University of Glasgow
Tumour growth patterns at the invasive front of colorectal cancer through integrated histopathological, transcriptomic and spatial-omic analysis
Abstract
dc:description.abstractColorectal cancer (CRC) is one of the most common cancers worldwide, yet patients with the same TNM stage often have very different outcomes, indicating that staging does not capture all of the biology that drives prognosis. Much of this missing information is concentrated at the invasive front. Here, the tumour growth pattern (TGP), which describes whether a tumour advances as a cohesive pushing border or dissects into the stroma as an infiltrative front, can be read on routine haematoxylin and eosin (H&E) sections and predicts survival. Why the infiltrative pattern carries a poorer prognosis and how cells at this interface are arranged in space have remained unclear. The central hypothesis of this thesis is that the spatial arrangement of cells at the invasive front, rather than their abundance alone, distinguishes the growth patterns and carries independent prognostic information. This was tested across two cohorts, the Glasgow Royal Infirmary test cohort (n = 538) and the TransSCOT validation cohort (n = 1,037), using histological TGP scoring, bulk transcriptomics (TempO-Seq), multiplex immunofluorescence with density-corrected proximity analysis (the Nearest Distance Index, NDI), and compartment-resolved spatial transcriptomics (GeoMx) with cell-state and ligand-receptor inference. A three-tier TGP classification was reproducible (weighted κ = 0.78) and was an independent prognostic factor for colon cancer-specific survival (HR = 1.65, p < 0.001), validated in TransSCOT. Bulk transcriptomics revealed stromal, epithelial-mesenchymal transition, and inflammatory programmes in infiltrative tumours, and proliferative programmes in pushing tumours. Importantly, cellular composition and density did not differ between patterns. Instead, spatial analysis showed that FAP+ cancer-associated fibroblasts (CAFs) lay closer to tumour cells and further from T cells in infiltrative tumours, and closer proximity of αSMA+FAP+ CAFs to tumour cells was independently associated with worse survival (HR = 2.46, p = 0.05). Spatial transcriptomics further linked the separation of CAFs from T cells to candidate inhibitory axes involving TIGIT-PVR/NECTIN2 and hyaluronan (HAS2-CD44). These findings reframe the TGP from a morphological descriptor into a spatial phenotype in which prognosis is determined by where cells sit rather than how many are present. The TGP offers a practical, H&E-based entry point into the spatial biology of the invasive front. As the spatial findings come from a single cohort and rely on inferred mechanisms, they are hypothesis-generating and require external validation.
Degree
thesis:*- Level dc:type.qualificationlevel
- PhD
- Grantor dc:publisher.institution
- University of Glasgow
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Woraharn, Walaiphorn
Subjects
dc:subject × 1Rights
- Language dc:language
- en
Identifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.5525/gla.thesis.86245