Critical review of OCT in clinical practice for the assessment of oral lesions.

Background: Optical Coherence Tomography (OCT) is an advanced imaging technique that is widely used in ophthalmology and is increasingly being applied in other fields of medicine. In oral oncology, OCT offers high-resolution, non-invasive (uses non-ionizing light), label-free, real-time imaging, providing detailed insights into tissue microanatomy and cellular structures, thus having the potential to improve early detection, monitoring and cost-effective screening of high-risk populations. However, significant challenges remain in applying OCT to OSCC and OPMDs, particularly in clinical practice. Methods: A comprehensive search of PUBMED, SCOPUS, and Web of Science databases was performed up to October 2024. Additional manual searches were conducted by screening article bibliographies. Inclusion criteria encompassed studies published in English involving human subjects and evaluating the role of OCT in OSCC and OPMD assessment, OCT utilization for margin resection, and artificial intelligence (AI)-assisted interpretation of OCT images. After removal of duplicates and screening of titles and abstracts, full-text analysis was conducted on eligible studies. Results: The technique has been investigated for its accuracy in identifying malignant changes in tissues before surgery and/or evaluating resection margins during surgery. Although early studies, primarily in animal models, have been extended to humans and have demonstrated the potential of OCT to accurately assess resection margins and identify precancerous lesions, significant limitations persist. The high cost of OCT equipment reduces its accessibility, availability and widespread use as a common investigation methodology in non-experimental settings. In addition, there is significant heterogeneity in the methodologies used to interpret OCT data, which is strictly operator dependent and may affect standardization and reproducibility of results. This is further complicated by the introduction and increased trend to adopt artificial intelligence (AI) algorithms in imaging evaluation. Machine learning and deep learning algorithms have shown superior diagnostic sensitivity and accuracy compared to clinician judgment. However, especially when used to assess resection margins, these algorithms may be significantly affected by sample extension and preparation, which remains a barrier to the routine clinical application of OCT systems. Conclusion: Addressing the advantages and challenges of this emerging technique may help focus future research on standardizing application protocols and enhancing AI-assisted analysis to improve diagnostic performance and facilitate clinical translation.