A deep learning approach to segmentation of nasopharyngeal carcinoma using computed tomography. (February 2021)

• Record Type: Journal Article
• Authors: Bai, Xiaoyu; Hu, Yan; Gong, Guanzhong; Yin, Yong; Xia, Yong
• Is Part Of: Biomedical signal processing and control. Volume 64(2021)
• Publication Date: 2021-02
• Subjects: Nasopharyngeal carcinoma segmentation -- Deep learning -- ResNeXt-50 U-Net -- Computed tomography; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2020.102246
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Propose a deep learning-based location-to-segmentation method for NPC segmentation. Location: define recall preserved loss for ResNeXt-50 U-Net to ensure over-segmentation. Segmentation: focus adaptively on NPC-related 3D volumes in each CT scan. Achieve Dice of 62.88 ± 8.12% on StructSeg-NPC 2019 training set and mean DSC of 61.81% in online testing. Abstract: Automated segmentation of Nasopharyngeal carcinoma (NPC) plays a critical role in the radiotherapy or chemo-radiotherapy for this cancer. Despite their improved performance, most deep learning models designed for this segmentation task use either magnetic resonance imaging (MRI) or multimodality data as input. In this paper, we propose a deep learning based algorithm called NPC-Seg for the segmentation of NPC using computed tomography (CT), which is less expensive and more available than MRI. This algorithm uses the location-to-segmentation framework. In the location step, it fine-tunes the pre-trained ResNeXt-50 U-Net with a newly proposed recall preserved loss to roughly segment the gross tumor volume (GTV) of each NPC. In the segmentation step, it fine-tunes the ResNeXt-50 U-Net again, but using the Dice loss, to segment the bounding box region detected in the location step on a patch-by-patch basis. We have evaluated the proposed NPC-Seg algorithm on the StructSeg-NPC dataset. Our algorithm achieves the Dice similarity coefficient (DSC) of 62.88 ± 8.12 % on 50 training data in the ten-fold cross-validation, substantially outperforming three existing deep learning methods, and also achieves an average DSC of 61.81% on the testing dataset in the online validation.