Knowledge-based Collaborative Deep Learning for Benign-Malignant Lung Nodule Classification on Chest CT

Xie, Yutong; Xia, Yong; Zhang, Jianpeng; Song, Yang; Feng, David Dagan; Fulham, Michael; Cai, Weidong

doi:10.1109/tmi.2018.2876510

Cited by 393 publications

(184 citation statements)

References 58 publications

(82 reference statements)

Supporting

Mentioning

182

Contrasting

Unclassified

Order By: Relevance

“…They finetunned all the layers of their four CNN architectures, which were pre-trained over the natural images [28]. The multiview knowledge-based collaborative (MV-KBC) deep model is proposed by Xie et al [29] for malignancy classification. They achieved an accuracy score of 91.60% on the LIDC-IDRI database.…”

Section: Small Database Issues In Lung Nodule Classificationmentioning

confidence: 99%

“…Author et al: Preparation of Papers for IEEE TRANSACTIONS and JOURNALS compared our texture CNN with deep learning-based models to prove the effectiveness of the model. Table 8 shows the performance comparison of the proposed model with various existing state-of-the-art deep learning-based models, like Fuse-TSD algorithm [25], deep fully convolutional neural network (DFCNet) [45], MV-KBC learning model [29], MK-SSAC model [30] and GD network [23] etc. The results presented in Table 8 show that the proposed model performs better as compared to all other deep learning techniques except LdcNet-FL which has a bit higher accuracy and specificity score.…”

Section: B Performance Evaluation With Lidc-idri Databasementioning

confidence: 99%

See 1 more Smart Citation

Efficient Lung Nodule Classification Using Transferable Texture Convolutional Neural Network

et al. 2020

View full text Add to dashboard Cite

Lung nodules are vital indicators for the presence of lung cancer. An early detection enhances the survival rate of the patient by starting treatment at the right time. The detection and classification of malignancy in Computed Tomography (CT) images is a very time-consuming and difficult task for radiologists which lead the researchers to develop algorithms for Computer-Aided Diagnosis (CAD) systems to mitigate this burden. The performance of CAD systems is continuously improving by using various deep learning techniques for screening of lung cancer. In this paper, we proposed transferable texture Convolutional Neural Networks (CNN) to improve the classification performance of pulmonary nodules in CT scans. An Energy Layer (EL) is incorporated in our scheme, which extracts texture features from the convolutional layer. The inclusion of EL reduces the number of learnable parameters of the network, which further reduces the memory requirements and computational complexity. The proposed model has only three convolutional layers and one EL, instead of pooling layer. Overall proposed CNN architecture comprises of nine layers for automatic feature extraction and classification of pulmonary nodule candidates as malignant or benign. Furthermore, the pre-trained model of proposed CNN is also used to handle the smaller dataset classification problem by using transfer learning. This work has been evaluated on publicly available LIDC-IDRI and the LUNGx Challenge database through different evaluation matrices, such as; the accuracy, specificity, error rate and AUC. The proposed model is trained by six-fold cross-validation and achieved an accuracy score of 96.69% ± 0.72% with only 3.30% ± 0.72% error rate. Whereas, the measured AUC and recall is 99.11% ± 0.45% and 97.19% ± 0.57%, respectively. Moreover, we also tested our proposed technique on the MNIST dataset and achieved state-of-the-art results in terms of accuracy and error rate.

show abstract

Section: Small Database Issues In Lung Nodule Classificationmentioning

confidence: 99%

Section: B Performance Evaluation With Lidc-idri Databasementioning

confidence: 99%

Efficient Lung Nodule Classification Using Transferable Texture Convolutional Neural Network

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Xie et al, [24] proposed to use minimum Chest CT data, a multi-visual, collaborative, and Knowledge-based (MV-KBC) model to differentiate malignant from benign nodules. This design develops the characteristics of the 3-D pulmonary nodule through the decomposition into 9 fixed views.…”

Section: Survey About Previous Workmentioning

confidence: 99%

Design of Automatic Lung Nodule Detection System Based on Multi-Scene Deep Learning Framework

Zhang

Kong

2020

IEEE Access

View full text Add to dashboard Cite

Nowadays, the efficient identification of the lung nodule greatly leads to the chance of lung cancer risk assessment. Hence, the exact locations of lung nodules are a critical and complicated task. Researchers in this area have been working widely for almost two years. However, previous computer-aided detection (CAD) modules, such as transforming CT, segmenting the lung nodule and extracting the features are mostly complex and time-consuming, because more modules will require the creation of a complete image processing system. In addition, certain state-of-the-art deep learning systems are specified in the database standard. For this purpose, this paper suggests an efficient identification system for lung nodules based on Multi-Scene Deep Learning Framework (MSDLF) by the vesselness filter. A four-channel CNN model is designed to enhance the radiologist's knowledge in the detection of four-stage nodules by integrating two image Scenes. This model can be applied in two different classes. The results show that the Multi-Scene Deep Learning Framework (MSDLF) is efficient for increasing the accuracy and significantly reducing false positives in an enormous amount of image data in the detection of lung nodules. INDEX TERMS Multi-scene deep learning framework, vesselness filter, convolutional neural network (CNN), lung nodules. QINGHAI ZHANG was born in 1965. He received the degree. He is currently a Deputy Chief Physician. In 1986, he joined in imaging diagnosis, for more than 30 years. He successively went to the Jinan Military Region General Hospital, Shandong Provincial Institute of Imaging, and 301 Hospital to study CT / MR diagnostics and CT interventional therapy. In practical work, he is also good at learning and summarizing, is diligent in studying and thinking, dares to challenge traditional theories with innovative thinking, and has accumulated rich clinical and practical experience. On the basis of fully mastering common and frequently-occurring imaging diagnosis, in obstetric diseases, we have our own unique insights. He has authored more than ten national academic articles, including four national core journals and two scientific research achievements. XIAOJING KONG was born in 1979. He received the degree from college. He has participated in 2003, involved in imaging diagnosis, for more than 17 years, and successively studied at the

show abstract

“…Therefore, many attempts have been made to develop computer-aided diagnosis (CAD) systems for automatic discrimination. [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] Conventional CAD systems first use classical image processing techniques, such as morphologic operators, 3-5 region growing, 6 energy optimization, 7,8 and statistical learning, 9,10 to segment a region of interest (ROI) that includes the nodule. Then, handcrafted features are extracted from the ROI, which are then fed to a classifier for nodule classification.…”

Section: Introductionmentioning

confidence: 99%

Feature‐shared adaptive‐boost deep learning for invasiveness classification of pulmonary subsolid nodules in CT images

Wang

Chen

et al. 2020

Medical Physics

View full text Add to dashboard Cite

Purpose: In clinical practice, invasiveness is an important reference indicator for differentiating the malignant degree of subsolid pulmonary nodules. These nodules can be classified as atypical adenomatous hyperplasia (AAH), adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), or invasive adenocarcinoma (IAC). The automatic determination of a nodule's invasiveness based on chest CT scans can guide treatment planning. However, it is challenging, owing to the insufficiency of training data and their interclass similarity and intraclass variation. To address these challenges, we propose a two-stage deep learning strategy for this task: prior-feature learning followed by adaptive-boost deep learning. Methods: The adaptive-boost deep learning is proposed to train a strong classifier for invasiveness classification of subsolid nodules in chest CT images, using multiple 3D convolutional neural network (CNN)-based weak classifiers. Because ensembles of multiple deep 3D CNN models have a huge number of parameters and require large computing resources along with more training and testing time, the prior-feature learning is proposed to reduce the computations by sharing the CNN layers between all weak classifiers. Using this strategy, all weak classifiers can be integrated into a single network. Results: Tenfold cross validation of binary classification was conducted on a total of 1357 nodules, including 765 noninvasive (AAH and AIS) and 592 invasive nodules (MIA and IAC). Ablation experimental results indicated that the proposed binary classifier achieved an accuracy of 73:4% AE 1:4 with an AUC of 81.3% AE 2:2. These results are superior compared to those achieved by three experienced chest imaging specialists who achieved an accuracy of 69:1% , 69:3% , and 67:9% , respectively. About 200 additional nodules were also collected. These nodules covered 50 cases for each category (AAH, AIS, MIA, and IAC, respectively). Both binary and multiple classifications were performed on these data and the results demonstrated that the proposed method definitely achieves better performance than the performance achieved by nonensemble deep learning methods. Conclusions: It can be concluded that the proposed adaptive-boost deep learning can significantly improve the performance of invasiveness classification of pulmonary subsolid nodules in CT images, while the prior-feature learning can significantly reduce the total size of deep models. The promising results on clinical data show that the trained models can be used as an effective lung cancer screening tool in hospitals. Moreover, the proposed strategy can be easily extended to other similar classification tasks in 3D medical images.

show abstract

Knowledge-based Collaborative Deep Learning for Benign-Malignant Lung Nodule Classification on Chest CT

Cited by 393 publications

References 58 publications

Efficient Lung Nodule Classification Using Transferable Texture Convolutional Neural Network

Efficient Lung Nodule Classification Using Transferable Texture Convolutional Neural Network

Design of Automatic Lung Nodule Detection System Based on Multi-Scene Deep Learning Framework

Feature‐shared adaptive‐boost deep learning for invasiveness classification of pulmonary subsolid nodules in CT images

Contact Info

Product

Resources

About