Montella, Annalaura (2025) Morphological biomarkers for early diagnosis in cancer. [Tesi di dottorato]

	Documento PDF montella_annalaura_37_COMPLETO.pdf Visibile a [TBR] Amministratori dell'archivio Download (5MB) | Richiedi una copia
	Documento PDF montella_annalaura_37_PARZIALE.pdf Download (1MB)

Abstract

Circulating tumor cells (CTCs) offer a minimally invasive approach for early cancer diagnosis and therapeutic responses monitoring, while also capturing intratumoral heterogeneity. However, their scarcity in the bloodstream and the absence of universal tumor-specific biomarkers pose significant challenges for their accurate detection and classification, particularly for label-dependent strategies. An emerging focus in liquid biopsy research involves leveraging morphological biomarkers to distinguish CTCs from normal blood cells, especially using label-independent approaches that integrate advanced imaging and machine learning technologies. This study evaluates the potential of combining morphological parameters to differentiate CTCs from white blood cells (WBCs). Cellular and nuclear morphology of tumor cell lines (neuroblastoma, ovarian, and breast cancer) and WBCs were analyzed using the Amnis ImageStreamX Mk II flow cytometer, a high-specific, single-cell, label-dependent system. Following DRAQ5 staining, at least 1,000 single cells per type were assessed using IDEAS software, which measured a total of 34 morphological features related to size, shape, and inner complexity. Principal component analysis (PCA) and hierarchical clustering revealed distinct tumor cell line clusters separated from WBCs. Further analysis of CTCs enriched from the peripheral blood of breast cancer patients—via CD45-positive cell depletion, EpCAM-positive cell sorting, and DRAQ5 staining—confirmed that CTCs shared similar morphological clustering patterns with tumor cell lines, distinct from WBCs. These findings suggest that nuclear and cellular morphology, along with inner complexity, represent promising biomarkers for CTCs detection. Integrating these morphological features with machine learning (ML) algorithms allowed the robust classification of tumor and normal cells, confirming that the synergic use of these characteristics and advanced detection technologies has the potential to improve CTCs identification, uncovering subpopulations missed by label-dependent methods, and enhancing early cancer detection, monitoring, and treatment strategies.

Downloads

Actions (login required)

Modifica documento