Understanding how the architecture of fibrous scaffolds influences cellular behavior and tissue development is a fundamental challenge in tissue engineering and regenerative medicine. The three-dimensional organization of the fibrous network play a key role in shaping the cellular microenvironment, regulating processes such as cell adhesion, migration, proliferation, extracellular matrix remodeling, and the mechanical behavior of engineered constructs. Consequently, structural descriptors including fiber diameter, spatial orientation, connectivity, and intersection density are essential for correlating scaffold architecture with its biological and biomechanical performance. Despite their importance, the quantitative characterization of complex fibrous networks remains challenging, especially when moving from planar images to volumetric datasets. Irregular fibers geometry, branching, overlapping structures, and imaging-related intensity variations complicate the reliable extraction of morpho-topological descriptors. Furthermore, most existing approaches are limited to two-dimensional analyses, providing only partial representation of the actual spatial organization of fibrous architecture. The aim of this thesis is to develop and validate an automated image-based framework for the three-dimensional morpho-topological characterization of fibrous scaffolds and biological tissues imaged by multiphoton microscopy. The proposed methodology extends an established two-dimensional graph-based approach to volumetric data, providing a quantitative and reproducible framework for network reconstruction and structural analysis. The workflow combines synthetic phantom generation for validation under controlled geometrical conditions with an analysis pipeline for experimental volumetric datasets. The computational framework includes images denoising, segmentation, three-dimensional skeletonization, graph reconstruction and automatic identification of nodes, branches, and intersections. These steps enable the extraction of quantitative descriptors such as fiber diameter, connectivity, real intersection density, spatial fiber orientation. After validation on synthetic phantoms, the pipeline was applied to volumetric datasets of electrospun polycaprolactone, fibroin, cellulose, and human cardiovascular tissues, including the ascending aorta and tricuspid valves. The results that the proposed approach enable the automatic reconstruction and characterization of complex three-dimensional fibrous networks, providing a quantitative description of sample architecture. In this perspective, this work offers a valuable tool for the objective assessment of biomimetic scaffolds, biological tissues, and engineered constructs, supporting future investigations into the relationship between fibrous architecture, tissue function, and regenerative processes.
La caratterizzazione tridimensionale dell’architettura fibrosa degli scaffold ingegnerizzati rappresenta un aspetto cruciale della tissue engineering e della medicina rigenerativa. In particolare, l’organizzazione tridimensionale della rete fibrosa influenza le proprietà meccaniche e biologiche del materiale, nonché processi cellulari quali adesione, migrazione e proliferazione.Parametri micro-architetturali quali diametro delle fibre, orientamento spaziale, connettività e densità di intersezione costituiscono pertanto descrittori fondamentali per correlare la morfologia dello scaffold alla sua funzionalità biologica e biomeccanica. Nonostante la rilevanza di tali parametri, la caratterizzazione quantitativa tridimensionale di reti fibrose rappresenta ancora una sfida metodologica, soprattutto quando l’analisi viene estesa da immagini planari a dati volumetrici tridimensionali. La geometria irregolare delle fibre, la presenza di intersezioni, sovrapposizioni e variazioni locali di intensità rendono complessa l’estrazione automatica di informazioni morfometriche affidabili da immagini volumetriche sperimentali. Inoltre, molti approcci comunemente utilizzati si basano su analisi bidimensionali, che non consentono di descrivere in modo completo la reale organizzazione spaziale dello scaffold e possono introdurre una semplificazione significativa della sua architettura. Il presente lavoro di tesi si propone di sviluppare e validare una pipeline automatizzata image-based per l’analisi tridimensionale di scaffold fibrosi e tessuti biologici studiati mediante microscopia multifotone. L’obiettivo è fornire uno strumento quantitativo e riproducibile per la ricostruzione, modellazione e caratterizzazione morfometrica e topologica di reti fibrose complesse, superando i limiti delle analisi planari e consentendo una descrizione più fedele dell’architettura tridimensionale del campione. La metodologia sviluppata comprende un modulo per la generazione di phantom tridimensionali sintetici per la validazione del metodo in condizioni geometriche controllate, e una pipeline di analisi di stack volumetrici sperimentali. Quest’ultima include fasi di denoising, segmentazione ed estrazione dello skeleton tridimensionale. A partire dalla struttura scheletrizzata, l’algoritmo identifica automaticamente nodi, ramificazioni e intersezioni, permettendo l’estrazione di descrittori quantitativi quali diametro delle fibre, connettività, densità di intersezioni e orientamento tridimensionale. Dopo la validazione su phantom sintetici, la pipeline è stata applicata a stack reali di scaffold e tessuti fibrosi costituiti da materiali di interesse per la tissue engineering, tra cui policaprolattone, fibroina, cellulosa e tessuto vascolare umano, con particolare riferimento ad aorta ascendente e valvole tricuspidi. I risultati ottenuti, confrontati con software e metodi di analisi esistenti, dimostrano che l’approccio proposto è in grado di ricostruire e caratterizzare automaticamente reti fibrose tridimensionali complesse, fornendo una descrizione morfometrica completa dell’architettura del campione. Nel complesso, la metodologia sviluppata rappresenta uno strumento quantitativo e riproducibile per la valutazione morfometrica e topologica di scaffold biomimetici, tessuti biologici e costrutti ingegnerizzati, contribuendo a una migliore comprensione delle relazioni tra architettura fibrosa, funzione tissutale e processi rigenerativi.
Sviluppo di una pipeline image-based tridimensionale scale-adaptive per la caratterizzazione morfo-topologica di reti fibrose in tessuti nativi e scaffold ingegnerizzati
RICCHIARI, MARCO
2025/2026
Abstract
Understanding how the architecture of fibrous scaffolds influences cellular behavior and tissue development is a fundamental challenge in tissue engineering and regenerative medicine. The three-dimensional organization of the fibrous network play a key role in shaping the cellular microenvironment, regulating processes such as cell adhesion, migration, proliferation, extracellular matrix remodeling, and the mechanical behavior of engineered constructs. Consequently, structural descriptors including fiber diameter, spatial orientation, connectivity, and intersection density are essential for correlating scaffold architecture with its biological and biomechanical performance. Despite their importance, the quantitative characterization of complex fibrous networks remains challenging, especially when moving from planar images to volumetric datasets. Irregular fibers geometry, branching, overlapping structures, and imaging-related intensity variations complicate the reliable extraction of morpho-topological descriptors. Furthermore, most existing approaches are limited to two-dimensional analyses, providing only partial representation of the actual spatial organization of fibrous architecture. The aim of this thesis is to develop and validate an automated image-based framework for the three-dimensional morpho-topological characterization of fibrous scaffolds and biological tissues imaged by multiphoton microscopy. The proposed methodology extends an established two-dimensional graph-based approach to volumetric data, providing a quantitative and reproducible framework for network reconstruction and structural analysis. The workflow combines synthetic phantom generation for validation under controlled geometrical conditions with an analysis pipeline for experimental volumetric datasets. The computational framework includes images denoising, segmentation, three-dimensional skeletonization, graph reconstruction and automatic identification of nodes, branches, and intersections. These steps enable the extraction of quantitative descriptors such as fiber diameter, connectivity, real intersection density, spatial fiber orientation. After validation on synthetic phantoms, the pipeline was applied to volumetric datasets of electrospun polycaprolactone, fibroin, cellulose, and human cardiovascular tissues, including the ascending aorta and tricuspid valves. The results that the proposed approach enable the automatic reconstruction and characterization of complex three-dimensional fibrous networks, providing a quantitative description of sample architecture. In this perspective, this work offers a valuable tool for the objective assessment of biomimetic scaffolds, biological tissues, and engineered constructs, supporting future investigations into the relationship between fibrous architecture, tissue function, and regenerative processes.| File | Dimensione | Formato | |
|---|---|---|---|
|
Ricchiari_Marco.pdf
embargo fino al 13/07/2029
Dimensione
3.84 MB
Formato
Adobe PDF
|
3.84 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/110017