Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT

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

doi:10.1007/978-3-319-66179-7_75

Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT

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

School of Computer Science

Northwestern Polytechnical University Xian
The University of Sydney
Shanghai Jiao Tong University
Royal Prince Alfred Hospital

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

66 Scopus citations

Abstract

The classification of benign versus malignant lung nodules using chest CT plays a pivotal role in the early detection of lung cancer and this early detection has the best chance of cure. Although deep learning is now the most successful solution for image classification problems, it requires a myriad number of training data, which are not usually readily available for most routine medical imaging applications. In this paper, we propose the transferable multi-model ensemble (TMME) algorithm to separate malignant from benign lung nodules using limited chest CT data. This algorithm transfers the image representation abilities of three ResNet-50 models, which were pre-trained on the ImageNet database, to characterize the overall appearance, heterogeneity of voxel values and heterogeneity of shape of lung nodules, respectively, and jointly utilizes them to classify lung nodules with an adaptive weighting scheme learned during the error back propagation. Experimental results on the benchmark LIDC-IDRI dataset show that our proposed TMME algorithm achieves a lung nodule classification accuracy of 93.40%, which is markedly higher than the accuracy of seven state-of-the-art approaches.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
Editors	Lena Maier-Hein, Alfred Franz, Pierre Jannin, Simon Duchesne, Maxime Descoteaux, D. Louis Collins
Publisher	Springer Verlag
Pages	656-664
Number of pages	9
ISBN (Print)	9783319661780
DOIs	https://doi.org/10.1007/978-3-319-66179-7_75
State	Published - 2017
Event	20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 - Quebec City, Canada Duration: 11 Sep 2017 → 13 Sep 2017

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10435 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017
Country/Territory	Canada
City	Quebec City
Period	11/09/17 → 13/09/17

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Computed tomography (CT)
Deep learning
Ensemble learning
Lung nodule classification

Access to Document

10.1007/978-3-319-66179-7_75

Cite this

Xie, Y., Xia, Y., Zhang, J., Feng, D. D., Fulham, M., & Cai, W. (2017). Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT. In L. Maier-Hein, A. Franz, P. Jannin, S. Duchesne, M. Descoteaux, & D. L. Collins (Eds.), Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings (pp. 656-664). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10435 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-66179-7_75

Xie, Yutong ; Xia, Yong ; Zhang, Jianpeng et al. / Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT. Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. editor / Lena Maier-Hein ; Alfred Franz ; Pierre Jannin ; Simon Duchesne ; Maxime Descoteaux ; D. Louis Collins. Springer Verlag, 2017. pp. 656-664 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{44f51bc43daa44d0814d2e6f6bbfd1f5,

title = "Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT",

abstract = "The classification of benign versus malignant lung nodules using chest CT plays a pivotal role in the early detection of lung cancer and this early detection has the best chance of cure. Although deep learning is now the most successful solution for image classification problems, it requires a myriad number of training data, which are not usually readily available for most routine medical imaging applications. In this paper, we propose the transferable multi-model ensemble (TMME) algorithm to separate malignant from benign lung nodules using limited chest CT data. This algorithm transfers the image representation abilities of three ResNet-50 models, which were pre-trained on the ImageNet database, to characterize the overall appearance, heterogeneity of voxel values and heterogeneity of shape of lung nodules, respectively, and jointly utilizes them to classify lung nodules with an adaptive weighting scheme learned during the error back propagation. Experimental results on the benchmark LIDC-IDRI dataset show that our proposed TMME algorithm achieves a lung nodule classification accuracy of 93.40\%, which is markedly higher than the accuracy of seven state-of-the-art approaches.",

keywords = "Computed tomography (CT), Deep learning, Ensemble learning, Lung nodule classification",

author = "Yutong Xie and Yong Xia and Jianpeng Zhang and Feng, \{David Dagan\} and Michael Fulham and Weidong Cai",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 ; Conference date: 11-09-2017 Through 13-09-2017",

year = "2017",

doi = "10.1007/978-3-319-66179-7\_75",

language = "英语",

isbn = "9783319661780",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "656--664",

editor = "Lena Maier-Hein and Alfred Franz and Pierre Jannin and Simon Duchesne and Maxime Descoteaux and Collins, \{D. Louis\}",

booktitle = "Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings",

}

Xie, Y, Xia, Y, Zhang, J, Feng, DD, Fulham, M & Cai, W 2017, Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT. in L Maier-Hein, A Franz, P Jannin, S Duchesne, M Descoteaux & DL Collins (eds), Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10435 LNCS, Springer Verlag, pp. 656-664, 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017, Quebec City, Canada, 11/09/17. https://doi.org/10.1007/978-3-319-66179-7_75

Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT. / Xie, Yutong; Xia, Yong; Zhang, Jianpeng et al.
Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. ed. / Lena Maier-Hein; Alfred Franz; Pierre Jannin; Simon Duchesne; Maxime Descoteaux; D. Louis Collins. Springer Verlag, 2017. p. 656-664 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10435 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT

AU - Xie, Yutong

AU - Xia, Yong

AU - Zhang, Jianpeng

AU - Feng, David Dagan

AU - Fulham, Michael

AU - Cai, Weidong

N1 - Publisher Copyright: © Springer International Publishing AG 2017.

PY - 2017

Y1 - 2017

N2 - The classification of benign versus malignant lung nodules using chest CT plays a pivotal role in the early detection of lung cancer and this early detection has the best chance of cure. Although deep learning is now the most successful solution for image classification problems, it requires a myriad number of training data, which are not usually readily available for most routine medical imaging applications. In this paper, we propose the transferable multi-model ensemble (TMME) algorithm to separate malignant from benign lung nodules using limited chest CT data. This algorithm transfers the image representation abilities of three ResNet-50 models, which were pre-trained on the ImageNet database, to characterize the overall appearance, heterogeneity of voxel values and heterogeneity of shape of lung nodules, respectively, and jointly utilizes them to classify lung nodules with an adaptive weighting scheme learned during the error back propagation. Experimental results on the benchmark LIDC-IDRI dataset show that our proposed TMME algorithm achieves a lung nodule classification accuracy of 93.40%, which is markedly higher than the accuracy of seven state-of-the-art approaches.

AB - The classification of benign versus malignant lung nodules using chest CT plays a pivotal role in the early detection of lung cancer and this early detection has the best chance of cure. Although deep learning is now the most successful solution for image classification problems, it requires a myriad number of training data, which are not usually readily available for most routine medical imaging applications. In this paper, we propose the transferable multi-model ensemble (TMME) algorithm to separate malignant from benign lung nodules using limited chest CT data. This algorithm transfers the image representation abilities of three ResNet-50 models, which were pre-trained on the ImageNet database, to characterize the overall appearance, heterogeneity of voxel values and heterogeneity of shape of lung nodules, respectively, and jointly utilizes them to classify lung nodules with an adaptive weighting scheme learned during the error back propagation. Experimental results on the benchmark LIDC-IDRI dataset show that our proposed TMME algorithm achieves a lung nodule classification accuracy of 93.40%, which is markedly higher than the accuracy of seven state-of-the-art approaches.

KW - Computed tomography (CT)

KW - Deep learning

KW - Ensemble learning

KW - Lung nodule classification

UR - https://www.scopus.com/pages/publications/85029468681

U2 - 10.1007/978-3-319-66179-7_75

DO - 10.1007/978-3-319-66179-7_75

M3 - 会议稿件

AN - SCOPUS:85029468681

SN - 9783319661780

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 656

EP - 664

BT - Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings

A2 - Maier-Hein, Lena

A2 - Franz, Alfred

A2 - Jannin, Pierre

A2 - Duchesne, Simon

A2 - Descoteaux, Maxime

A2 - Collins, D. Louis

PB - Springer Verlag

T2 - 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017

Y2 - 11 September 2017 through 13 September 2017

ER -

Xie Y, Xia Y, Zhang J, Feng DD, Fulham M, Cai W. Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT. In Maier-Hein L, Franz A, Jannin P, Duchesne S, Descoteaux M, Collins DL, editors, Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Springer Verlag. 2017. p. 656-664. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-66179-7_75

Transferable multi-model ensemble for benign-malignant lung nodule classification on chest CT

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this