Development of a self-constrained 3D DenseNet model in automatic detection and segmentation of nasopharyngeal carcinoma using magnetic resonance images. (November 2020)

• Record Type: Journal Article
• Authors: Ke, Liangru; Deng, Yishu; Xia, Weixiong; Qiang, Mengyun; Chen, Xi; Liu, Kuiyuan; Jing, Bingzhong; He, Caisheng; Xie, Chuanmiao; Guo, Xiang; Lv, Xing; Li, Chaofeng
• Is Part Of: Oral oncology. Volume 110(2020)
• Publication Date: 2020-11
• Subjects: Nasopharyngeal carcinoma -- Magnetic resource images -- Deep learning -- Detection -- Automatic segmentation; NPC Nasopharyngeal carcinoma -- MRI magnetic resonance image -- RT radiotherapy -- GPU Graphic Processing Units -- CNN convolutional neural network -- SC-DenseNet self-constrained 3D DenseNet -- DSC dice similarity coefficient -- PACS picture archiving and communication system -- DICOM Digital Imaging and Communications in Medicine -- PPV positive predictive value -- NPV negative predictive value -- CI confidence interval -- AUC area under curve -- ROC receiver operating characteristic curve -- SPSS Statistical Program for Social Sciences -- EBV Epstein-Barr virus -- GTV gross tumor volume; Mouth -- Cancer -- Periodicals; Mouth -- Tumors -- Periodicals; Mouth Diseases -- Periodicals; Mouth Neoplasms -- Periodicals; Bouche -- Cancer -- Périodiques; Bouche -- Tumeurs -- Périodiques; Tumeurs -- Périodiques; Electronic journals; 616.9943105
• DOI: 10.1016/j.oraloncology.2020.104862
• Languages: English
• ISSNs: 1368-8375

Abstract

Highlights: A dual-task AI model was developed to detect and segment NPC in MRI automatically. The model showed higher overall accuracy than experienced radiologists to detect NPC. The model showed encouraging dice similarity coefficient in segmentation of NPC. Abstract: Objectives: We aimed to develop a dual-task model to detect and segment nasopharyngeal carcinoma (NPC) automatically in magnetic resource images (MRI) based on deep learning method, since the differential diagnosis of NPC and atypical benign hyperplasia was difficult and the radiotherapy target contouring of NPC was labor-intensive. Materials and methods: A self-constrained 3D DenseNet (SC-DenseNet) architecture was improved using separated training and validation sets. A total of 4100 individuals were finally enrolled and split into the training, validation and test sets at a proximate ratio of 8:1:1 using simple randomization. The diagnostic metrics of the established model against experienced radiologists was compared in the test set. The dice similarity coefficient (DSC) of manual and model-defined tumor region was used to evaluate the efficacy of segmentation. Results: Totally, 3142 nasopharyngeal carcinoma (NPC) and 958 benign hyperplasia were included. The SC-DenseNet model showed encouraging performance in detecting NPC, attained a higher overall accuracy, sensitivity and specificity than those of the experienced radiologists (97.77% vs 95.87%, 99.68% vs 99.24% and 91.67% vs 85.21%, respectively). Moreover, the model also exhibited promising performance in automatic segmentation of tumor region in NPC, with an average DSC at 0.77 ± 0.07 in the test set. Conclusions: The SC-DenseNet model showed competence in automatic detection and segmentation of NPC in MRI, indicating the promising application value as an assistant tool in clinical practice, especially in screening project.