A brush biopsy—a simple scrape of cells from the mouth—could one day tell a clinician whether a suspicious white patch is on its way to becoming cancer, without needing to cut out tissue. Oral squamous cell carcinoma kills more than half of patients once it spreads to lymph nodes, but catching it while the lesion is still pre-malignant is difficult: most pre-malignant lesions never turn cancerous, and repeated surgical biopsies are painful and impractical. Clinicians currently rely on intuition and basic imaging to decide where to sample, meaning high-risk lesions are often missed. This project aims to identify the molecular and metabolic changes that mark a pre-malignant lesion as high-risk, using a mouse model that mimics the stepwise progression of the disease. The researcher will track which cells give rise to tumours, analyse the immune environment and metabolism of transforming lesions, and build an AI tool that automatically scans stained tissue slides to flag dangerous sub-classes. If successful, the work could replace repeated surgical biopsies with a brush-based test, making surveillance far less invasive and more widely accessible. The same biomarkers might later become targets for precision therapies that intercept cancer before it invades.
View original technical description
Background: The prognosis for advanced-stage oral squamous cell carcinoma (OSCC) is very poor, with patient survival dropping below 30% once the tumor spreads to the lymph-nodes. Early detection and treatment are crucial for preventing disease progression, improving treatment outcomes, and enhancing patients’ survival and quality of life. This underscores the importance of early diagnosis, ideally before the tumor becomes invasive, while the lesion is still pre-malignant. However, not all pre-malignant lesions develop into cancer, and repeat biopsy can discourage patient participation and may not be feasible in scenarios where the entire mucosa could be altered. With limited information on which lesions would have higher / the highest transforming risk, clinicians usually rely on intuition and basic imaging to decide where to take biopsies. This highlights the urgent need for reliable biomarkers that can identify high-risk lesions, improving lesion assessment and guiding biopsies. Developing accurate and efficient diagnostic tools requires a deeper understanding of oral mucosal biology in its normal and premalignant state to pinpoint the precise cellular and molecular changes involved in cancer progression. Aims: The study aims to identify changes in the transcriptome, cellular metabolism, and immune microenvironment of pre-malignant oral lesions that are linked to a higher risk of malignant transformation. Methods: The 4NQO mouse model will be utilized to study clonal evolution and the stepwise progression of OSCC through various pre-malignant stages. I will employ lineage tracing methods by labeling oral epithelial stem cells and differentiating cells in these mouse models. Tissue and brush biopsies from the lesions will be histologically assessed, and high-content imaging, along with untargeted LC-MS analysis, spatial transcriptomics, and spectral cytometry, will be used to identify changes unique to high-risk transforming lesions. I will examine mitochondrial structural and functional changes, alterations in cellular and metabolic phenotypes, and shifts in the immune landscape in the high-risk transforming lesions. Additionally, an AI-based deep learning tool will be developed for automated analysis of histological slides combined with immunofluorescence-based markers to detect different lesion sub-classes with accuracy. How this research will be used: The identified biomarkers will be used to develop a robust system for detecting high-risk lesions using brush-biopsies, eliminating the need for frequent tissue-biopsies for histological evaluation. By employing AI-tools, I plan to enhance the system’s accuracy for identifying potential lesions. The next possible application would be identifying biomarkers overexpressed in high-risk pre-malignant lesions and using them to develop precision oral cancer interception strategies.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know