Combining Radiomics and CNNs to Classify Benign and Malignant GIST

DOI

10.2991/ncce-18.2018.44How to use a DOI?

Keywords

GIST, combining features, radiomics, DNN, classification, RF, GBDT.

Abstract

This paper studies the classification of gastrointestinal stromal tumor (GIST) with machine learning and radiomics methods. The highlight of the paper is that combining features extracted from computed tomography (CT) images by radiomics and convolutional neural networks (CNNs) are used in classification task. It’s different from the previous radiomics-only methodology including tumor intensity, tumor shape, tumor texture and wavelet features. In the paper, deep neural networks (DNN) U-Net and Res-Net realize high-level features extraction. The classifiers based on random forest (RF) and gradient boosting decision tree (GBDT) are trained with combining features extracted from computed tomography images. The experiments show that the combining features are effective biomarkers to discriminate benign and malignant GIST. The performance (accuracy 81%, precision 78%, recall 83%) of the classifier is validated on 50 instances.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Tianqi Zhuo
AU  - Xin Li
AU  - Hong Zhou
PY  - 2018/05
DA  - 2018/05
TI  - Combining Radiomics and CNNs to Classify Benign and Malignant GIST
BT  - Proceedings of the 2018 International Conference on Network, Communication, Computer Engineering (NCCE 2018)
PB  - Atlantis Press
SP  - 281
EP  - 287
SN  - 1951-6851
UR  - https://doi.org/10.2991/ncce-18.2018.44
DO  - 10.2991/ncce-18.2018.44
ID  - Zhuo2018/05
ER  -