This review aimed to evaluate the evidence from previous studies on smartphone-based imaging and artificial intelligence (AI)–assisted mobile health (mHealth) technologies for the early detection of oral potentially malignant disorders (OPMDs) and oral cancer, focusing on their diagnostic performance and feasibility in low-resource settings. A structured literature review was conducted, including 23 peer-reviewed articles published between 2015 and 2024. Eligible studies investigated smartphone-based imaging, tele-dentistry platforms, dual-modality autofluorescence systems, mobile cytology tools, mHealth community programs, and AI-driven models. Data extraction was based on study design, sample size, assessment tool, reference standard, and diagnostic outcomes. The 23 studies demonstrated that smartphone-based imaging and mHealth platforms are not only feasible but also effective for OPMDs and oral cancer screening in both community and clinical settings. These platforms reported sensitivity ranged from 70% to 99%, specificity from 64% to 100%, and accuracy from 81% to 97%, depending on the device used, the screening context, and operator expertise. Moreover, AI-driven models such as DenseNet, HRNet, MobileNet, and CNN-based ensembles achieved diagnostic accuracies of 84–95%, in some cases even approaching specialist-level performance. Similarly, dual-modality autofluorescence and white-light imaging systems enhanced classification accuracy, reaching 79–87%. Importantly, community-based mHealth programs involving frontline health workers showed high diagnostic agreement with specialists (κ up to 0.92), thereby enabling large-scale, cost-effective screening. Moreover, MeMoSA® applications demonstrated strong concordance with conventional oral examination (sensitivity 92–94%, specificity up to 95.5%), thus supporting their integration into structured referral pathways.
Alm et al. (Sat,) studied this question.