Non-Cochrane · DTA · PM42774089
Diagnostic performance of artificial intelligence in oral squamous cell carcinoma detection: a higher-order evidence synthesis through umbrella review and meta-meta-analysis.
Jodalli Praveen, Vinay Vineet, Chavan Mahesh, et al. · Frontiers in dental medicine · 2026 · PubMed · doi:10.3389/fdmed.2026.1915105
Findings
9 findings
| Population | ||||
|---|---|---|---|---|
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Artificial intelligence (machine and deep learning) diagnostic models, any image modality | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Implementation feasibility (cost-effectiveness, workflow integration) Any clinical setting; no geographic restriction | Implementation feasibility: Qualitative evidence only; No. of reviews (primary studies): 7 (Mixed); Very Low; Major barriers identified; minimal real-world implementation Very low Downgraded: imprecision, indirectness 7 studies |
| People evaluated for OSCC | AI diagnostic systems | Histopathological diagnosis | Sensitivity | Sensitivity 0.90 [0.81, 0.99] GRADE not reported 6 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Histopathology-based AI (whole slide images, tissue microarrays) | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Sensitivity for OSCC detection Any clinical setting; no geographic restriction | Sensitivity 0.87 [0.82, 0.91] 87 [82, 91] per 100 patients with OSCC correctly identified Moderate Downgraded: inconsistency 9 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Histopathology-based AI (whole slide images, tissue microarrays) | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Specificity for OSCC detection Any clinical setting; no geographic restriction | Specificity 0.86 [0.81, 0.90] 86 [81, 90] per 100 patients without OSCC correctly identified Moderate Downgraded: inconsistency 9 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Clinical photography-based AI | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Specificity for OSCC detection Any clinical setting; no geographic restriction | Specificity 0.85 [0.80, 0.89] 85 [80, 89] per 100 patients without OSCC correctly identified Low Downgraded: imprecision, inconsistency 6 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Clinical photography-based AI | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Sensitivity for OSCC detection Any clinical setting; no geographic restriction | Sensitivity 0.87 [0.83, 0.90] 87 [83, 90] per 100 patients with OSCC correctly identified Low Downgraded: imprecision, inconsistency 6 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Optical coherence tomography (OCT)-based AI | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Specificity for OSCC detection Any clinical setting; no geographic restriction | Specificity 0.88 [0.82, 0.93] 88 [82, 93] per 100 patients without OSCC correctly identified Very low Downgraded: imprecision, indirectness 3 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Optical coherence tomography (OCT)-based AI | Histopathological diagnosis (biopsy/surgical specimen) as reference standard | Sensitivity for OSCC detection Any clinical setting; no geographic restriction | Sensitivity 0.90 [0.85, 0.94] 90 [85, 94] per 100 patients with OSCC correctly identified Very low Downgraded: imprecision, indirectness 3 studies |
| Patients with oral lesions assessed for oral squamous cell carcinoma (OSCC); evidence from systematic reviews of diagnostic accuracy studies | Histopathology-based AI | Clinical photography-based AI | Diagnostic odds ratio for OSCC detection Any clinical setting; no geographic restriction | Ratio of diagnostic odds ratios 2.15 [1.42, 3.26] Low 15 studies |
Included primary studies 15
Open in Studies| Study | Year | Also included by | Outcomes it informs here | Trial data |
|---|---|---|---|---|
| Kim 2022 Efficacy of Artificial Intelligence-Assisted Discrimination of Oral Cancerous Lesions from Normal Mucosa Based on the Oral Mucosal Image:… PMID:35884560 | 2022 | 2 other reviews | not stated | not extracted yet |
| Li 2024 Diagnostic accuracy of artificial intelligence assisted clinical imaging in the detection of oral potentially malignant disorders and oral… PMID:38652301 | 2024 | 2 other reviews | not stated | not extracted yet |
| Elmakaty 2022 Accuracy of artificial intelligence-assisted detection of Oral Squamous Cell Carcinoma: A systematic review and meta-analysis. PMID:35931404 | 2022 | 1 other review | not stated | not extracted yet |
| Sahoo 2024 Diagnostic performance of artificial intelligence in detecting oral potentially malignant disorders and oral cancer using medical… PMID:39568787 | 2024 | 1 other review | not stated | not extracted yet |
| Alabi 2021 Machine learning application for prediction of locoregional recurrences in early oral tongue cancer: a Web-based prognostic tool. PMID:31422502 | 2019 | this review only | not stated | not extracted yet |
| Chowdhury 2025 Role of Artificial Intelligence in the Diagnosis of Oral Squamous Cell Carcinoma: A Systematic Review. PMID:40330404 | 2025 | this review only | not stated | not extracted yet |
| Jerjes 2024 Enhancing Oral Cancer Detection: A Systematic Review of the Diagnostic Accuracy and Future Integration of Optical Coherence Tomography… PMID:39407882 | 2024 | this review only | not stated | not extracted yet |
| Khanagar 2021 Application and Performance of Artificial Intelligence Technology in Oral Cancer Diagnosis and Prediction of Prognosis: A Systematic Review. PMID:34072804 | 2021 | this review only | not stated | not extracted yet |
| Khanagar 2023 Application and Performance of Artificial Intelligence (AI) in Oral Cancer Diagnosis and Prediction Using Histopathological Images: A… PMID:37371706 | 2023 | this review only | not stated | not extracted yet |
| Mahmood 2020 Use of artificial intelligence in diagnosis of head and neck precancerous and cancerous lesions: A systematic review. PMID:32674040 | 2020 | this review only | not stated | not extracted yet |
| Malhotra 2024 Diagnostic Accuracy of Artificial Intelligence Compared to Biopsy in Detecting Early Oral Squamous Cell Carcinoma: A Systematic Review and… PMID:39205556 | 2024 | this review only | not stated | not extracted yet |
| Nieri 2025 Diagnosis of Oral Cancer With Deep Learning. A Comparative Test Accuracy Systematic Review. PMID:40163741 | 2025 | this review only | not stated | not extracted yet |
| Pirayesh 2024 Deep Learning-Based Image Classification and Segmentation on Digital Histopathology for Oral Squamous Cell Carcinoma: A Systematic Review… PMID:39256895 | 2024 | this review only | not stated | not extracted yet |
| Ren 2025 The diagnostic value of artificial intelligence in oral squamous cell carcinoma: A systematic review and meta-analysis. PMID:40518015 | 2025 | this review only | not stated | not extracted yet |
| Thakuria 2024 Deep learning for early diagnosis of oral cancer via smartphone and DSLR image analysis: a systematic review. PMID:39587051 | 2024 | this review only | not stated | not extracted yet |
Overlap with other reviews
Reviews that include the same primary studies are not independent evidence. Shared counts match studies by PMID, DOI or NCT, and otherwise by author and year (a probable match). Only the non-Cochrane dental reviews have study lists, so overlap with Cochrane reviews is not counted yet.
Review question
| Element | Planned (methods) | Found (results) |
|---|---|---|
| Population | Systematic reviews and/or meta-analyses on diagnosis of histologically confirmed OSCC in humans (OPMD reviews only with separate OSCC data) | 15 systematic reviews (2020 to 2025) synthesizing 341 primary study inclusions (about 241 unique primary studies; CCA 9.06%) |
| Intervention | Any AI-based diagnostic technology (traditional machine learning, deep learning, hybrid models) on histopathology, clinical photographs, OCT, autofluorescence, cytology, radiography or hyperspectral imaging | AI on histopathology (n=9 reviews), clinical photography (n=6), OCT (n=3), mixed/other modalities (n=5); deep learning in 12 reviews |
| Comparison | Histopathological diagnosis (biopsy or surgical specimen) as reference standard; expert clinical examination or conventional non-AI methods | Histopathology reference standard; clinical photography-based AI as comparator for histopathology-based AI (relative DOR) |
| Outcomes | Primary: sensitivity, specificity, DOR, AUC, PLR/NLR. Secondary: clinical utility, implementation feasibility, economic evaluations, barriers to adoption | Pooled sensitivity by meta-meta-analysis (n=6 reviews); GRADE SoF for sensitivity and specificity by modality (histopathology n=9, clinical photography n=6, OCT n=3), relative DOR (n=15), implementation feasibility (n=7, qualitative); specificity not pooled by meta-meta-analysis |
| Study designs | Completed peer-reviewed systematic reviews and/or meta-analyses with search of at least two databases | Systematic reviews only (n=7) and systematic reviews with meta-analysis (n=8) |
| Timing | Not specified | Not specified |