Literature DB >> 32550134

Combination of generative adversarial network and convolutional neural network for automatic subcentimeter pulmonary adenocarcinoma classification.

Yunpeng Wang1, Lingxiao Zhou1,2, Mingming Wang3, Cheng Shao3, Lili Shi1, Shuyi Yang1, Zhiyong Zhang1, Mingxiang Feng4, Fei Shan1, Lei Liu1,5.   

Abstract

BACKGROUND: The efficient and accurate diagnosis of pulmonary adenocarcinoma before surgery is of considerable significance to clinicians. Although computed tomography (CT) examinations are widely used in practice, it is still challenging and time-consuming for radiologists to distinguish between different types of subcentimeter pulmonary nodules. Although there have been many deep learning algorithms proposed, their performance largely depends on vast amounts of data, which is difficult to collect in the medical imaging area. Therefore, we propose an automatic classification system for subcentimeter pulmonary adenocarcinoma, combining a convolutional neural network (CNN) and a generative adversarial network (GAN) to optimize clinical decision-making and to provide small dataset algorithm design ideas.
METHODS: A total of 206 nodules with postoperative pathological labels were analyzed. Among them were 30 adenocarcinomas in situ (AISs), 119 minimally invasive adenocarcinomas (MIAs), and 57 invasive adenocarcinomas (IACs). Our system consisted of two parts, a GAN-based image synthesis, and a CNN classification. First, several popular existing GAN techniques were employed to augment the datasets, and comprehensive experiments were conducted to evaluate the quality of the GAN synthesis. Additionally, our classification system processes were based on two-dimensional (2D) nodule-centered CT patches without the need of manual labeling information.
RESULTS: For GAN-based image synthesis, the visual Turing test showed that even radiologists could not tell the GAN-synthesized from the raw images (accuracy: primary radiologist 56%, senior radiologist 65%). For CNN classification, our progressive growing wGAN improved the performance of CNN most effectively (area under the curve =0.83). The experiments indicated that the proposed GAN augmentation method improved the classification accuracy by 23.5% (from 37.0% to 60.5%) and 7.3% (from 53.2% to 60.5%) in comparison with training methods using raw and common augmented images respectively. The performance of this combined GAN and CNN method (accuracy: 60.5%±2.6%) was comparable to the state-of-the-art methods, and our CNN was also more lightweight.
CONCLUSIONS: The experiments revealed that GAN synthesis techniques could effectively alleviate the problem of insufficient data in medical imaging. The proposed GAN plus CNN framework can be generalized for use in building other computer-aided detection (CADx) algorithms and thus assist in diagnosis. 2020 Quantitative Imaging in Medicine and Surgery. All rights reserved.

Entities:  

Keywords:  Subcentimeter pulmonary adenocarcinoma diagnosis; computed tomography; data augmentation; deep convolutional neural networks; generative adversarial network (GAN)

Year:  2020        PMID: 32550134      PMCID: PMC7276356          DOI: 10.21037/qims-19-982

Source DB:  PubMed          Journal:  Quant Imaging Med Surg        ISSN: 2223-4306


  27 in total

1.  Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge.

Authors:  Arnaud Arindra Adiyoso Setio; Alberto Traverso; Thomas de Bel; Moira S N Berens; Cas van den Bogaard; Piergiorgio Cerello; Hao Chen; Qi Dou; Maria Evelina Fantacci; Bram Geurts; Robbert van der Gugten; Pheng Ann Heng; Bart Jansen; Michael M J de Kaste; Valentin Kotov; Jack Yu-Hung Lin; Jeroen T M C Manders; Alexander Sóñora-Mengana; Juan Carlos García-Naranjo; Evgenia Papavasileiou; Mathias Prokop; Marco Saletta; Cornelia M Schaefer-Prokop; Ernst T Scholten; Luuk Scholten; Miranda M Snoeren; Ernesto Lopez Torres; Jef Vandemeulebroucke; Nicole Walasek; Guido C A Zuidhof; Bram van Ginneken; Colin Jacobs
Journal:  Med Image Anal       Date:  2017-07-13       Impact factor: 8.545

2.  Sharpness-Aware Low-Dose CT Denoising Using Conditional Generative Adversarial Network.

Authors:  Xin Yi; Paul Babyn
Journal:  J Digit Imaging       Date:  2018-10       Impact factor: 4.056

3.  The Cancer Imaging Archive (TCIA): maintaining and operating a public information repository.

Authors:  Kenneth Clark; Bruce Vendt; Kirk Smith; John Freymann; Justin Kirby; Paul Koppel; Stephen Moore; Stanley Phillips; David Maffitt; Michael Pringle; Lawrence Tarbox; Fred Prior
Journal:  J Digit Imaging       Date:  2013-12       Impact factor: 4.056

4.  Computer-Aided Diagnosis of Ground-Glass Opacity Nodules Using Open-Source Software for Quantifying Tumor Heterogeneity.

Authors:  Ming Li; Vivek Narayan; Ritu R Gill; Jyothi P Jagannathan; Maria F Barile; Feng Gao; Raphael Bueno; Jagadeesan Jayender
Journal:  AJR Am J Roentgenol       Date:  2017-10-18       Impact factor: 3.959

5.  Global cancer statistics.

Authors:  Ahmedin Jemal; Freddie Bray; Melissa M Center; Jacques Ferlay; Elizabeth Ward; David Forman
Journal:  CA Cancer J Clin       Date:  2011-02-04       Impact factor: 508.702

6.  Biomarkers in Non-Small Cell Lung Cancer: Perspectives of Individualized Targeted Therapy.

Authors:  Mateus N Aoki; Marla K Amarante; Carlos E C de Oliveira; Maria A E Watanabe
Journal:  Anticancer Agents Med Chem       Date:  2018       Impact factor: 2.505

7.  Persistent pure ground-glass opacity lung nodules ≥ 10 mm in diameter at CT scan: histopathologic comparisons and prognostic implications.

Authors:  Hyun-Ju Lim; Soomin Ahn; Kyung Soo Lee; Joungho Han; Young Mog Shim; Sookyoung Woo; Jae-Hun Kim; Miyeon Yie; Ho Yun Lee; Chin A Yi
Journal:  Chest       Date:  2013-10       Impact factor: 9.410

8.  Quantitative features can predict further growth of persistent pure ground-glass nodule.

Authors:  Zhe Shi; Jiajun Deng; Yunlang She; Lei Zhang; Yijiu Ren; Weiyan Sun; Hang Su; Chenyang Dai; Gening Jiang; Xiwen Sun; Dong Xie; Chang Chen
Journal:  Quant Imaging Med Surg       Date:  2019-02

Review 9.  Development of PD-1/PD-L1 Pathway in Tumor Immune Microenvironment and Treatment for Non-Small Cell Lung Cancer.

Authors:  Jiabei He; Ying Hu; Mingming Hu; Baolan Li
Journal:  Sci Rep       Date:  2015-08-17       Impact factor: 4.379

10.  Automated Classification of Pulmonary Nodules through a Retrospective Analysis of Conventional CT and Two-phase PET Images in Patients Undergoing Biopsy.

Authors:  Atsushi Teramoto; Masakazu Tsujimoto; Takahiro Inoue; Tetsuya Tsukamoto; Kazuyoshi Imaizumi; Hiroshi Toyama; Kuniaki Saito; Hiroshi Fujita
Journal:  Asia Ocean J Nucl Med Biol       Date:  2019
View more
  5 in total

1.  Pix2Pix generative adversarial network for low dose myocardial perfusion SPECT denoising.

Authors:  Jingzhang Sun; Yu Du; ChienYing Li; Tung-Hsin Wu; BangHung Yang; Greta S P Mok
Journal:  Quant Imaging Med Surg       Date:  2022-07

2.  Models of Artificial Intelligence-Assisted Diagnosis of Lung Cancer Pathology Based on Deep Learning Algorithms.

Authors:  Su Chen
Journal:  J Healthc Eng       Date:  2022-03-26       Impact factor: 2.682

3.  Design and Implementation of Obstetric Central Monitoring System Based on Medical Image Segmentation Algorithm.

Authors:  Yuanyuan Zhang; Hua Nie
Journal:  J Healthc Eng       Date:  2022-04-28       Impact factor: 3.822

4.  [Chinese Experts Consensus on Artificial Intelligence Assisted Management for 
Pulmonary Nodule (2022 Version)].

Authors: 
Journal:  Zhongguo Fei Ai Za Zhi       Date:  2022-03-28

5.  Multi-channel multi-task deep learning for predicting EGFR and KRAS mutations of non-small cell lung cancer on CT images.

Authors:  Yunyun Dong; Lina Hou; Wenkai Yang; Jiahao Han; Jiawen Wang; Yan Qiang; Juanjuan Zhao; Jiaxin Hou; Kai Song; Yulan Ma; Ntikurako Guy Fernand Kazihise; Yanfen Cui; Xiaotang Yang
Journal:  Quant Imaging Med Surg       Date:  2021-06
  5 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.