Diagnostic Accuracy of AI-Based Tools in Detecting Oral Cancer: An Umbrella Review
1
Public Health Dentistry, Amrita School of Dentistry, Kochi, Kochi, Kerala, India
Popul. Med. 2026;8(Supplement Supplement 1):
ABSTRACT
BACKGROUND:
Oral cancer is the 13th most common cancer globally in 2020, with high mortality and poor prognosis due to late-stage diagnoses. Despite known risk factors like tobacco and alcohol, over 50% of cases are detected at advanced stages, significantly reducing survival rates. Delays in diagnosis, both patient- and clinician-related, contribute to disease progression. Emerging artificial intelligence (AI) technologies offer potential in faster, more standardized, and early detection of oral cancer, particularly in resource-limited settings.
OBJECTIVE:
To evaluate the diagnostic accuracy of AI-based image analysis tools compared to conventional (non-AI) diagnostic methods for detecting oral cancer in adults presenting with suspicious oral lesions. Summary of the
METHODS:
A systematic search was conducted across MEDLINE (Ovid), Scopus, Google Scholar, Cochrane, and ProQuest databases. Duplicates were removed using Zotero, and studies were screened using Rayyan by two independent reviewers. Disagreements were resolved by a third reviewer. Critical appraisal of included study was assessed using JBI Critical Appraisal Checklist for Systematic Reviews and QUADAS-2 tool. The findings were summarised in tabulation form and discussed using the narrative method. The Corrected Covered Area (CCA) method was used to assess study overlap. Summary of the
RESULTS:
This umbrella review included 35 systematic reviews/meta-analyses. AI tools—particularly deep learning models such as convolutional neural networks—demonstrated high diagnostic accuracy, with sensitivity ranging from 62.5% to 100%, specificity from 57% to 100%, and overall accuracy from 64% to 100%. Deep learning approaches outperformed traditional machine learning models. Heterogeneity in datasets, model validation, and study design were noted. The CCA revealed moderate overlap among studies.
CONCLUSIONS:
AI-based diagnostic tools show promising accuracy for early detection of oral cancer and may complement clinical decision-making. Further prospective studies with standardized methodologies and real-world validation studies are needed before these tools can be routinely implemented in clinical settings.
Oral cancer is the 13th most common cancer globally in 2020, with high mortality and poor prognosis due to late-stage diagnoses. Despite known risk factors like tobacco and alcohol, over 50% of cases are detected at advanced stages, significantly reducing survival rates. Delays in diagnosis, both patient- and clinician-related, contribute to disease progression. Emerging artificial intelligence (AI) technologies offer potential in faster, more standardized, and early detection of oral cancer, particularly in resource-limited settings.
OBJECTIVE:
To evaluate the diagnostic accuracy of AI-based image analysis tools compared to conventional (non-AI) diagnostic methods for detecting oral cancer in adults presenting with suspicious oral lesions. Summary of the
METHODS:
A systematic search was conducted across MEDLINE (Ovid), Scopus, Google Scholar, Cochrane, and ProQuest databases. Duplicates were removed using Zotero, and studies were screened using Rayyan by two independent reviewers. Disagreements were resolved by a third reviewer. Critical appraisal of included study was assessed using JBI Critical Appraisal Checklist for Systematic Reviews and QUADAS-2 tool. The findings were summarised in tabulation form and discussed using the narrative method. The Corrected Covered Area (CCA) method was used to assess study overlap. Summary of the
RESULTS:
This umbrella review included 35 systematic reviews/meta-analyses. AI tools—particularly deep learning models such as convolutional neural networks—demonstrated high diagnostic accuracy, with sensitivity ranging from 62.5% to 100%, specificity from 57% to 100%, and overall accuracy from 64% to 100%. Deep learning approaches outperformed traditional machine learning models. Heterogeneity in datasets, model validation, and study design were noted. The CCA revealed moderate overlap among studies.
CONCLUSIONS:
AI-based diagnostic tools show promising accuracy for early detection of oral cancer and may complement clinical decision-making. Further prospective studies with standardized methodologies and real-world validation studies are needed before these tools can be routinely implemented in clinical settings.
| eISSN: | 2654-1459 |
We process personal data collected when visiting the website. The function of obtaining information about users and their behavior is carried out by voluntarily entered information in forms and saving cookies in end devices. Data, including cookies, are used to provide services, improve the user experience and to analyze the traffic in accordance with the Privacy policy. Data are also collected and processed by Google Analytics tool (more).
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
