Differentiation between acute and chronic myocardial infarction by means of texture analysis of late gadolinium enhancement and cine cardiac magnetic resonance imaging
Differentiation between acute and chronic myocardial infarction by means of texture analysis of late gadolinium enhancement and cine cardiac magnetic resonance imaging
复制标题
DOI：
10.1016/j.ejrad.2017.04.024
复制
发表时间：
2017-07-01
影响因子：
3.3
通讯作者：
Moratal, David
中科院分区：
文献类型：
--
作者：
Larroza, Andres;Materka, Andrzej;Moratal, David
本研究的目的是利用机器学习技术和从心脏磁共振成像 (MRI) 中提取的纹理特征来区分急性和慢性心肌梗死。研究组包括 22 例急性心肌梗死（AMI）患者和 22 例慢性心肌梗死（CMI）患者。独立分析电影和晚期钆增强 (LGE) MRI，以区分 AMI 和 CMI。从预定义的感兴趣区域 (ROI) 中提取了总共 279 个纹理特征：LGE MRI 上的梗塞区域和电影 MRI 上的整个心肌。分类性能通过嵌套交叉验证方法进行评估，该方法将特征选择技术与三种预测模型相结合：随机森林、具有高斯核的支持向量机 (SVM) 以及具有多项式核的 SVM。多项式 SVM 产生了最好的分类性能。使用 72 个特征的 LGE MRI 上的受试者工作特征曲线提供了 0.86 +/- 0.06 的曲线下面积 (AUC)（平均值 +/- 标准差）； AMI 敏感性 = 0.81 +/- 0.08，特异性 = 0.84 +/- 0.09。在电影 MRI 上，使用 75 个特征，AUC = 0.82 +/- 0.06； AMI 敏感性 = 0.79 +/- 0.10，特异性 = 0.80 +/- 0.10。我们得出的结论是，纹理分析可用于在心脏 LGE MRI 上区分 AMI 和 CMI，也可用于在大多数情况下视觉上难以察觉梗塞的标准电影序列。
The purpose of this study was to differentiate acute from chronic myocardial infarction using machine learning techniques and texture features extracted from cardiac magnetic resonance imaging (MRI). The study group comprised 22 cases with acute myocardial infarction (AMI) and 22 cases with chronic myocardial infarction (CMI). Cine and late gadolinium enhancement (LGE) MRI were analyzed independently to differentiate AMI from CMI. A total of 279 texture features were extracted from predefined regions of interest (ROIs): the infarcted area on LGE MRI, and the entire myocardium on cine MRI. Classification performance was evaluated by a nested cross-validation approach combining a feature selection technique with three predictive models: random forest, support vector machine (SVM) with Gaussian Kernel, and SVM with polynomial kernel. The polynomial SVM yielded the best classification performance. Receiver operating characteristic curves provided area-under-thecurve (AUC) (mean +/- standard deviation) of 0.86 +/- 0.06 on LGE MRI using 72 features; AMI sensitivity = 0.81 +/- 0.08 and specificity = 0.84 +/- 0.09. On cine MRI, AUC = 0.82 +/- 0.06 using 75 features; AMI sensitivity = 0.79 +/- 0.10 and specificity = 0.80 +/- 0.10. We concluded that texture analysis can be used for differentiation of AMI from CMI on cardiac LGE MRI, and also on standard cine sequences in which the infarction is visually imperceptible in most cases.