Construction of a Nomogram Prediction Model for Prognosis in Patients with Large Artery Occlusion-Acute Ischemic Stroke.

Abstract

Patients with large artery occlusion-acute ischemic stroke (LAO-AIS) can experience adverse outcomes, such as brain herniation due to complications. This study aimed to construct a nomogram prediction model for prognosis in patients with LAO-AIS in order to maximize the benefits for clinical patients. Retrospective analysis of 243 patients with LAO-AIS from January 2019 to January 2022 with medical history data and blood examination at admission. Univariate and multivariate analyses were conducted through binary logistic regression equation analysis, and a nomogram prediction model was constructed. Results of this study showed that hyperlipidemia (odds ratio [OR]= 2.849, 95% confidence interval [CI]= 1.100-7.375, P= 0.031), right cerebral infarction (OR=2.144, 95% CI=1.106-4.156, P= 0.024), D-Dimer>500ng/mL (OR=2.891, 95% CI=1.398-5.980, P=0.004), and neutrophil-lymphocyte ratio >7.8 (OR=2.149, 95% CI=1.093-4.225, P= 0.027) were independent risk factors for poor early prognosis in patients with LAO-AIS. In addition, hypertension (OR=1.947, 95% CI=1.114-3.405, P= 0.019), hyperlipidemia (OR=2.594, 95% CI=1.281-5.252, P=0.008), smoking (OR=2.414, 95% CI=1.368-4.261, P=0.002), D-dimer>500ng/mL (OR=3.170, 95% CI=1.533-6.553, P= 0.002), and neutrophil-lymphocyte ratio >7.8 (OR=2.144, 95% CI=1.231-3.735, P= 0.007) were independent risk factors for poor long-term prognosis. The early prognosis nomogram receiver operating characteristic curve area under the curve value was 0.688 for the training set and 0.805 for the validation set, which was highly differentiated. The mean error was 0.025 for the training set calibration curve and 0.016 for the validation set calibration curve. Both the training and validation set decision curve analyses indicated that the clinical benefit of the nomogram was significant. The long-term prognosis nomogram receiver operating characteristic curve area under the curve values was 0.697 for the training set and 0.735 for the validation set, showing high differentiation. The mean error was 0.041 for the training set calibration curve and 0.021 for the validation set calibration curve. Both of the training and validation set decision curve analyses demonstrated a substantial clinical benefit of the nomogram. The nomogram prediction model based on admission history data and blood examination are easy-to-use tools that provide an accurate individualized prediction for patients with LAO-AIS and can assist in early clinical decisions and in obtaining an early prognosis.