , Wang Ruixin , Lin Zhongqiu , Liu Peng , Xu Lin TITLE=Development and validation of a nomogram for predicting significant coronary artery stenosis in suspected non-ST-segment elevation acute coronary artery syndrome with low-to-intermediate risk stratification JOURNAL=Frontiers in Cardiovascular Medicine VOLUME=9 YEAR=2022 URL=https://www.frontiersin.org/journals/cardiovascular-medicine/articles/10.3389/fcvm.2022.1013563 DOI=10.3389/fcvm.2022.1013563 ISSN=2297-055X ABSTRACT=Background

Patients with non-ST-segment coronary artery syndrome (NSTE-ACS) have significant heterogeneity in their coronary arteries. A better assessment of significant coronary artery stenosis (SCAS) in low-to-intermediate risk NSTE-ACS patients would help identify who might benefit from invasive coronary angiography (ICA). Our study aimed to develop a multivariable-based model for pretesting SCAS in suspected NSTE-ACS with low-to-intermediate risk.

Methods

This prediction nomogram was constructed retrospectively in 469 suspected NSTE-ACS patients with low-to-intermediate risk. Patients were divided into a development group (n = 331, patients admitted to hospital before 1 May 2021) and a temporal validation group (n = 138, patients admitted to hospital since 1 May 2021). The outcome was existing SCAS, including left main artery stenosis ≥50% or any subepicardial coronary artery stenosis ≥70%, all confirmed by invasive coronary angiography. Pretest predictors were selected using Least Absolute Shrinkage and Selection Operator (LASSO) and stepwise logistic regression.

Results

Derivation analyses from the development group (n = 331, admitted before 1 May 2021) generated the 7 strongest predictors out of 25 candidate variables comprising smoker, diabetes, heart rate, cardiac troponin T, N-terminal pro-B-type natriuretic peptide, high-density lipoprotein cholesterol, and left atrial diameter. This nomogram model showed excellent discrimination ability with an area under the receiver operating characteristic curve (AUC) of 0.83 in the development set and 0.79 in the validation dataset. Good calibration was generally displayed, although it slightly overestimated patients’ SCAS risk in the validation group. Decision curve analysis demonstrated the clinical benefit of this model, indicating its value in clinical practice. Furthermore, an optimal cut-off of prediction probability was assigned as 0.61 according to the Youden index.

Conclusion

A prediction nomogram consisting of seven readily available clinical parameters was established to pretest the probability of SCAS in suspected NSTE-ACS patients with low-to-intermediate risk, which may serve as a cost-effective risk stratification tool and thus assist in initial decision making.