Department of Radiology, Peking University Third Hospital, 49 North Garden Road, Haidian District, Beijing, 100191, People's Republic of China.
Huiying Medical Technology Co., Ltd., Dongsheng Science and Technology Park, HaiDian District, Beijing, 100192, People's Republic of China.
Eur Radiol. 2022 Jan;32(1):572-581. doi: 10.1007/s00330-021-08150-y. Epub 2021 Jul 13.
This study aimed to use the most frequent features to establish a vertebral MRI-based radiomics model that could differentiate multiple myeloma (MM) from metastases and compare the model performance with different features number.
We retrospectively analyzed conventional MRI (T1WI and fat-suppression T2WI) of 103 MM patients and 138 patients with metastases. The feature selection process included four steps. The first three steps defined as conventional feature selection (CFS), carried out 50 times (ten times with 5-fold cross-validation), included variance threshold, SelectKBest, and least absolute shrinkage and selection operator. The most frequent fixed features were selected for modeling during the last step. The number of events per independent variable (EPV) is the number of patients in a smaller subgroup divided by the number of radiomics features considered in developing the prediction model. The EPV values considered were 5, 10, 15, and 20. Therefore, we constructed four models using the top 16, 8, 6, and 4 most frequent features, respectively. The models constructed with features selected by CFS were also compared.
The AUCs of 20EPV-Model, 15EPV-Model, and CSF-Model (AUC = 0.71, 0.81, and 0.78) were poor than 10EPV-Model (AUC = 0.84, p < 0.001). The AUC of 10EPV-Model was comparable with 5EPV-Model (AUC = 0.85, p = 0.480).
The radiomics model constructed with an appropriate small number of the most frequent features could well distinguish metastases from MM based on conventional vertebral MRI. Based on our results, we recommend following the 10 EPV as the rule of thumb for feature selection.
• The developed radiomics model could distinguish metastases from multiple myeloma based on conventional vertebral MRI. • An accurate model based on just a handful of the most frequent features could be constructed by utilizing multiple feature reduction techniques. • An event per independent variable value of 10 is recommended as a rule of thumb for modeling feature selection.
本研究旨在利用最常见的特征建立一个基于椎体 MRI 的放射组学模型，以区分多发性骨髓瘤（MM）与转移瘤，并比较不同特征数量下的模型性能。
我们回顾性分析了 103 例 MM 患者和 138 例转移瘤患者的常规 MRI（T1WI 和脂肪抑制 T2WI）。特征选择过程包括四个步骤。前三个步骤定义为常规特征选择（CFS），进行了 50 次（10 次 5 倍交叉验证），包括方差阈值、SelectKBest 和最小绝对值收缩和选择算子。在最后一步中，选择最常见的固定特征进行建模。每个自变量的事件数（EPV）是较小亚组中的患者数除以开发预测模型时考虑的放射组学特征数。考虑的 EPV 值为 5、10、15 和 20。因此，我们分别使用前 16、8、6 和 4 个最常见的特征构建了四个模型。使用 CFS 选择的特征构建的模型也进行了比较。
20EPV-Model、15EPV-Model 和 CSF-Model 的 AUC 值（AUC = 0.71、0.81 和 0.78）均低于 10EPV-Model（AUC = 0.84，p < 0.001）。10EPV-Model 的 AUC 与 5EPV-Model 相当（AUC = 0.85，p = 0.480）。
基于常规椎体 MRI，利用适当数量的最常见特征构建的放射组学模型可以很好地区分转移瘤与 MM。基于我们的结果，我们建议遵循 10EPV 作为特征选择的经验法则。
开发的放射组学模型可以基于常规椎体 MRI 区分转移瘤与多发性骨髓瘤。
通过利用多种特征降维技术，可以构建基于少数最常见特征的准确模型。
建议将每个自变量的事件数（EPV）值设为 10 作为特征选择建模的经验法则。
