10081 research outputs found
Sort by
Whole Body MRI (WB-MRI) in patients affected by monoclonal plasma cell disorders
Get PDFPurpose
This PhD research evaluates the impact of multiparametric Whole Body-Magnetic Resonance Imaging (WB-MRI) in myeloma patients, comparing its diagnostic accuracy with 18F-FDG PET-CT and assessing its influence on clinical management. Additionally, it explores quantitative WB-MRI using radiomics.
Materials and Methods
Within the AccuMRI IRST protocol, 177 patients were enrolled (October 2020–January 2024), with 134 undergoing both WB-MRI and PET-CT within one month. Bone marrow involvement (BMI) detection was eveluated in multiple myeloma (MM) and high-risk smoldering multiple myeloma (HR-SMM). Inter-reader agreement among three radiologists was assessed in 52 patients, and WB-MRI acceptancy was evaluated in 134 patients.
Quantitative WB-MRI using radiomics was analyzed in 84 patients (45 MM, 39 HR-SMM), focusing on Apparent Diffusion Coefficient (ADC) and fat fraction (FF) sequences. Six volumes of interest (VOIs) were placed in pelvic bones and spine. Logistic regression with LASSO selection was used for feature analysis.
Results
WB-MRI showed 100% sensitivity and 97% specificity for BMI detection in MM, outperforming PET-CT (sensitivity 89%, p=0.02). Patients with BMI had higher blood paraprotein levels (p=0.007 in MM, p=0.01 in HR-SMM) and lower hemoglobin (p=0.002 in MM). WB-MRI findings influenced treatment in 97% of cases vs. 61% for PET-CT.
Inter-reader agreement was excellent (ICC > 0.9 for FF and ADC, except limbs: ICC=0.78). WB-MRI was preferred by 68.7% of patients, with preference influenced by age (p=0.011).
Radiomics models distinguished HR-SMM from MM (AUC=0.80 training, 0.70 test), with RS VOI showing highest accuracy (AUC=0.76). FF and ADC correlated with plasma cell percentage.
Conclusion
WB-MRI and PET-CT play key roles in evaluating myeloma patients. WB-MRI demonstrated superior sensitivity in detecting BMI and had a greater influence on therapeutic decision-making. MY-RADS criteria ensure reproducibility and radiomics could enhance risk stratification. WB-MRI is a reliable, patient-friendly imaging tool for myeloma
Photoelectrochemical biomass valorisation for renewable energy conversion
Get PDFNatural Photosynthesis has become a guiding model for developing sustainable solutions aiding in the transition from fossil fuels to renewable energy sources. For over a century, scientists have sought to mimic plants by creating devices capable of capturing energy and storing it in chemical bonds. Among the approaches explored to date, splitting water into molecular hydrogen and oxygen offers a promising breakthrough for clean energy production. However, the slow kinetics of the oxygen evolution reaction (OER) consistently hampers the efficiency of this process, limiting its competitiveness against widely available fossil-based technologies. To overcome this limitation, the focus of the scientific community is shifting towards alternative oxidation reactions, characterised by lower energy requirements and inexpensive starting compounds. In this thesis, the photoelectrochemical conversion of biomass derivatives to useful chemicals is investigated. The following anodic reactions were explored: i) Titanium doped hematite (Ti:Fe2O3) photoanodes modified with cobalt- or nickel-based co-catalysts, for the conversion of 5-hydroxymethylfurfural (HMF) into 2,5- furan dicarboxylic acid (FDCA); ii) bismuth vanadate (BiVO4) photoanodes for glycerol oxidation reaction (GOR). Overall, a comprehensive understanding of the optimal conditions for both reaction and photoanode stability was crucial for maximizing process performance. This knowledge can also pave the way for successfully coupling valuable cathodic reaction, such as the hydrogen evolution reaction (HER) or CO2 reduction, thereby greatly enhancing the overall functionality and effectiveness of the PEC device
Nutritional and health care aspects of several flours with different rheological proprieties
Get PDFThis study conducted a three-year field trial to evaluate the composition of gluten proteins in wheat and cell proliferation before and after partial digestion. First year, free polyphenols in 18 ancient durum wheat flours showed higher FPC, TPC, and flavonoid levels, enhancing DPPH scavenging compared to subsequent seasons. Partial digestion resulted in high-molecular-weight (HMW) proteins breaking down into smaller peptides, with significant overall protein reductions, varying among gluten proteins and subunits. The gliadins to glutenins ratio remained high with minor fluctuations.
Second year, bound polyphenols and flavonoids decreased, leading to lower antioxidant activity and DPPH scavenging ability. The protein profile of 16 ancient durum wheat types demonstrated minimal changes from the previous year, though overall protein levels declined without significantly impacting gluten and HMW-GS ratios. An increase in ω-gliadin coincided with a decrease in LMW-GS, influencing the gliadins/glutenins and HMW-GS/LMW-GS ratios. Following partial digestion, the second year's wheat varieties presented reduced HMW-GS and LMW-GS bands, contrasting with the first year's results. The effects of partial digestion on gliadins/glutenins and HMW-GS/LMW-GS ratios varied, often showing a decline.
Third year, polyphenol and flavonoid levels in wheat flour remained stable, with polyphenols primarily combined and a slight increase in antioxidant activity. The protein profiles of 22 ancient wheat samples varied, mainly displaying low molecular weight glutenin subunits (LMW-GS). Notable changes in gluten protein composition were observed, including a decrease in high molecular weight glutenin subunits (HMW-GS) and ω-gliadins or reduced LMW-GS, while protein degradation post-digestion persisted. The total protein content decreased from the first to the third year but was higher than in the second year. Post-digestion protein levels fell, yet ancient wheat varieties consistently showed higher protein content than the modern variety Claudio. Ratios of gluten and its subunits, indicative of dough rheological properties (e.g., gliadin/glutenin and HMW-GS/LMW-GS), typically matched historical ranges
Essential oils as a source of natural compounds for the modulation of intestinal physiology
Get PDFEssential oils and their components have demonstrated significant therapeutic properties, including anti-inflammatory, antioxidant, antibacterial, antifungal, antiviral, and antitumor activities. This thesis presents a series of studies conducted during the PhD, aimed at exploring the potential applications of various essential oil components in human health, with a particular focus on intestinal health. The studies address topics such as the properties related to irritable bowel syndrome, the effects on obesity, the antifungal activity against clinical strains of Candida spp., and the antitumor activities against colorectal cancer and glioblastoma. The experimental approach included both in vitro and in vivo models to assess the mechanisms of action and therapeutic efficacy, with special attention to microbiota modulation and molecular analysis of the active components. The overall objective was to deepen the understanding of the therapeutic potential of these natural substances by evaluating their mechanisms of action and efficacy in combination with other natural compounds, given the limited literature available on the subject
Deep reinforcement learning and creativity
Get PDFGenerative artificial intelligence (AI) is among the most exciting developments in computer science over the last decade. In several fields, it is not only complementing but also replacing the creative abilities that were once solely in humans’ hands. However, current generative models are limited by their learning schemes, which merely aim to imitate training data.
To develop more creativity-oriented models, new approaches should be considered. Among them, reinforcement learning (RL) represents a promising direction. RL is an inherently learning-by-acting approach and can capture a greater variety of target behaviors, making it ideal for modeling how humans learn to behave creatively.
Studying RL together with creativity can be of crucial importance for both fields. This thesis explores whether creativity can enhance the design of RL algorithms and, vice versa, whether RL can help develop more creative generative models.
First, we study if dreaming can help RL agents better generalize, as suggested for humans. We leverage generative augmentations to transform predicted trajectories into dream-like experiences for training agents and evaluate generalization capabilities in different scenarios.
Then, we develop a new creativity score that quantifies the originality and value of artifacts. We use it as a reward in an RL framework, and we propose to fine-tune pre-trained models toward more creative solutions. We validate our method in two different domains: poetry generation and problem solving. In addition, we present new sampling schemes to better simulate the human creative process by working at the response generation and validation levels.
Finally, we conclude with a deep analysis of three main social and practical issues: whether current models are creative and their implications; whether they can be entitled to agency and what happens to human agency when collaborating with them; how copyright laws can manage the complexity of generative AI to protect human- and machine-generated artworks
Applications of delay differential equations to the physics of complex systems
Get PDFIn the present Thesis we introduce Delay Differential Equations as an approach to model emergent states arising from the interaction of different time scales in Complex Systems. We devote our attention to models related to neuronal phenomena and epidemiological forecasting. Our work on neural models consists in demonstrating how the interplay of neural spiking and refractoriness time scales, and cyclic structures in a directed interaction network can give rise to self-sustained traveling waves. We also build a description of the bifurcation phenomenon and the dynamical steady state in terms of a single Delay Differential Equation, the solution of which can be interpolated to obtain single node trajectories. We subsequently formulate a normal form model to explain how the interplay of a delayed feedback and the time scale of the system can stabilize an orbit near privileged locations in phase space. Due to the interest in neural computation and information processing we also construct a discrete stochastic model that preserves the activation statistics of a noisy nonlinear neuron, focusing on the different time scales involved in the process. The results on epidemiological modeling are concerned with the usage of Distributed Delay Differential Equations in the forecast of epidemic events on the short and medium term on a metropolitan scale in Bologna. We show that these models can cover for the shortcomings of models based on Ordinary Differential Equations at the considered scales, despite the more complicated nature of Delayed Equations. In particular we show that road traffic data can be used as a proxy for the rate of contacts in a population, during periods when other factors are approximately unchanging. We also set a basis for quantitative predictivity analysis, by obtaining linear response laws for the model and in particular for the variables that could be used in regression tasks
Constructing thought spontaneously and deliberately: neural bases and adaptive strategies
Get PDFHuman thought spans a continuum from spontaneous to deliberate forms. Spontaneous thought occurs when attention shifts from ongoing activities to task-unrelated content (mind-wandering), whereas deliberate thought involves voluntarily imagining future events (mental time travel; MTT) or setting goals (prospective memory; PM). Both forms of thought engage multiple brain networks – notably, the Default Mode Network (DMN) with key hubs in the ventromedial prefrontal cortex (vmPFC) and hippocampus – while external attention is primarily mediated by the Dorsal and Ventral Attention Networks. However, there is still no consensus on the neural and cognitive mechanisms underlying spontaneous and deliberate thought, their relationship with external attention, and optimal strategies for promoting adaptive thinking. Chapter 1 explores the neural bases of mind-wandering and external attention, adopting a tDCS protocol that assesses both simultaneously. Findings reveal that the posterior parietal cortex mediates both internal and external attention, while the vmPFC uniquely supports future-oriented mind-wandering. Chapter 2 examines whether activating mental representations (self-schema and the future time) promotes adaptive mind-wandering. Results show that activating representations of future time increases aware mind-wandering, whereas activating self-schema promotes unaware mind-wandering. Chapter 3 investigates the role of the vmPFC in deliberate thought (MTT), examining the temporal unfolding of memory structures (“think aloud” method) during event construction. Patients with vmPFC damage show degraded personal semantic information and atypical, “backward” transitions from lower- to higher-level memory structures, reporting fewer specific events. Finally, Chapter 4 explores the simultaneous effect of intention offloading (use of external reminders) on both spontaneous (mind-wandering) and deliberate (PM) thought. Results indicate that offloading reduces pupil diameter, especially under high memory load, and decreases off-task thinking over time. Collectively, this work clarifies the cognitive, neural, and physiological mechanisms underlying internal (spontaneous and deliberate thought) and external attention, identifying strategies to foster adaptive thinking
More-than-music. Echosystems, acoustemologies and histories of listening from Sápmi
Get PDFThe challenge of defining ‘music’ and ‘sound’ across cultures has been a persistent concern in ethnomusicological and anthropological research from the outset of these fields. Top-down academic applications of these concepts often remain unquestioned, affecting the rich plurality of local classifications for diverse sounding practices alongside their associated bodies of knowledge. This dissertation critically examines the relevance of these overarching categories in the context of Sámi acoustemologies, proposing a novel theoretical paradigm derived from and informed by Indigenous sound ontologies: ‘more-than-music’. ‘More-than-music’ seeks to challenge the ethno-anthropocentric characterizations of sonic relationships across societies and environments which emerge from academic discourses. Its bottom-up nature underlines the necessity to acknowledge the complexity of local onto-epistemologies and the agencies of human and other-than-human beings in academic research practices. To measure the validity and applicability of the ‘more-than-music’ paradigm, this study places emphasis on juoiggus as a Sámi more-than-musical expression and biocultural heritage bridging human performativity and aesthetics with other-than-human voices. The analysis is guided by the research questions: How can juoiggus be explained as something more-than-music? What can we learn from the sonic relationships between individuals, communities and place through juoiggus? How have these interconnections been altered by the transformation of landscapes and traditional knowledge resulted from settler colonialism and environmental crises? While this dissertation builds on mixed methods at the convergence of ethnography and ecology, the methodological choices have been tailored to align with decolonizing research practices appropriate for a thesis developed in collaboration with Indigenous peoples and centered on Indigenous Knowledge. The manuscript is accompanied by a collaborative audio-anthology that serves both as a multimedia output and a means of returning the collected knowledge and field-recordings. The audio-anthology weaves together a plurality of listening experiences and soundscape compositions from Sápmi, integrating the thesis’ contents while delving into the notion of ‘more-than-music’
Identification of outcome-oriented cut-offs for copy number alterations in multiple myeloma: predictive biomarkers with improved prognostic accuracy
No full textAim of the present study was to develop a statistical approach to define the best cut-off Copy number alterations (CNAs) calling from genomic data provided by high throughput experiments, able to predict a specific clinical end-point (early relapse, 18 months) in the context of Multiple Myeloma (MM).
743 newly diagnosed MM patients with SNPs array-derived genomic and clinical data were included in the study.
CNAs were called both by a conventional (classic, CL) and an outcome-oriented (OO) method, and Progression Free Survival (PFS) hazard ratios of CNAs called by the two approaches were compared.
The OO approach successfully identified patients at higher risk of relapse and the univariate survival analysis showed stronger prognostic effects for OO-defined high-risk alterations, as compared to that defined by CL approach, statistically significant for 12 CNAs.
Overall, 155/743 patients relapsed within 18 months from the therapy start. A small number of OO-defined CNAs were significantly recurrent in early-relapsed patients (ER-CNAs) - amp1q, amp2p, del2p, del12p, del17p, del19p -. Two groups of patients were identified either carrying or not ≥1 ER-CNAs (249 vs. 494, respectively), the first one with significantly shorter PFS and overall survivals (OS) (PFS HR 2.15, p<0001; OS HR 2.37, p<0.0001). The risk of relapse defined by the presence of ≥1 ER-CNAs was independent from those conferred both by R-IIS 3 (HR=1.51; p=0.01) and by low quality (< stable disease) clinical response (HR=2.59 p=0.004). Notably, the type of induction therapy was not descriptive, suggesting that ER is strongly related to patients’ baseline genomic architecture.
In conclusion, the OO- approach employed allowed to define CNAs-specific dynamic clonality cut-offs, improving the CNAs calls’ accuracy to identify MM patients with the highest probability to ER. As being outcome-dependent, the OO-approach is dynamic and might be adjusted according to the selected outcome variable of interest
Development of mechanically reinforced porous ceramics for bone regeneration
No full textHydroxyapatite (HA) is widely recognized as suitable material to develop scaffolds for bone regeneration, mainly due to optimal chemical mimicry with the natural inorganic component of bones. Despite such a great potential, a major drawback associated to hydroxyapatite, typical of all ceramic materials, refers to its intrinsic brittleness, posing significant limitation to their clinical use mainly due to the risk of irreparable failure of the scaffolds. This is a major reason making the regeneration of load-bearing bone defects such as maxillofacial regions and spine region, a still unmet clinical need of great socio-economic relevance. Among the various approaches for apatite reinforcement, the use of carbon fibers is an interesting strategy due to their inherent biocompatibility, high strength-to-weight ratio, thermophysical properties, sorption, and high elastic modulus. They can bridge cracks and arrest their propagation, thus raising the toughness and helping to prevent the ceramic scaffold from sudden failure. In this respect, my Ph.D. project is focused on the development of new processes to achieve calcium phosphate-based scaffolds featuring enhanced mechanical properties. The first part of my PhD thesis involves the development of mechanically reinforced porous apatitic ceramic scaffolds obtained by 3D forming of hydroxyapatite powders and sintering. The second part of my PhD activity involves the development of porous apatitic bone cements obtained by isothermal reaction of metastable calcium phosphate precursors at body temperature, mechanically reinforced with carbon fibers. This second part includes three sub-chapters dealing with: i) apatitic cements obtained from inorganic precursors synthesized by low and high-temperature treatments, ii) magnesium-doped apatitic cements with intrinsic antibacterial properties and iii) reinforcement of magnesium and strontium-doped apatitic cements with carbon fibers