ABSTRACT
OBJECTIVE
Little is known about the efficacy of using artificial intelligence to identify laryngeal carcinoma from images of vocal lesions taken in different hospitals with multiple laryngoscope systems. This multicenter study aimed to establish an artificial intelligence system and provide a reliable auxiliary tool to screen for laryngeal carcinoma.
Study Design
Multicenter case-control study
Setting
Six tertiary care centers
Participants
Laryngoscopy images were collected from 2179 patients with vocal fold lesions.
Outcome Measures
An automatic detection system of laryngeal carcinoma was established and used to distinguish malignant and benign vocal lesions in 2179 laryngoscopy images acquired from 6 hospitals with 5 types of laryngoscopy systems. Pathological examination was the gold standard for identifying malignant and benign vocal lesions.
Results
Out of 89 cases in the malignant group, the classifier was able to correctly identify laryngeal carcinoma in 66 patients (74.16%, sensitivity). Out of 640 cases in the benign group, the classifier was able to accurately assess the laryngeal lesion in 503 cases (78.59%, specificity). Furthermore, the region-based convolutional neural network(R-CNN) classifier achieved an overall accuracy of 78.05%, with a 95.63% negative predictive and a 32.51% positive predictive value for the testing dataset.
Conclusion
This automatic diagnostic system has the potential to assist clinical laryngeal carcinoma diagnosis which may improve and standardize the diagnostic capacity of laryngologists using different laryngoscopes.
This article is protected by copyright. All rights reserved.
No comments:
Post a Comment