Knowledge-guided synthetic medical image adversarial augmentation for ultrasonography thyroid nodule classification. (November 2020)

• Record Type: Journal Article
• Authors: Shi, Guohua; Wang, Jiawen; Qiang, Yan; Yang, Xiaotang; Zhao, Juanjuan; Hao, Rui; Yang, Wenkai; Du, Qianqian; Kazihise, Ntikurako Guy-Fernand
• Is Part Of: Computer methods and programs in biomedicine. Volume 196(2020)
• Publication Date: 2020-11
• Subjects: Domain knowledge -- Generative adversarial network -- Image synthesis -- Data augmentation -- Thyroid nodule -- Classification; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2020.105611
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: This work is the first to integrate domain knowledge and deep learning into synthetic medical image augmentation. This work is the first to extract domain knowledge from standardized terminology rather than images. Design a novel deep learning model for data augmentation task, and the results are comparable to the state-of-the-art methods. Integrate domain knowledge into deep learning model to improve the classification performance of ultrasonography thyroid nodules. Abstract: Background and objective: Image classification is an important task in many medical applications. Methods based on deep learning have made great achievements in the computer vision domain. However, they typically rely on large-scale datasets which are annotated. How to obtain such great datasets is still a serious problem in medical domain. Methods: In this paper, we propose a knowledge-guided adversarial augmentation method for synthesizing medical images. First, we design Term and Image Encoders to extract domain knowledge from radiologists, then we use domain knowledge as novel condition to constrain the Auxiliary Classifier Generative Adversarial Network (ACGAN) framework for the synthesis of high-quality thyroid nodule images. Finally, we demonstrate our method on the task of classifying ultrasonography thyroid nodule. Our method can make effective use of the high-quality diagnostic experience of advanced radiologists. In addition, we creatively choose to extract domain knowledge from standardized terms rather than ultrasound images. Results: Our novel method is demonstrated on a limited dataset of 1937 clinical thyroid ultrasound images and corresponding standardized terms. The accuracy of the proposed model for thyroid nodules is 91.46%, the sensitivity is 90.63%, the specificity is 92.65%, and the AUC is 95.32%, which is better than the current classification methods for thyroid nodules. The experimental results show the model has better generalization and robustness. Conclusions: We believe that the proposed method can alleviate the problem of insufficient data in the medical domain, and other medical problems can benefit from using synthetic augmentation.