An Appraisal of Nodule Diagnosis for Lung Cancer in CT Images

Zhang, Guobin; Yang, Zhiyong; Gong, Li; Jiang, Shan; Wang, Lu; Cao, Xi; Lin, Wei; Zhang, Hongyun; Liu, Ziqi

doi:10.1007/s10916-019-1327-0

Cited by 26 publications

(13 citation statements)

References 112 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Statistical/classical machine learning algorithms have been extensively used for lung classification [5]- [8] and nodule diagnosis from computed tomography (CT) images [9]. For instance, in [10], three statistical features were calculated from lung texture to discriminate between malignant and benign lung nodules using support vector machines (SVM) classifier.…”

Section: Introductionmentioning

confidence: 99%

“…Deep learning algorithms have also demonstrated their great success in different domains including bioinformatics [17]- [19], plant identification [20], medical image analysis [3], [21], [22] and wind power prediction [23]. With the availability of enough annotated images, deep learning approaches [2], [24]- [30] have demonstrated their superiority over the statistical machine learning approaches [9]. In [31], two deep learning approaches, deep belief network (DBN) and restricted Boltzmann machine (RBM) were used to classify the lung nodules from CT into malignant or benign based on three feature extraction methods (GLCM, histogram features, and wavelet transformation).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DeTrac: Transfer Learning of Class Decomposed Medical Images in Convolutional Neural Networks

2020

View full text Add to dashboard Cite

Due to the high availability of large-scale annotated image datasets, paramount progress has been made in deep convolutional neural networks (CNNs) for image classification tasks. CNNs enable learning highly representative and hierarchical local image features directly from data. However, the availability of annotated data, especially in the medical imaging domain, remains the biggest challenge in the field. Transfer learning can provide a promising and effective solution by transferring knowledge from generic image recognition tasks to the medical image classification. However, due to irregularities in the dataset distribution, transfer learning usually fails to provide a robust solution. Class decomposition facilitates easier to learn class boundaries of a dataset, and consequently can deal with any irregularities in the data distribution. Motivated by this challenging problem, the paper presents Decompose, Transfer, and Compose (DeTraC) approach, a novel CNN architecture based on class decomposition to improve the performance of medical image classification using transfer learning and class decomposition approach. DeTraC enables learning at the subclass level that can be more separable with a prospect to faster convergence. We validated our proposed approach with three different cohorts of chest X-ray images, histological images of human colorectal cancer, and digital mammograms. We compared DeTraC with the state-of-the-art CNN models to demonstrate its high performance in terms of accuracy, sensitivity, and specificity.INDEX TERMS Convolution neural networks, class decomposition, data irregularity, medical image classification, transfer learning.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

DeTrac: Transfer Learning of Class Decomposed Medical Images in Convolutional Neural Networks

2020

View full text Add to dashboard Cite

show abstract

“…There are no characteristic clinical symptoms in the early stages of lung cancer, delaying its diagnosis and treatment ( Jacobsen et al., 2017 ; Hong et al., 2015 ). We rely on low-dose chest CT scan more than clinical symptoms, chest X-rays, or tumor markers for the early diagnosis of lung cancer ( Cainap et al., 2020 ; Zhang et al., 2019a ). Therefore, the application of low-dose chest CT scans in the screening of early lung disease is becoming more common.…”

Section: Introductionmentioning

confidence: 99%

Signature identification of relapse-related overall survival of early lung adenocarcinoma after radical surgery

Han

Yue

Kong

et al. 2021

PeerJ

View full text Add to dashboard Cite

Background The widespread use of low-dose chest CT screening has improved the detection of early lung adenocarcinoma. Radical surgery is the best treatment strategy for patients with early lung adenocarcinoma; however, some patients present with postoperative recurrence and poor prognosis. Through this study, we hope to establish a model that can identify patients that are prone to recurrence and have poor prognosis after surgery for early lung adenocarcinoma. Materials and Methods We screened prognostic and relapse-related genes using The Cancer Genome Atlas (TCGA) database and the GSE50081 dataset from the Gene Expression Omnibus (GEO) database. The GSE30219 dataset was used to further screen target genes and construct a risk prognosis signature. Time-dependent ROC analysis, calibration degree analysis, and DCA were used to evaluate the reliability of the model. We validated the TCGA dataset, GSE50081, and GSE30219 internally. External validation was conducted in the GSE31210 dataset. Results A novel four-gene signature (INPP5B, FOSL2, CDCA3, RASAL2) was established to predict relapse-related survival outcomes in patients with early lung adenocarcinoma after surgery. The discovery of these genes may reveal the molecular mechanism of recurrence and poor prognosis of early lung adenocarcinoma. In addition, ROC analysis, calibration analysis and DCA were used to verify the genetic signature internally and externally. Our results showed that our gene signature had a good predictive ability for recurrence and prognosis. Conclusions We established a four-gene signature and predictive model to predict the recurrence and corresponding survival rates in patients with early lung adenocarcinoma after surgery. These may be helpful for reforumulating post-operative consolidation treatment strategies.

show abstract

“…Imaging analysis is still the mainstream detection method [ 3 , 4 ]. It is also a main method to predict the development trend of benign and malignant pulmonary nodules from the perspective of imaging [ 5 , 6 ]. “Biopsy” needs to sample the suspected lung lesions for detection.…”

Section: Introductionmentioning

confidence: 99%

Spiculation Sign Recognition in a Pulmonary Nodule Based on Spiking Neural P Systems

Qiu

Sun

Zhou

et al. 2020

BioMed Research International

View full text Add to dashboard Cite

The spiculation sign is one of the main signs to distinguish benign and malignant pulmonary nodules. In order to effectively extract the image feature of a pulmonary nodule for the spiculation sign distinguishment, a new spiculation sign recognition model is proposed based on the doctors’ diagnosis process of pulmonary nodules. A maximum density projection model is established to fuse the local three-dimensional information into the two-dimensional image. The complete boundary of a pulmonary nodule is extracted by the improved Snake model, which can take full advantage of the parallel calculation of the Spike Neural P Systems to build a new neural network structure. In this paper, our experiments show that the proposed algorithm can accurately extract the boundary of a pulmonary nodule and effectively improve the recognition rate of the spiculation sign.

show abstract

An Appraisal of Nodule Diagnosis for Lung Cancer in CT Images

Cited by 26 publications

References 112 publications

DeTrac: Transfer Learning of Class Decomposed Medical Images in Convolutional Neural Networks

DeTrac: Transfer Learning of Class Decomposed Medical Images in Convolutional Neural Networks

Signature identification of relapse-related overall survival of early lung adenocarcinoma after radical surgery

Spiculation Sign Recognition in a Pulmonary Nodule Based on Spiking Neural P Systems

Contact Info

Product

Resources

About