The Value of the Gensini Score For Prognostic Assessment in Patients with Acute Coronary Syndrome--A Retrospective Cohort Study Based on Machine Learning Methods

BackgroundThe Gensini score (GS) provides a good assessment of the degree of coronary plate loading. However, its clinical significance has been little explored. MethodsIn this retrospective cohort study, we implemented model development and performance comparison on database of The Fourth Affiliated Hospital of Zhejiang University School of Medicine (2019.1-2020.12). The patients were followed up for 2 years. Follow-up endpoint was the occurrence of MACCEs. We extracted clinical baseline data from each ACS patient within 24 hours of hospital admission and randomly divided the datasets into 70% for model training and 30% for model validation. Area under the curve (AUC) was used to compare the prediction performance of XGBoost, SGD and KNN. A decision tree model was constructed to predict the probability of MACCEs using a combination of weight features picked by XGBoost and clinical significance. ResultsA total of 361 ACS patients who met the study criteria were included in this study. It could be observed that the probability of a recurrent MACCEs within 2 years was 25.2%. XGboost had the best predictive efficacy (AUC:0.97). GS has high clinical significance. Then we used GS, Age and CK-MB to construct a decision tree model to predict the probability model of MACCEs reoccurring, and the final AUC value reached 0.771. ConclusionsGS is a powerful indicator for assessing the prognosis of patients with ACS. The cut-off value of GS in the decision tree model provides a reference standard for grading the risk level of patients with ACS.