MRI 放射组学的 Combat 均衡化：机器学习对非二进制组织分类的影响。
ComBat Harmonization for MRI Radiomics: Impact on Nonbinary Tissue Classification by Machine Learning.
机构信息
From the Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY.
Department of Biomedical Imaging and Image-Guided Therapy, Medical University of Vienna, Vienna, Austria.
出版信息
Invest Radiol. 2023 Sep 1;58(9):697-701. doi: 10.1097/RLI.0000000000000970.
OBJECTIVES
The aims of this study were to determine whether ComBat harmonization improves multiclass radiomics-based tissue classification in technically heterogeneous MRI data sets and to compare the performances of 2 ComBat variants.
MATERIALS AND METHODS
One hundred patients who had undergone T1-weighted 3D gradient echo Dixon MRI (2 scanners/vendors; 50 patients each) were retrospectively included. Volumes of interest (2.5 cm 3 ) were placed in 3 disease-free tissues with visually similar appearance on T1 Dixon water images: liver, spleen, and paraspinal muscle. Gray-level histogram (GLH), gray-level co-occurrence matrix (GLCM), gray-level run-length matrix (GLRLM), and gray-level size-zone matrix (GLSZM) radiomic features were extracted. Tissue classification was performed on pooled data from the 2 centers (1) without harmonization, (2) after ComBat harmonization with empirical Bayes estimation (ComBat-B), and (3) after ComBat harmonization without empirical Bayes estimation (ComBat-NB). Linear discriminant analysis with leave-one-out cross-validation was used to distinguish among the 3 tissue types, using all available radiomic features as input. In addition, a multilayer perceptron neural network with a random 70%:30% split into training and test data sets was used for the same task, but separately for each radiomic feature category.
RESULTS
Linear discriminant analysis-based mean tissue classification accuracies were 52.3% for unharmonized, 66.3% for ComBat-B harmonized, and 92.7% for ComBat-NB harmonized data. For multilayer perceptron neural network, mean classification accuracies for unharmonized, ComBat-B-harmonized, and ComBat-NB-harmonized test data were as follows: 46.8%, 55.1%, and 57.5% for GLH; 42.0%, 65.3%, and 71.0% for GLCM; 45.3%, 78.3%, and 78.0% for GLRLM; and 48.1%, 81.1%, and 89.4% for GLSZM. Accuracies were significantly higher for both ComBat-B- and ComBat-NB-harmonized data than for unharmonized data for all feature categories (at P = 0.005, respectively). For GLCM ( P = 0.001) and GLSZM ( P = 0.005), ComBat-NB harmonization provided slightly higher accuracies than ComBat-B harmonization.
CONCLUSIONS
ComBat harmonization may be useful for multicenter MRI radiomics studies with nonbinary classification tasks. The degree of improvement by ComBat may vary among radiomic feature categories, among classifiers, and among ComBat variants.
目的
本研究旨在确定 ComBat 均衡是否能提高技术异质 MRI 数据集的多类放射组学组织分类，并比较 2 种 ComBat 变体的性能。
材料和方法
回顾性纳入了 100 名接受 T1 加权 3D 梯度回波 Dixon MRI（2 台扫描仪/供应商；每台 50 名患者）的患者。在 T1 Dixon 水图像上具有视觉相似外观的 3 种无疾病组织中放置感兴趣体积（2.5 cm 3 ）：肝、脾和脊柱旁肌肉。提取灰度直方图（GLH）、灰度共生矩阵（GLCM）、灰度游程长度矩阵（GLRLM）和灰度大小区域矩阵（GLSZM）放射组学特征。在（1）未均衡、（2）采用经验贝叶斯估计的 ComBat 均衡（ComBat-B）和（3）无经验贝叶斯估计的 ComBat 均衡（ComBat-NB）的情况下，对来自 2 个中心的合并数据进行组织分类。使用线性判别分析进行分类，使用所有可用的放射组学特征作为输入。此外，还使用多层感知器神经网络（随机 70%：30% 分为训练和测试数据集）进行相同的任务，但分别针对每个放射组学特征类别。
结果
基于线性判别分析的平均组织分类准确率为：未均衡组为 52.3%，ComBat-B 均衡组为 66.3%，ComBat-NB 均衡组为 92.7%。对于多层感知器神经网络，未均衡、ComBat-B 均衡和 ComBat-NB 均衡测试数据的平均分类准确率如下：GLH 为 46.8%、55.1%和 57.5%；GLCM 为 42.0%、65.3%和 71.0%；GLRLM 为 45.3%、78.3%和 78.0%；GLSZM 为 48.1%、81.1%和 89.4%。对于所有特征类别，ComBat-B 和 ComBat-NB 均衡数据的准确率均明显高于未均衡数据（均为 P=0.005）。对于 GLCM（P=0.001）和 GLSZM（P=0.005），ComBat-NB 均衡的准确率略高于 ComBat-B 均衡。
结论
ComBat 均衡可能有助于具有非二进制分类任务的多中心 MRI 放射组学研究。ComBat 改善的程度可能因放射组学特征类别、分类器和 ComBat 变体而异。
Zotero 插件