Non-contrast Cine Cardiac Magnetic Resonance image radiomics features and machine learning algorithms for myocardial infarction detection
Non-contrast Cine Cardiac Magnetic Resonance image radiomics features and machine learning algorithms for myocardial infarction detection
复制标题
DOI：
10.1016/j.compbiomed.2021.105145
复制
发表时间：
2022-02-01
影响因子：
7.7
通讯作者：
Zaidi, Habib
中科院分区：
文献类型：
--
作者：
Avard, Elham;Shiri, Isaac;Zaidi, Habib
目的：准确区分梗死组织和正常组织对于临床诊断和精准医疗具有重要意义。这项工作的目的是调查的radiomic功能，并开发一个机器学习算法的区分心肌梗死（MI）和存活组织/正常情况下，在左心室心肌的非对比电影心脏磁共振（Cine-CMR）images.Methods：72例患者（52与MI和20名健康对照患者）参加了这项研究。在1.5 T MRI上进行MR成像，参数为：TR = 43.35 ms，TE = 1.22 ms，翻转角= 65 °，时间分辨率为30-40 ms，采用N4偏置场校正算法校正图像的不均匀性。所有图像均由两名心脏成像专家同时进行分割和验证。随后，在舒张末期容积相的整个左心室心肌（3D容积）内进行特征提取。对MR图像重新采样至1 x 1 x 1 mm（3）体素。将MR图像VOI内的所有强度离散化为64个箱。将放射组学特征标准化以获得Z分数，然后进行Student t检验统计分析以进行比较。p值< 0.05用作统计学显著差异的阈值，并进行错误发现率（FDR）校正以报告q值（FDR调整的p值）。对提取的特征采用MSVM-RFE算法进行排序，然后进行特征间的斯皮尔曼相关性分析，剔除相关性较高的特征（R2 > 0.80）。十种不同的机器学习算法用于分类，不同的度量用于评估，各种参数用于模型评估。在单变量分析中，最大2D直径切片（M2 DS）形状特征的受试者工作特征（ROC）值的曲线下面积（AUC）最高（AUC = 0.88，q值= 1.02 E-7），而单变量AUC的平均值为0.62 +/- 0.08。在多变量分析中，Logistic回归（AUC = 0.93 +/- 0.03，准确度= 0.86 +/- 0.05，召回率= 0.87 +/- 0.1，精密度= 0.93 +/- 0.03和F1评分= 0.90 +/- 0.04）和SVM（AUC = 0.92 +/- 0.05，准确度= 0.85 +/- 0.04，召回率= 0.92 +/- 0.01，精确度= 0.88 +/- 0.04和F1评分= 0.90 +/- 0.02）作为用于该放射组学分析的最佳机器学习算法产生了最佳性能。本研究表明，在非造影Cine-CMR图像上使用放射组学分析能够准确检测MI，这可能用作晚期钆增强心脏磁共振（LGE-CMR）的替代诊断方法。
Objective: Robust differentiation between infarcted and normal tissue is important for clinical diagnosis and precision medicine. The aim of this work is to investigate the radiomic features and to develop a machine learning algorithm for the differentiation of myocardial infarction (MI) and viable tissues/normal cases in the left ventricular myocardium on non-contrast Cine Cardiac Magnetic Resonance (Cine-CMR) images.Methods: Seventy-two patients (52 with MI and 20 healthy control patients) were enrolled in this study. MR imaging was performed on a 1.5 T MRI using the following parameters: TR = 43.35 ms, TE = 1.22 ms, flip angle = 65 degrees, temporal resolution of 30-40 ms. N4 bias field correction algorithm was applied to correct the inhomogeneity of images. All images were segmented and verified simultaneously by two cardiac imaging experts in consensus. Subsequently, features extraction was performed within the whole left ventricular myocardium (3D volume) in end-diastolic volume phase. Re-sampling to 1 x 1 x 1 mm(3) voxels was performed for MR images. All intensities within the VOI of MR images were discretized to 64 bins. Radiomic features were normalized to obtain Z-scores, followed by Student's t-test statistical analysis for comparison. A p-value < 0.05 was used as a threshold for statistically significant differences and false discovery rate (FDR) correction performed to report q-value (FDR adjusted p-value). The extracted features were ranked using the MSVM-RFE algorithm, then Spearman correlation between features was performed to eliminate highly correlated features (R2 > 0.80). Ten different machine learning algorithms were used for classification and different metrics used for evaluation and various parameters used for models' evaluation.Results: In univariate analysis, the highest area under the curve (AUC) of receiver operating characteristic (ROC) value was achieved for the Maximum 2D diameter slice (M2DS) shape feature (AUC = 0.88, q-value = 1.02E-7), while the average of univariate AUCs was 0.62 +/- 0.08. In multivariate analysis, Logistic Regression (AUC = 0.93 +/- 0.03, Accuracy = 0.86 +/- 0.05, Recall = 0.87 +/- 0.1, Precision = 0.93 +/- 0.03 and F1 Score = 0.90 +/- 0.04) and SVM (AUC = 0.92 +/- 0.05, Accuracy = 0.85 +/- 0.04, Recall = 0.92 +/- 0.01, Precision = 0.88 +/- 0.04 and F1 Score = 0.90 +/- 0.02) yielded optimal performance as the best machine learning algorithm for this radiomics analysis.Conclusion: This study demonstrated that using radiomics analysis on non-contrast Cine-CMR images enables to accurately detect MI, which could potentially be used as an alternative diagnostic method for Late Gadolinium Enhancement Cardiac Magnetic Resonance (LGE-CMR).