Development and Validation of a Deep Neural Network Model for Prediction of Postoperative In-hospital Mortality.
Development and Validation of a Deep Neural Network Model for Prediction of Postoperative In-hospital Mortality.
复制标题
DOI：
10.1097/aln.0000000000002186
复制
发表时间：
2018-10
期刊：
影响因子：
8.8
通讯作者：
Cannesson M
中科院分区：
文献类型：
--
作者：
Lee CK;Hofer I;Gabel E;Baldi P;Cannesson M
我们测试了基于术中特征训练的深度神经网络可以预测术后住院死亡率的假设。用于训练和验证算法的数据由59,985名患者组成，在手术结束时提取了87个特征。采用带动量的随机梯度下降法训练具有逻辑输出的前馈网络。深度神经网络在80%的数据上进行了训练，剩下20%用于测试。我们通过将美国麻醉师协会的物理状态分类和深度神经网络的鲁棒性添加到简化的特征集来评估深度神经网络的改进。然后将这些网络与身体状况、逻辑回归和其他已公布的临床评分进行比较，包括外科Apgar、术前预测术后死亡率评分、风险量化指数和风险分层指数。训练组和测试组的住院死亡率分别为0.81%和0.73%。简化特征集和物理状态分类的深度神经网络在接收者工作特征曲线下的面积最大，为0.91 （95% CI, 0.88 - 0.93）。曲线下的最高逻辑回归面积发现特征集和ASA状态减少（0.90,95% CI）。0.87 - 0.93)。风险分层指数在受试者工作特征曲线下的面积最大，为0.97 （95% CI, 0.94 ~ 0.99）。深度神经网络可以基于自动提取的术中数据预测住院死亡率，但（尚）不优于现有方法。
We tested the hypothesis that deep neural networks trained on intraoperative features can predict postoperative in-hospital mortality. The data used to train and validate the algorithm consists of 59,985 patients with 87 features extracted at the end of surgery. Feed-forward networks with a logistic output were trained using stochastic gradient descent with momentum. The deep neural networks were trained on 80% of the data, with 20% reserved for testing. We assessed improvement of the deep neural network by adding American Society of Anesthesiologists Physical Status Classification and robustness of the deep neural network to a reduced feature set. The networks were then compared to Physical Status, logistic regression, and other published clinical scores including the Surgical Apgar, PreOperative Score to Predict PostOperative Mortality, Risk Quantification Index, and the Risk Stratification Index. In-hospital mortality in the training and test sets were 0.81% and 0.73%. The deep neural network with a reduced feature set and Physical Status classification had the highest area under the receiver operating characteristics curve, 0.91 (95% CI, 0.88 – 0.93). The highest logistic regression area under the curve was found with a reduced feature set and ASA status (0.90, 95% CI. 0.87 – 0.93). The Risk Stratification Index had the highest area under the receiver operating characteristics curve, at 0.97 (95% CI, 0.94 – 0.99). Deep neural networks can predict in-hospital mortality based on automatically extractable intraoperative data, but are not (yet) superior to existing methods.