Ecografia con mezzo di contrasto dalla ricerca alla pratica clinica: diagnosi di carcinoma epatocellulare con sistema CEUS LI-RADS e con software di quantificazione della perfusione tissutale

Il carcinoma epatocellulare (HCC) rappresenta il tumore epatico primitivo più comune con una incidenza fino all’85%. È uno dei tumori più frequenti al mondo ed è noto per l’elevata letalità soprattutto in stadio avanzato. La diagnosi precoce attraverso la sorveglianza ecografica è necessaria per migliorare la sopravvivenza dei pazienti a rischio. Il mezzo di contrasto ecografico migliora la sensibilità e la specificità diagnostica dell’ecografia convenzionale. L’ecografia con mezzo di contrasto (contrast-enhanced ultrasound, CEUS) è pertanto considerata una metodica valida per la diagnosi di HCC a livello globale per la sua ottima specificità anche a fronte di una sensibilità subottimale. L’aspetto contrastografico delle lesioni focali epatiche ha portato un team di esperti allo sviluppo del sistema Liver Imaging Reporting and Data System (LI-RADS) con l’obiettivo di standardizzare la raccolta dati e la refertazione delle metodiche di imaging per la diagnosi di HCC. La CEUS è una metodica operatore-dipendente e le discordanze diagnostiche con gli imaging panoramici lasciano spazio a nuove tecniche (Dynamic Contrast Enhanced UltraSound, DCE-US) volte a migliorare l’accuratezza diagnostica della metodica e in particolare la sensibilità. Un software di quantificazione della perfusione tissutale potrebbe essere di aiuto nella pratica clinica per individuare il wash-out non visibile anche all’occhio dell’operatore più esperto. Il nostro studio ha due obiettivi: 1) validare il sistema CEUS LI-RADS nella diagnosi di carcinoma epatocellulare in pazienti ad alto rischio di HCC usando come gold-standard l’istologia quando disponibile oppure metodiche di imaging radiologico accettate da tutte le linee guida (tomografia computerizzata o risonanza magnetica con aspetto tipico) eseguite entro quattro settimane dalla CEUS; 2) valutare l’efficacia di un software di quantificazione della perfusione tissutale nel riscontro di wash-out per la diagnosi di HCC in CEUS.

Sensor networks optimization and software development for Structural Health Monitoring based on ultrasonic guided waves

The Structural Health Monitoring (SHM) research area is increasingly investigated due to its high potential in reducing the maintenance costs and in ensuring the systems safety in several industrial application fields. A growing demand of new SHM systems, permanently embedded into the structures, for savings in weight and cabling, comes from the aeronautical and aerospace application fields. As consequence, the embedded electronic devices are to be wirelessly connected and battery powered. As result, a low power consumption is requested. At the same time, high performance in defects or impacts detection and localization are to be ensured to assess the structural integrity. To achieve these goals, the design paradigms can be changed together with the associate signal processing. The present thesis proposes design strategies and unconventional solutions, suitable both for real-time monitoring and periodic inspections, relying on piezo-transducers and Ultrasonic Guided Waves. In the first context, arrays of closely located sensors were designed, according to appropriate optimality criteria, by exploiting sensors re-shaping and optimal positioning, to achieve improved damages/impacts localisation performance in noisy environments. An additional sensor re-shaping procedure was developed to tackle another well-known issue which arises in realistic scenario, namely the reverberation. A novel sensor, able to filter undesired mechanical boundaries reflections, was validated via simulations based on the Green's functions formalism and FEM. In the active SHM context, a novel design methodology was used to develop a single transducer, called Spectrum-Scanning Acoustic Transducer, to actively inspect a structure. It can estimate the number of defects and their distances with an accuracy of 2[cm]. It can also estimate the damage angular coordinate with an equivalent mainlobe aperture of 8[deg], when a 24[cm] radial gap between two defects is ensured. A suitable signal processing was developed in order to limit the computational cost, allowing its use with embedded electronic devices.

Real-time anomaly detection of gamma-ray bursts for the Cherenkov Telescope Array using deep learning

The Cherenkov Telescope Array (CTA) will be the next-generation ground-based observatory to study the universe in the very-high-energy domain. The observatory will rely on a Science Alert Generation (SAG) system to analyze the real-time data from the telescopes and generate science alerts. The SAG system will play a crucial role in the search and follow-up of transients from external alerts, enabling multi-wavelength and multi-messenger collaborations. It will maximize the potential for the detection of the rarest phenomena, such as gamma-ray bursts (GRBs), which are the science case for this study. This study presents an anomaly detection method based on deep learning for detecting gamma-ray burst events in real-time. The performance of the proposed method is evaluated and compared against the Li&Ma standard technique in two use cases of serendipitous discoveries and follow-up observations, using short exposure times. The method shows promising results in detecting GRBs and is flexible enough to allow real-time search for transient events on multiple time scales. The method does not assume background nor source models and doe not require a minimum number of photon counts to perform analysis, making it well-suited for real-time analysis. Future improvements involve further tests, relaxing some of the assumptions made in this study as well as post-trials correction of the detection significance. Moreover, the ability to detect other transient classes in different scenarios must be investigated for completeness. The system can be integrated within the SAG system of CTA and deployed on the onsite computing clusters. This would provide valuable insights into the method's performance in a real-world setting and be another valuable tool for discovering new transient events in real-time. Overall, this study makes a significant contribution to the field of astrophysics by demonstrating the effectiveness of deep learning-based anomaly detection techniques for real-time source detection in gamma-ray astronomy.