An LSTM machine learning model predicted future increases in remifentanil flow rate 1 minute in advance with an ROC-AUC of 0.753, sensitivity of 0.659, and specificity of 0.732.
Observational (n=210)
No
Can machine learning models accurately predict anesthesiologists' decisions to increase the flow rate of remifentanil during surgery?
Machine learning models, particularly LSTM, demonstrate potential in predicting anesthesiologists' decisions to increase remifentanil flow rates during surgery.
Recently, research has been conducted to automatically control anesthesia using machine learning, with the aim of alleviating the shortage of anesthesiologists. In this study, we address the problem of predicting decisions made by anesthesiologists during surgery using machine learning; specifically, we formulate a decision making problem by increasing the flow rate at each time point in the continuous administration of analgesic remifentanil as a supervised binary classification problem. The experiments were conducted to evaluate the prediction performance using six machine learning models: logistic regression, support vector machine, random forest, LightGBM, artificial neural network, and long short-term memory (LSTM), using 210 case data collected during actual surgeries. The results demonstrated that when predicting the future increase in flow rate of remifentanil after 1 min, the model using LSTM was able to predict with scores of 0.659 for sensitivity, 0.732 for specificity, and 0.753 for ROC-AUC; this demonstrates the potential to predict the decisions made by anesthesiologists using machine learning. Furthermore, we examined the importance and contribution of the features of each model using Shapley additive explanations-a method for interpreting predictions made by machine learning models. The trends indicated by the results were partially consistent with known clinical findings.
Miyaguchi et al. (Wed,) conducted a observational in Patients undergoing surgery with general anesthesia (n=210). Long short-term memory (LSTM) machine learning model vs. Other machine learning models (Logistic regression, SVM, Random forest, LightGBM, ANN) was evaluated on ROC-AUC for predicting remifentanil flow-increase events 1 minute in the future. An LSTM machine learning model predicted future increases in remifentanil flow rate 1 minute in advance with an ROC-AUC of 0.753, sensitivity of 0.659, and specificity of 0.732.