Welcome to PracticeUpdate! We hope you are enjoying access to a selection of our top-read and most recent articles. Please register today for a free account and gain full access to all of our expert-selected content.
Already Have An Account? Log in Now
Deep Learning for the Automated Detection of Oral Lesions
TAKE-HOME MESSAGE
- This study evaluated the use of deep learning for the automated detection of oral lesions. A YOLOX-based convolutional neural network model for the detection and identification of oral lichen planus, oral leukoplakia, and oral squamous cell carcinoma (OSCC) on patients' photographs demonstrated its capacity to function as a valuable diagnostic tool to aid both experienced specialists and general dentists, especially those working in community hospitals, in achieving precise diagnoses of oral mucosal diseases. The model-based mobile application is a valuable, noninvasive, and cost-effective tool for nonspecialists to promptly identify oral lichen planus, oral leukoplakia, and OSCC, enabling early intervention and treatment.
- The timely identification and treatment of OSCC and precancerous lesions are crucial for enhancing the prognosis of patients. The authors present a model to improve the diagnostic accuracy of oral mucosal lesions, which can assist healthcare professionals and address the shortage of expertise and healthcare resources.
abstract
This abstract is available on the publisher's site.
Access this abstract now Full Text Available for ClinicalKey SubscribersOBJECTIVES
We aim to develop a YOLOX-based convolutional neural network model for the precise detection of multiple oral lesions, including OLP, OLK, and OSCC, in patient photos.
MATERIALS AND METHODS
We collected 1419 photos for model development and evaluation, conducting both a comparative analysis to gauge the model's capabilities and a multicenter evaluation to assess its diagnostic aid, where 24 participants from 14 centers across the nation were invited. We further integrated this model into a mobile application for rapid and accurate diagnostics.
RESULTS
In the comparative analysis, our model overperformed the senior group (comprising three most experienced experts with more than 10 years of experience) in macro-average recall (85 % vs 77.5 %), precision (87.02 % vs 80.29 %), and specificity (95 % vs 92.5 %). In the multicenter model-assisted diagnosis evaluation, the dental, general, and community hospital groups showed significant improvement when aided by the model, reaching a level comparable to the senior group, with all macro-average metrics closely aligning or even surpassing with those of the latter (recall of 78.67 %, 74.72 %, 83.54 % vs 77.5 %, precision of 80.56 %, 76.42 %, 85.15 % vs 80.29 %, specificity of 92.89 %, 91.57 %, 94.51 % vs 92.5 %).
CONCLUSION
Our model exhibited a high proficiency in detection of oral lesions, surpassing the performance of highly experienced specialists. The model can also help specialists and general dentists from dental and community hospitals in diagnosing oral lesions, reaching the level of highly experienced specialists. Moreover, our model's integration into a mobile application facilitated swift and precise diagnostic procedures.
Additional Info
Utilizing deep learning for automated detection of oral lesions: A multicenter study
Oral Oncol 2024 Aug 01;155(xx)106873, YJ Ye, Y Han, Y Liu, ZL Guo, MW HuangFrom MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.
Integrating artificial intelligence into dental medicine offers significant potential for enhancing the detection and diagnosis of oral lesions, ultimately improving patient care. In this multicenter study, the authors collected 1419 oral photos from smartphones and digital cameras from one hospital site between 2014 and 2023, consisting of 262 photos of oral lichen planus (OLP) lesions, 328 photos of oral leukoplakia (OLK) lesions, 799 photos of oral cancer (oral squamous cell carcinoma [OSCC]) lesions, and 30 photos of the normal mucosa. In this study, the authors developed YOLOX, a convolutional neural network (CNN) model, to classify these three lesions — OLP, OLK, and OSCC — by analyzing the captured digital photos.
The YOLOX/CNN model outperformed dental specialists in all major classification metrics, including sensitivity/recall (85.0% vs 77.5%), specificity (95.0% vs 92.5%), and precision (87.02% vs 80.29%). Additionally, when general dentists in community practice used the algorithm, the detection rate of OLP, OLK, and OSCC improved by 14% or more. The authors highlighted that a major limitation of the CNN model is its lack of consideration for additional clinical information such as symptoms, lesion location and texture, and disease duration.
The key considerations in evaluating the potential use and efficacy of the CNN model include the following:
Given these considerations, the CNN model could be used in low-resource settings by nonspecialists to diagnose OLP, OLK, and OSCC. However, biopsy with histopathologic examination is still the gold standard for confirmation. Future efforts should focus on integrating image analysis into intraoral scanners for initial lesion identification and expanding lesion classification categories to enhance clinical management.
References