출혈성 쇼크를 일으킨 흰쥐에서 인공 신경망을 이용한 생존율 예측

Abstract

Purpose: To achieve early diagnosis of hemorrhagic shock using a survival prediction model in rats. Methods: We measured heart rate, mean arterial pressure, respiration rate and temperature in 45 Sprague-Dawley rats, and obtained an artificial neural network model for predicting survival rates. Results: Area under the receiver operating characteristic (ROC) curves was 0.992. Applying the determined optimal boundary value of 0.47, the sensitivity and specificity of survival prediction were 98.4 and 96.6%, respectively. Conclusion: Because this artificial neural network predicts quite accurate survival rates for rats subjected to fixed-volume hemorrhagic shock, and does so with simple measurements of systolic blood pressure (SBP), mean arterial pressure (MAP), heart rate (HR), respiration rate (RR), and temperature (TEMP), it could provide early diagnosis and effective treatment for hemorrhagic shock if this artificial neural network is applicable to humans.