Development of Optical Coherence Tomography and Angiography for Clinical Applications

Abstract

Skin and oral cancers present significant health challenges that require early detection for effective treatment. Traditional diagnostic methods are often invasive and resource-intensive, while non-invasive imaging techniques can lack the necessary capabilities for early detection. Optical Coherence Tomography (OCT) and OCT-base angiography (OCTA) offer high-resolution, non-invasive imaging abilities, enabling detailed structural and vascular assessments.

This thesis presents the development and clinical transition of two lab-built OCT systems, including developing a fast-acquisition system and an existing one, designed for enhanced in vivo applications in dermatology and dentistry. Significantly, several novel OCT image processing algorithms were developed to enhance imaging outcomes. The colour-depth encoding projection algorithm integrates depth information into OCTA data, facilitating improved visual interpretation of microvascular structures. The windowed Eigen-Decomposition algorithm effectively reduces motion artefacts during OCTA reconstruction, enhancing image clarity. Additionally, the Robust Ultrafast Projection Pipeline markedly accelerates the processing of OCT and OCTA projections, improving processing efficiency for clinical use despite minor trade-offs in image quality.

The OCT systems were applied in dermatology to monitor wound healing in a murine model and to establish a vasculature database of healthy human skin, utilising quantitative metrics. In dentistry, the system was adapted with a newly developed specialised intraoral scanning probe, successfully imaging multiple oral sites with angiography. Quantitative assessments of oral microvasculature confirmed the reliability of OCTA in providing reproducible data.

In conclusion, the development of fast-acquisition SSOCT systems with clinical approval provides a robust platform for further research and clinical application. The novel image processing algorithms contribute substantially to improved visualisation, motion artefact reduction, and processing efficiency. The automatic quantitative assessment methods developed represent an essential step towards clinical adoption, facilitating consistent and objective evaluations.

Date of Award	2025
Original language	English
Awarding Institution	University of Dundee
Supervisor	Chunhui Li (Supervisor) & Alessandro Perelli (Supervisor)