Decruyenaere Alexander, Decruyenaere Philippe, Peeters Patrick, Vermassen Frank, Dhaene Tom, Couckuyt Ivo
Department of Nephrology, Ghent University Hospital, Ghent, Belgium.
Department of Thoracic and Vascular Surgery, Ghent University Hospital, Ghent, Belgium.
BMC Med Inform Decis Mak. 2015 Oct 14;15:83. doi: 10.1186/s12911-015-0206-y.
Predictive models for delayed graft function (DGF) after kidney transplantation are usually developed using logistic regression. We want to evaluate the value of machine learning methods in the prediction of DGF.
497 kidney transplantations from deceased donors at the Ghent University Hospital between 2005 and 2011 are included. A feature elimination procedure is applied to determine the optimal number of features, resulting in 20 selected parameters (24 parameters after conversion to indicator parameters) out of 55 retrospectively collected parameters. Subsequently, 9 distinct types of predictive models are fitted using the reduced data set: logistic regression (LR), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machines (SVMs; using linear, radial basis function and polynomial kernels), decision tree (DT), random forest (RF), and stochastic gradient boosting (SGB). Performance of the models is assessed by computing sensitivity, positive predictive values and area under the receiver operating characteristic curve (AUROC) after 10-fold stratified cross-validation. AUROCs of the models are pairwise compared using Wilcoxon signed-rank test.
The observed incidence of DGF is 12.5 %. DT is not able to discriminate between recipients with and without DGF (AUROC of 52.5 %) and is inferior to the other methods. SGB, RF and polynomial SVM are mainly able to identify recipients without DGF (AUROC of 77.2, 73.9 and 79.8 %, respectively) and only outperform DT. LDA, QDA, radial SVM and LR also have the ability to identify recipients with DGF, resulting in higher discriminative capacity (AUROC of 82.2, 79.6, 83.3 and 81.7 %, respectively), which outperforms DT and RF. Linear SVM has the highest discriminative capacity (AUROC of 84.3 %), outperforming each method, except for radial SVM, polynomial SVM and LDA. However, it is the only method superior to LR.
The discriminative capacities of LDA, linear SVM, radial SVM and LR are the only ones above 80 %. None of the pairwise AUROC comparisons between these models is statistically significant, except linear SVM outperforming LR. Additionally, the sensitivity of linear SVM to identify recipients with DGF is amongst the three highest of all models. Due to both reasons, the authors believe that linear SVM is most appropriate to predict DGF.
肾移植后移植肾功能延迟（DGF）的预测模型通常采用逻辑回归进行构建。我们旨在评估机器学习方法在预测DGF中的价值。
纳入2005年至2011年间根特大学医院497例来自 deceased donors的肾移植病例。应用特征消除程序来确定最佳特征数量，从55个回顾性收集的参数中选出20个参数（转换为指示参数后为24个参数）。随后，使用缩减后的数据集拟合9种不同类型的预测模型：逻辑回归（LR）、线性判别分析（LDA）、二次判别分析（QDA）、支持向量机（SVM；使用线性、径向基函数和多项式核）、决策树（DT）、随机森林（RF）和随机梯度提升（SGB）。在10倍分层交叉验证后，通过计算敏感性、阳性预测值和受试者工作特征曲线下面积（AUROC）来评估模型性能。使用Wilcoxon符号秩检验对模型的AUROC进行两两比较。
观察到的DGF发生率为12.5%。DT无法区分发生和未发生DGF的受者（AUROC为52.5%），且不如其他方法。SGB、RF和多项式SVM主要能够识别未发生DGF的受者（AUROC分别为77.2%、73.9%和79.8%），仅优于DT。LDA、QDA、径向SVM和LR也有能力识别发生DGF的受者，具有更高的判别能力（AUROC分别为82.2%、79.6%、83.3%和81.7%），优于DT和RF。线性SVM具有最高的判别能力（AUROC为84.3%），除径向SVM、多项式SVM和LDA外，优于其他每种方法。然而，它是唯一优于LR的方法。
LDA、线性SVM、径向SVM和LR的判别能力是仅有的高于80%的。这些模型之间的两两AUROC比较均无统计学意义，除了线性SVM优于LR。此外，线性SVM识别发生DGF受者的敏感性在所有模型中位列前三。基于这两个原因，作者认为线性SVM最适合预测DGF。
