DiSegNet: A deep dilated convolutional encoder-decoder architecture for lymph node segmentation on PET/CT images. (March 2021)

• Record Type: Journal Article
• Authors: Xu, Guoping; Cao, Hanqiang; Udupa, Jayaram K.; Tong, Yubing; Torigian, Drew A.
• Is Part Of: Computerized medical imaging and graphics. Volume 88(2021)
• Publication Date: 2021-03
• Subjects: Convolutional neural network -- Lymph node segmentation -- Positron emission tomography/computed tomography (PET/CT) -- Dilated convolution -- Imbalance class; Diagnostic imaging -- Periodicals; Imaging systems in medicine -- Periodicals; Diagnosis, Radioscopic -- Data processing -- Periodicals; Diagnostic Imaging -- Periodicals; Imagerie pour le diagnostic -- Périodiques; Diagnostic imaging; Periodicals; Electronic journals; Electronic journals; 616.0754
• DOI: 10.1016/j.compmedimag.2020.101851
• Languages: English
• ISSNs: 0895-6111

Abstract

Highlights: A multi-stage Atrous spatial pyramid pooling (MS-ASPP) sub-module that can be integrated into the SegNet architecture to improve the semantically accurate predictions and detailed segmentation along LN boundaries owing to the ability of multi-scale feature learning. A simple but effective cosine-sine (CS) loss function as an objective function for training different networks to deal with the problem of class imbalance for LN segmentation. The CS loss function can focus on learning to detect the hard-to-classify (misclassified) voxels but down-weight the well-classified voxels at the same time. Four-fold cross-validation is performed on 63 PET/CT data sets. The results show that we reach an average 77% Dice similarity coefficient score with CS loss function by trained DiSegNet. Abstract: Purpose: Automated lymph node (LN) recognition and segmentation from cross-sectional medical images is an important step for the automated diagnostic assessment of patients with cancer. Yet, it is still a difficult task owing to the low contrast of LNs and surrounding soft tissues as well as due to the variation in nodal size and shape. In this paper, we present a novel LN segmentation method based on a newly designed neural network for positron emission tomography/computed tomography (PET/CT) images. Methods: This work communicates two problems involved in LN segmentation task. Firstly, an efficient loss function named cosine-sine (CS) is proposed for the voxel class imbalance problem in the convolution network training process. Second, a multi-stage and multi-scale Atrous (Dilated) spatial pyramid pooling sub-module, named MS-ASPP, is introduced to the encoder-decoder architecture (SegNet), which aims to make use of multi-scale information to improve the performance of LN segmentation. The new architecture is named DiSegNet (Dilated SegNet). Results: Four-fold cross-validation is performed on 63 PET/CT data sets. In each experiment, 10 data sets are selected randomly for testing and the other 53 for training. The results show that we reach an average 77 % Dice similarity coefficient score with CS loss function by trained DiSegNet, compared to a baseline method SegNet by cross-entropy (CE) with 71 % Dice similarity coefficient. Conclusions: The performance of the proposed DiSegNet with CS loss function suggests its potential clinical value for disease quantification.