Ungan Gülnur, Pons-Escoda Albert, Ulinic Daniel, Arús Carles, Vellido Alfredo, Julià-Sapé Margarida
Centro de Investigación Biomédica en Red (CIBER), 28029 Madrid, Spain.
Departament de Bioquímica i Biologia Molecular and Institut de Biotecnologia i Biomedicina (IBB), Universitat Autònoma de Barcelona (UAB), 08193 Barcelona, Spain.
Cancers (Basel). 2023 Jul 21;15(14):3709. doi: 10.3390/cancers15143709.
In vivo magnetic resonance spectroscopy (MRS) has two modalities, single-voxel (SV) and multivoxel (MV), in which one or more contiguous grids of SVs are acquired.
To test whether MV grids can be classified with models trained with SV.
Retrospective study. Training dataset: Multicenter multiformat SV INTERPRET, 1.5T. Testing dataset: MV eTumour, 3T. Two classification tasks were completed: 3-class (meningioma vs. aggressive vs. normal) and 4-class (meningioma vs. low-grade glioma vs. aggressive vs. normal). Five different methods were tested for feature selection. The classification was implemented using linear discriminant analysis (LDA), random forest, and support vector machines. The evaluation was completed with balanced error rate (BER) and area under the curve (AUC) on both sets. The accuracy in class prediction was calculated by developing a solid tumor index (STI) and segmentation accuracy with the Dice score.
The best method was sequential forward feature selection combined with LDA, with AUCs = 0.95 (meningioma), 0.89 (aggressive), 0.82 (low-grade glioma), and 0.82 (normal). STI was 66% (4-class task) and 71% (3-class task) because two cases failed completely and two more had suboptimal STI as defined by us.
The reasons for failure in the classification of the MV test set were related to the presence of artifacts.
体内磁共振波谱（MRS）有两种模式，即单体素（SV）和多体素（MV），其中获取的是一个或多个相邻的单体素网格。
测试多体素网格是否可以用通过单体素训练的模型进行分类。
回顾性研究。训练数据集：多中心多格式单体素INTERPRET，1.5T。测试数据集：多体素eTumour，3T。完成了两项分类任务：3分类（脑膜瘤vs侵袭性肿瘤vs正常）和4分类（脑膜瘤vs低级别胶质瘤vs侵袭性肿瘤vs正常）。测试了五种不同的特征选择方法。使用线性判别分析（LDA）、随机森林和支持向量机进行分类。通过平衡错误率（BER）和两组的曲线下面积（AUC）完成评估。通过建立实体肿瘤指数（STI）和用Dice分数计算分割准确率来计算类别预测的准确性。
最佳方法是顺序向前特征选择结合LDA，AUC分别为0.95（脑膜瘤）、0.89（侵袭性肿瘤）、0.82（低级别胶质瘤）和0.82（正常）。实体肿瘤指数在4分类任务中为66%，在3分类任务中为71%，因为有两例完全失败，另外两例实体肿瘤指数未达我们定义的最优水平。
多体素测试集分类失败的原因与伪影的存在有关。
