Epithelial tissue thickness improves optical coherence tomography's ability in detecting oral cancer

Background: OCT is a non-invasive imaging technique that enables the measurement of epithelial thickness and architectural changes, which can help in the diagnosis of pre-cancerous and cancerous lesions. The purpose of the study was to assess whether epithelial tissue thickness improves optical coherence tomography's ability in detecting oral cancer. Patients and methods: Surgically resected oral margins from 60 patients diagnosed with oral squamous cell carcinoma were subjected to OCT. Three OCT measurements (immediate, 1 h and 24 h post-resection) were conducted per resected tissue specimen to look at the effect of saline and formalin on the specimen and its effect on the reproducibility of the OCT. OCT was, then, used to measure the epithelial tissue thickness in cancer-free and cancer-involved margins in eight oral anatomical locations. This data was, then, combined with architectural changes data to calculate the sensitivity and specificity. Results: An overall of 189 cancer-free margins and 51 cancer-involved margins had their epithelial thickness measured using OCT and compared to histopathology. With regards to the validity of the OCT and histopathological measurements, epithelial thickness showed good correlation between different readings at all oral sites. With regards to the reproducibility of the OCT measurements, the mean epithelial thickness for all measurements at first (immediate) and second (1 h post-resection - saline preserved) measurements was not significantly different. Underestimation of the epithelial depth in cancer-free margins was 20 mu m, while in the cancer-involved margins was 10 mu m. Combining data from architectural changes and epithelial thickness, a sensitivity of 92% and a specificity of 94% was achieved. Conclusion: Oral epithelium measurements using OCT were valid compared to those made with gold standard pathology. Measurements made using OCT was also reproducible with minor underestimation. Epithelial thickness, combined with architectural changes, led to high accuracy in differentiating between cancer-free and cancer-involved margins.