Theses and Dissertations (Anatomical Pathology)

Permanent URI for this collectionhttp://hdl.handle.net/2263/32491

Browse

Recent Submissions

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Evaluation of tumour infiltrating lymphocytes using cytokine expression as biomarkers for prognosis of melanoma in black South Africans
    Bida, Meshack (University of Pretoria, 2025-11-24)
    Background: Cutaneous melanoma induces an immune response from the host, evidenced by infiltration of immune effector cells, primarily lymphocytes and macrophages, within the tumour microenvironment. Tumour-infiltrating lymphocytes (TILs) represent a key component of this immune response and are increasingly used as a biomarker of prognosis and treatment response. However, the prognostic value of TILs remains debated, particularly in melanomas, where their presence does not uniformly translate into tumour eradication. Recent research has turned to the tumour immune microenvironment (TIME), especially cytokine signalling pathways, to better understand and enhance immune-mediated tumour clearance. Despite this interest, the expression of key cytokines in formalin-fixed, paraffin-embedded (FFPE) melanoma tissue remains understudied, particularly in relation to TIL distribution and function. Methods: We conducted a multi-platform investigation on 205 FFPE melanoma samples, histologically classified into brisk (n=65), non-brisk (n=60), and absent (n=80) TIL groups using AACR criteria. Interobserver reliability was quantified using intra-class correlation coefficients (ICCs). Cytokine expression (IFN-γ and TNF-α) was assessed by immunohistochemistry (IHC), quantified via immune scores, and validated with TissueFAXS imaging. Statistical analyses included chi-square tests with Cramér’s V, Kruskal–Wallis tests, logistic regression, and diagnostic performance metrics. Outlier cases were further analysed by transcriptomic profiling (NextSeq 2000, TSO 500 assay; 31-gene expression profile testing). To integrate functional insights with therapeutic relevance, structure-based modelling of BRAF was performed using Schrödinger Maestro, including Glide docking of dabrafenib and Prime MM-GBSA rescoring. Results and Discussion: TIL categorization showed excellent reproducibility (ICC 0.90–0.95). Cytokine expression correlated significantly with TIL groups (χ² p<0.001; Cramér’s V 0.31–0.41), with ordered increases confirmed by Kruskal–Wallis tests (χ² (2) =31.39–46.87, p<0.001; ε²≈0.15–0.22). Logistic regression demonstrated 3–4-fold higher odds of high cytokine expression in non-brisk and 6–8 fold higher odds in brisk cases versus absent. Diagnostic performance showed moderate sensitivity (0.66–0.72), specificity (0.62–0.66), and high negative predictive values (0.80–0.84). Discordant subsets were identified: 17% of brisk cases lacked TNF-α, 9% lacked IFN-γ, and a minority of absent TILs showed strong cytokine expression. Transcriptomic profiling of discordant cases confirmed cytokine expression in brisk outliers and revealed variable risk profiles among absent but cytokine-positive tumours. Structure-based analysis of BRAF confirmed canonical type-I binding of dabrafenib, with Glide SP docking yielding strong interaction scores (GlideScore −11.495; Emodel −115.824) and MM-GBSA rescoring supporting robust binding (ΔG_bind = −84.93 kcal·mol⁻¹). The pose demonstrated hallmark hinge hydrogen bonds to Cys532, hydrophobic packing at the gatekeeper/P-loop shelf, and π–π stacking with the DFG motif, consistent with a physiologically relevant binding mode. Conclusion: This is the first study to evaluate TIL categorisation in a South African melanoma cohort, underscoring its novelty and contribution to filling a critical gap in underrepresented populations. We demonstrate that histological TIL scoring using AACR tools in FFPE melanoma tissue is reproducible and functionally enhanced by cytokine IHC and transcriptomic profiling, uncovering immune activity not always reflected by morphology alone. Statistical analyses confirmed strong, graded associations between TIL infiltration and cytokine expression, while highlighting clinically relevant discordant subsets. Finally, structure-based modelling confirmed robust binding of dabrafenib to the BRAF ATP site, providing a mechanistic link between genomic profiling and potential therapeutic targeting. These findings underscore the value of a multi-step framework that integrates TIL histology, cytokine IHC, transcriptomics, and molecular docking to refine prognostic stratification and inform immunotherapy decision-making in melanoma.