A Comparative Study of Radiomics and Deep-Learning Based Methods for Pulmonary Nodule Malignancy Prediction in Low Dose CT Images.
A Comparative Study of Radiomics and Deep-Learning Based Methods for Pulmonary Nodule Malignancy Prediction in Low Dose CT Images.
复制标题
DOI：
10.3389/fonc.2021.737368
复制
发表时间：
2021
影响因子：
4.7
通讯作者：
Wang C
中科院分区：
文献类型：
--
作者：
Astaraki M;Yang G;Zakko Y;Toma-Dasu I;Smedby Ö;Wang C
放射组学和深度学习方法在各种基于图像的肿瘤学研究中预测病变恶性程度方面表现出巨大的前景。然而，目前还不清楚在获得相同数量的训练数据的情况下，针对特定的临床问题选择哪种方法。在这项研究中，我们尝试比较一系列精心选择的传统放射组学方法，端到端深度学习模型和基于深度特征的放射组学管道在由1297个手动描绘的肺结节组成的开放数据库上预测肺结节恶性肿瘤的性能。常规放射组学分析是通过从目标结节图像中提取标准手工特征来进行的。几个端到端的深度分类器网络，包括VGG，ResNet，DenseNet和EfficientNet也被用来识别肺结节恶性肿瘤。除了基线实现之外，我们还研究了特征选择和类平衡的重要性，以及分离在结节目标区域和背景/上下文区域中学习的特征。通过将放射组学和深度特征汇集在一个混合特征集中，我们研究了这两个集合在恶性肿瘤预测方面的兼容性。通过5倍交叉验证分析，最佳基线常规放射组学模型、深度学习模型和基于深度特征的放射组学模型的AUROC值（平均值±标准差）分别为0.792 ± 0.025、0.801 ± 0.018和0.817 ± 0.032。然而，在尝试了几种优化技术（如特征选择和数据平衡）以及添加上下文特征后，相应的最佳放射组学、端到端深度学习和基于深度特征的模型分别实现了0.921 ± 0.010、0.824 ± 0.021和0.936 ± 0.011的AUROC值。我们从混合特征集获得了最佳的预测准确度（AUROC：0.938 ± 0.010）。端到端的深度学习模型在没有太多微调的情况下，开箱即用就优于传统的放射组学。另一方面，微调模型导致预测性能的显着改善，其中传统的和基于深层特征的放射组学模型取得了可比的结果。在这项比较研究中，混合放射组学方法似乎是最有前途的肺结节恶性预测模型。
Both radiomics and deep learning methods have shown great promise in predicting lesion malignancy in various image-based oncology studies. However, it is still unclear which method to choose for a specific clinical problem given the access to the same amount of training data. In this study, we try to compare the performance of a series of carefully selected conventional radiomics methods, end-to-end deep learning models, and deep-feature based radiomics pipelines for pulmonary nodule malignancy prediction on an open database that consists of 1297 manually delineated lung nodules. Conventional radiomics analysis was conducted by extracting standard handcrafted features from target nodule images. Several end-to-end deep classifier networks, including VGG, ResNet, DenseNet, and EfficientNet were employed to identify lung nodule malignancy as well. In addition to the baseline implementations, we also investigated the importance of feature selection and class balancing, as well as separating the features learned in the nodule target region and the background/context region. By pooling the radiomics and deep features together in a hybrid feature set, we investigated the compatibility of these two sets with respect to malignancy prediction. The best baseline conventional radiomics model, deep learning model, and deep-feature based radiomics model achieved AUROC values (mean ± standard deviations) of 0.792 ± 0.025, 0.801 ± 0.018, and 0.817 ± 0.032, respectively through 5-fold cross-validation analyses. However, after trying out several optimization techniques, such as feature selection and data balancing, as well as adding context features, the corresponding best radiomics, end-to-end deep learning, and deep-feature based models achieved AUROC values of 0.921 ± 0.010, 0.824 ± 0.021, and 0.936 ± 0.011, respectively. We achieved the best prediction accuracy from the hybrid feature set (AUROC: 0.938 ± 0.010). The end-to-end deep-learning model outperforms conventional radiomics out of the box without much fine-tuning. On the other hand, fine-tuning the models lead to significant improvements in the prediction performance where the conventional and deep-feature based radiomics models achieved comparable results. The hybrid radiomics method seems to be the most promising model for lung nodule malignancy prediction in this comparative study.
登录
查看更多内容
影响因子：
4.4
作者：
Han, Fangfang;Wang, Huafeng;Liang, Zhengrong
通讯作者：
Liang, Zhengrong
影响因子：
3.7
作者：
Chen CH;Chang CK;Tu CY;Liao WC;Wu BR;Chou KT;Chiou YR;Yang SN;Zhang G;Huang TC
通讯作者：
Huang TC
影响因子：
3.7
作者：
Metz S;Ganter C;Lorenzen S;van Marwick S;Holzapfel K;Herrmann K;Rummeny EJ;Wester HJ;Schwaiger M;Nekolla SG;Beer AJ
通讯作者：
Beer AJ
DOI：
10.1016/j.ejmp.2019.03.024
发表时间：
2019-04-01
影响因子：
3.4
作者：
Astaraki, Mehdi;Wang, Chunliang;Smedby, Orjan
通讯作者：
Smedby, Orjan
影响因子：
5.5
作者：
Erasmus, JJ;Connolly, JE;Roggli, VL
通讯作者：
Roggli, VL