The prognostic value of tertiary lymphoid structures in cancer; a digital pathology and artificial intelligence approach

TLS typically resemble the structure and cellular composition of secondary follicles found in lymph nodes and mimic their functionality by acting as a site for antigen presentation, immune cell maturation, and antibody production. Chronic inflammation leading to TLS formation can arise from autoimmune diseases such as rheumatoid arthritis, Crohn’s disease, and type 1 diabetes.

Increasing evidence has implicated roles for TLS in some cancers, showing they may be of use as a new prognostic and predictive biomarker. Despite this, TLS have been inconsistently defined which has led to some contradictory results, particularly as a result of variability in both detection methodology and patient stratification strategies. This thesis aims to develop and apply a consistent definition of TLS in a single medium that can be automated for robust and reproducible detection and annotation in digital H&E-stained slides.