Artificial Intelligence and Machine Learning solutions-based oral cancer screening and detection: A scoping review

Area of Science:

• Medical Informatics
• Computational Biology
• Oncology

Background:

• Oral cancer presents a substantial global health challenge, underscoring the critical need for early detection to improve patient prognoses.
• Conventional oral cancer screening methods are often subjective and labor-intensive, relying heavily on the expertise of clinicians.
• Artificial Intelligence (AI) and Machine Learning (ML) offer transformative potential by utilizing deep learning and image analysis to enhance diagnostic precision and streamline screening workflows.

Purpose of the Study:

• This scoping review comprehensively evaluates the application of AI and ML technologies in the detection of oral cancer.
• The review focuses on AI's capabilities in classification, segmentation, and early diagnosis of oral cancer.
• Additionally, it examines the role of AI in predicting patient risk, optimizing treatment planning, and developing prognostic models.

Main Methods:

• A systematic literature search was performed across major scientific databases to identify peer-reviewed studies detailing AI/ML techniques for oral cancer screening.
• Key AI models analyzed include Convolutional Neural Networks (CNNs), Fully Convolutional Networks (FCNNs), and ensemble learning methods.
• These models were applied to diverse data types, including intraoral images, histopathology slides, and fluorescence visualization data.

Main Results:

• AI-powered fluorescence visualization achieved high diagnostic performance, with 98.0% sensitivity and 92.7% specificity, outperforming traditional methods.
• Deep learning applied to histopathology analysis demonstrated improved risk stratification for oral leukoplakia, while cloud-based AI solutions enhanced diagnostic accessibility.
• AI models demonstrated comparable or superior accuracy to pathologists in lesion classification, and AI-driven immunohistochemical tools improved tumor marker detection.

Conclusions:

• AI/ML technologies represent a significant advancement in oral cancer screening, offering improved accuracy and efficiency.
• Further rigorous validation and real-world implementation studies are essential for seamless clinical integration of these AI tools.
• Future research should prioritize AI interpretability, ethical considerations, and robust multi-institutional collaborations to address dataset limitations and algorithmic bias, ultimately maximizing clinical impact.