Cribriform morphology in prostate cancer: an integrated approach combining digital pathology and transcriptomic analysis

• Ano de defesa: 2025
• Autor(a) principal: Araujo, Francisco Jonathan dos Santos
• Orientador(a): Melo, Vânia Maria Maciel
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Link de acesso: http://repositorio.ufc.br/handle/riufc/83184
• topic: Câncer de próstata; Morfologia cribriforme; Patologia digital; Expressão gênica; Biomarcadores prognósticos; Prostate cancer; Cribriform morphology; Digital pathology; Gene expression; Prognostic biomarkers

Cribriform morphology has been increasingly recognized as a histological marker of aggressiveness in prostate cancer, being associated with a higher risk of recurrence, metastasis, and cancer-specific mortality. However, the molecular mechanisms underlying this architectural pattern remain poorly understood. This thesis proposes a multidimensional approach to investigate the morphological and molecular determinants associated with tumor aggressiveness in prostate cancer, with a particular focus on cribriform architecture. In the first chapter, 72 radical prostatectomy samples were analyzed using quantitative digital pathology. Through whole-slide imaging and the QuPath software, immunohistochemical markers PTEN, Ki-67, ATM, and CD8 and key morphological features—such as Gleason pattern 4 percentage, tumor extent, and cribriform morphology—were quantified. The analysis identified a group of markers associated with increased tumor aggressiveness, with cribriform morphology showing significant correlation with higher pathological stage and absence of organ confinement. A logistic regression model combining Gleason pattern 4 percentage and ATM expression accurately predicted the presence of cribriform morphology (AUC = 0.79). In the second chapter, an in silico RNA-Seq analysis was conducted using publicly available data (GEO- NCBI) to characterize the gene expression profile associated with cribriform morphology. A total of 55 differentially expressed genes were identified between cribriform and non- cribriform tumors. Genes such as CPNE4, PLA2G7, PTPRJ, GCNT1, AMACR, AR, and ERG were linked to immune evasion, lipid metabolism, and hormonal signaling pathways. Functional enrichment analysis revealed transcriptional reprogramming consistent with aggressive tumor behavior. Although methodologically independent, the two chapters present complementary findings that reinforce the association between cribriform morphology and a biologically aggressive phenotype in prostate cancer, contributing to the identification of prognostic biomarkers and potential therapeutic targets