The Digital Assessment of Biomarkers in Breast and Colorectal Cancer

Digital pathology holds promise to transform biomarker analysis by improving accuracy, efficiency, and reducing turnaround times with image analysis and artificial intelligence tools. This thesis focuses on the development and utilization of image analysis tools on mismatch repair (MMR) protein in colorectal cancer and tumour infiltrating lymphocytes (TILs) in breast cancer (BC). Using a semi-automated image analysis workflow, we utilize a novel means, tumour to stroma ratio, to objectively classify MMR status. We deployed a similar semi-automated workflow to enumerate TILs. We utilized this to quantify TILs in BC cases treated with neoadjuvant chemotherapy or chemoradiation to explore potential predictors of response to therapy. We identified significance between chemotherapy and high TILs in patients achieving a pathological complete response (pCR). Chemoradiation resulted in increased TILs among patients achieving a pCR. Utilizing semi-automated workflows, it’s possible to augment existing biomarker workflows and novel markers to better prognostication and guide treatment decisions.