基于机器学习的腺苷负荷心肌灌注SPECT心脏死亡预测
研究基于8321例接受腺苷负荷心肌灌注SPECT患者的122个候选临床预测特征,训练多种机器学习模型预测心脏死亡(551例),并与基线逻辑回归比较。所有机器学习方法均优于基线逻辑回归(AUC=0.76;14个特征),LASSO模型所需特征最少(AUC=0.77;P=.045;6个特征),支持向量机模型准确度最高(AUC=0.83;P<.0001;49个特征)。作者认为机器学习在该应用中可能比逻辑回归更有效。
要点
- 支持向量机模型预测心脏死亡的AUC最高(0.83;P<.0001;49个特征)。
- LASSO模型仅需6个特征,AUC=0.77(P=.045),优于基线逻辑回归。
- 所有机器学习方法均优于基线逻辑回归(AUC=0.76;14个特征)。
核心结论
- LASSO模型在准确度和简洁性(特征数量)方面均优于逻辑回归,支持向量机模型在预测接受心肌灌注SPECT患者心脏死亡方面AUC最高。
- 结合展示风险评分背后的推理,结果提示机器学习在该应用中可能比逻辑回归更有效。
研究设计与样本
- 使用8321例患者的122个候选临床预测特征训练多种模型,其中包括551例后续心脏死亡病例。
- 准确度通过受试者工作特征曲线下面积(AUC)在交叉验证框架内计算。
- 开发了一种展示模型推理依据的方法以促进临床解读。
主要发现
- 基线逻辑回归AUC=0.76(14个特征),被所有其他方法超越。
- LASSO模型AUC=0.77(P=.045;6个特征),所需特征最少。
- 支持向量机模型AUC=0.83(P<.0001;49个特征),准确度最高。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: We developed machine-learning (ML) models to estimate a patient's risk of cardiac death based on adenosine myocardial perfusion SPECT (MPS) and associated clinical data, and compared their performance to baseline logistic regression (LR). We demonstrated an approach to visually convey the reasoning behind a patient's risk to provide insight to clinicians beyond that of a "black box." METHODS: We trained multiple models using 122 potential clinical predictors (features) for 8321 patients, including 551 cases of subsequent cardiac death. Accuracy was measured by area under the ROC curve (AUC), computed within a cross-validation framework. We developed a method to display the model's rationale to facilitate clinical interpretation. RESULTS: The baseline LR (AUC = 0.76; 14 features) was outperformed by all other methods. A least absolute shrinkage and selection operator (LASSO) model (AUC = 0.77; p = .045; 6 features) required the fewest features. A support vector machine (SVM) model (AUC = 0.83; p < .0001; 49 features) provided the highest accuracy. CONCLUSIONS: LASSO outperformed LR in both accuracy and simplicity (number of features), with SVM yielding best AUC for prediction of cardiac death in patients undergoing MPS. Combined with presenting the reasoning behind the risk scores, our results suggest that ML can be more effective than LR for this application.
引用
- Haro Alonso D, Wernick MN, Yang Y, et al. Prediction of cardiac death after adenosine myocardial perfusion SPECT based on machine learning. J Nucl Cardiol. 2018. PMID: 29542015. doi:10.1007/s12350-018-1250-7
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。