Jianjun Ren1,2, Xueping Jing3,4, Jing Wang1, Xue Ren5, Yang Xu1, Qiuyun Yang6, Lanzhi Ma7, Yi Sun7, Wei Xu8, Ning Yang9, Jian Zou1, Yongbo Zheng1, Min Chen1, Weigang Gan1, Ting Xiang1, Junnan An1, Ruiqing Liu10, Cao Lv11, Ken Lin12, Xianfeng Zheng1, Fan Lou12, Yufang Rao1, Hui Yang1, Kai Liu3, Geoffrey Liu2,13, Tao Lu14, Xiujuan Zheng3, Yu Zhao1. 1. Department of Otorhinolaryngology, West China Hospital, West China Medical School, Sichuan University, Chengdu, China. 2. Medical Oncology and Medical Biophysics, Princess Margaret Cancer Centre, Toronto, Ontario, Canada. 3. Department of Automation, College of Electrical Engineering and Information Technology, Sichuan University, Chengdu, China. 4. Department of Radiation Oncology, University Medical Center Groningen, Groningen, The Netherlands. 5. Department of Economic Statistics, School of Statistics and Management, Shanghai University of Finance and Economics, Shanghai, China. 6. Department of Forensics, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China. 7. Department of Preclinical Medicine, West China School of Preclinical and Forensic Medicine, Sichuan University, Chengdu, China. 8. Department of Biostatistics, Princess Margaret Cancer Centre, Toronto, Ontario, Canada. 9. College of Computer Science, Sichuan University, Chengdu, China. 10. Department of Otorhinolaryngology, Kunming City Women and Children Hospital, Kunming, China. 11. Department of Otorhinolaryngology, The Second Affiliated Hospital of Kunming Medical University, Kunming, China. 12. Department of Otorhinolaryngology, The Affiliated Children's Hospital of Kunming Medical University, Kunming, China. 13. Medicine and Epidemiology, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada. 14. Department of Otolaryngology and Head Neck Surgery, The First Affiliated Hospital of Kunming Medical University, Kunming, China.
Abstract
OBJECTIVES/HYPOTHESIS: To develop a deep-learning-based computer-aided diagnosis system for distinguishing laryngeal neoplasms (benign, precancerous lesions, and cancer) and improve the clinician-based accuracy of diagnostic assessments of laryngoscopy findings. STUDY DESIGN: Retrospective study. METHODS: A total of 24,667 laryngoscopy images (normal, vocal nodule, polyps, leukoplakia and malignancy) were collected to develop and test a convolutional neural network (CNN)-based classifier. A comparison between the proposed CNN-based classifier and the clinical visual assessments (CVAs) by 12 otolaryngologists was conducted. RESULTS: In the independent testing dataset, an overall accuracy of 96.24% was achieved; for leukoplakia, benign, malignancy, normal, and vocal nodule, the sensitivity and specificity were 92.8% vs. 98.9%, 97% vs. 99.7%, 89% vs. 99.3%, 99.0% vs. 99.4%, and 97.2% vs. 99.1%, respectively. Furthermore, when compared with CVAs on the randomly selected test dataset, the CNN-based classifier outperformed physicians for most laryngeal conditions, with striking improvements in the ability to distinguish nodules (98% vs. 45%, P < .001), polyps (91% vs. 86%, P < .001), leukoplakia (91% vs. 65%, P < .001), and malignancy (90% vs. 54%, P < .001). CONCLUSIONS: The CNN-based classifier can provide a valuable reference for the diagnosis of laryngeal neoplasms during laryngoscopy, especially for distinguishing benign, precancerous, and cancer lesions. LEVEL OF EVIDENCE: NA Laryngoscope, 130:E686-E693, 2020.
OBJECTIVES/HYPOTHESIS: To develop a deep-learning-based computer-aided diagnosis system for distinguishing laryngeal neoplasms (benign, precancerous lesions, and cancer) and improve the clinician-based accuracy of diagnostic assessments of laryngoscopy findings. STUDY DESIGN: Retrospective study. METHODS: A total of 24,667 laryngoscopy images (normal, vocal nodule, polyps, leukoplakia and malignancy) were collected to develop and test a convolutional neural network (CNN)-based classifier. A comparison between the proposed CNN-based classifier and the clinical visual assessments (CVAs) by 12 otolaryngologists was conducted. RESULTS: In the independent testing dataset, an overall accuracy of 96.24% was achieved; for leukoplakia, benign, malignancy, normal, and vocal nodule, the sensitivity and specificity were 92.8% vs. 98.9%, 97% vs. 99.7%, 89% vs. 99.3%, 99.0% vs. 99.4%, and 97.2% vs. 99.1%, respectively. Furthermore, when compared with CVAs on the randomly selected test dataset, the CNN-based classifier outperformed physicians for most laryngeal conditions, with striking improvements in the ability to distinguish nodules (98% vs. 45%, P < .001), polyps (91% vs. 86%, P < .001), leukoplakia (91% vs. 65%, P < .001), and malignancy (90% vs. 54%, P < .001). CONCLUSIONS: The CNN-based classifier can provide a valuable reference for the diagnosis of laryngeal neoplasms during laryngoscopy, especially for distinguishing benign, precancerous, and cancer lesions. LEVEL OF EVIDENCE: NA Laryngoscope, 130:E686-E693, 2020.
Authors: Blake S Wilson; Debara L Tucci; David A Moses; Edward F Chang; Nancy M Young; Fan-Gang Zeng; Nicholas A Lesica; Andrés M Bur; Hannah Kavookjian; Caroline Mussatto; Joseph Penn; Sara Goodwin; Shannon Kraft; Guanghui Wang; Jonathan M Cohen; Geoffrey S Ginsburg; Geraldine Dawson; Howard W Francis Journal: J Assoc Res Otolaryngol Date: 2022-04-20
Authors: Alberto Paderno; Cesare Piazza; Francesca Del Bon; Davide Lancini; Stefano Tanagli; Alberto Deganello; Giorgio Peretti; Elena De Momi; Ilaria Patrini; Michela Ruperti; Leonardo S Mattos; Sara Moccia Journal: Front Oncol Date: 2021-03-24 Impact factor: 6.244