Artificial Intelligence in the Screening of Oral Cancer: A Cross-Sectional Study on a Novel App-Based Approach for Primary Health Care Settings
Background: Oral cancer presents a significant public health challenge, particularly in resource-limited settings where early diagnosis is difficult. Integrating artificial intelligence (AI) into mobile health applications offers a promising solution. Objective: This study aimed to develop and evalu...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wolters Kluwer Medknow Publications
2025-04-01
|
| Series: | Journal of Indian Academy of Oral Medicine and Radiology |
| Subjects: | |
| Online Access: | https://journals.lww.com/10.4103/jiaomr.jiaomr_49_25 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Background:
Oral cancer presents a significant public health challenge, particularly in resource-limited settings where early diagnosis is difficult. Integrating artificial intelligence (AI) into mobile health applications offers a promising solution.
Objective:
This study aimed to develop and evaluate the diagnostic performance of an AI-powered mobile application, DiagnOCe, for oral cancer screening in primary healthcare settings.
Methods:
The DiagnOCe app was built using the MobileNet architecture via the Massachusetts Institute of Technology’s App Inventor platform. A dataset of 850 clinical images (425 cancerous, 425 normal mucosa) was divided into training (600), validation (100), and testing (150) sets. Diagnostic performance was assessed using sensitivity, specificity, accuracy, predictive values, and F1 score. Cohen’s kappa evaluated interobserver reliability.
Results:
The app demonstrated sensitivity, specificity, and accuracy of 85.33%, 88%, and 86.67%, respectively. The positive predictive value and negative predictive value were 87.67% and 85.71%. The receiver operating curve–area under the curve was 0.867, with an F1 score of 86.43%. Cohen’s kappa showed high interobserver agreement (0.88).
Conclusion:
DiagnOCe exhibits strong diagnostic accuracy and usability, making it a feasible tool for oral cancer screening in resource-limited settings. Future improvements could involve multicenter trials and enhanced AI models for greater precision. |
|---|---|
| ISSN: | 0972-1363 0975-1572 |