AUTHOR=Ding Gang-yu , Xu Jian-hua , He Ji-hong , Nie Zhi-yu TITLE=Clinical scoring model based on age, NIHSS, and stroke-history predicts outcome 3 months after acute ischemic stroke JOURNAL=Frontiers in Neurology VOLUME=13 YEAR=2022 URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2022.935150 DOI=10.3389/fneur.2022.935150 ISSN=1664-2295 ABSTRACT=Background

The clinical nomogram is a popular decision-making tool that can be used to predict patient outcomes, bringing benefits to clinicians and patients in clinical decision-making. This study established a simple and effective clinical prediction model to predict the 3-month prognosis of acute ischemic stroke (AIS), and based on the predicted results, improved clinical decision-making and improved patient outcomes.

Methods

From 18 December 2021 to 8 January 2022, a total of 146 hospitalized patients with AIS confirmed by brain MR were collected, of which 132 eligible participants constituted a prospective study cohort. The least absolute shrinkage and selection operator (LASSO) regression was applied to a nomogram model development dataset to select features associated with poor prognosis in AIS for inclusion in the logistic regression of our risk scoring system. On this basis, the nomogram was drawn, evaluated for discriminative power, calibration, and clinical benefit, and validated internally by bootstrap. Finally, the optimal cutoff point for each independent risk factor and nomogram was calculated using the Youden index.

Results

A total of 132 patients were included in this study, including 85 men and 47 women. Good outcome was found in 94 (71.212%) patients and bad outcome in 38 (28.788%) patients during the follow-up period. A total of eight (6.061%) deaths were reported over this period, of whom five (3.788%) died during hospitalization. Five factors affecting the 3-month prognosis of AIS were screened by LASSO regression, namely, age, hospital stay, previous stroke, atrial fibrillation, and NIHSS. Further multivariate logistic regression revealed three independent risk factors affecting patient outcomes, namely, age, previous stroke, and NIHSS. The area under the curve of the nomogram was 0.880, and the 95% confidence interval was 0.818–0.943, suggesting that the nomogram model has good discriminative power. The p-value for the calibration curve is 0.925, indicating that the nomogram model is well-calibrated. According to the decision curve analysis results, when the threshold probability is >0.01, the net benefit obtained by the nomogram is the largest. The concordance index for 1,000 bootstrapping calculations is 0.869. The age cutoff for predicting poor patient outcomes using the Youden index was 76.5 years (specificity 0.777 and sensitivity 0.684), the cutoff for the NIHSS was 7.5 (specificity 0.936, sensitivity 0.421), and the cutoff for total nomogram score was 68.8 (sensitivity 81.6% and specificity 79.8%).

Conclusion

The nomogram model established in this study had good discrimination, calibration, and clinical benefits. A nomogram composed of age, previous stroke, and NIHSS might predict the prognosis of stroke after AIS. It might intuitively and individually predict the risk of poor prognosis in 3 months of AIS and provide a reference basis for screening the treatment plan of patients.