Trattamento radioterapico ipofrazionato ad alte dosi del carcinoma prostatico con tecnica ad intensità modulata (IMRT) guidata dalle immagini (IGRT)

Obiettivo: valutare la tossicità ed il controllo di malattia di un trattamento radioterapico ipofrazionato ad alte dosi con tecnica ad intensità modulata (IMRT) guidata dalle immagini (IGRT) in pazienti affetti da carcinoma prostatico a rischio intermedio, alto ed altissimo di recidiva. Materiali e metodi: tutti i pazienti candidati al trattamento sono stati stadiati e sottoposti al posizionamento di tre “markers” fiduciali intraprostatici necessari per l’IGRT. Mediante tecnica SIB – IMRT sono stati erogati alla prostata 67,50 Gy in 25 frazioni (EQD2 = 81 Gy), alle vescichette 56,25 Gy in 25 frazioni (EQD2 = 60,35 Gy) e ai linfonodi pelvici, qualora irradiati, 50 Gy in 25 frazioni. La tossicità gastrointestinale (GI) e genitourinaria (GU) è stata valutata mediante i CTCAE v. 4.03. Per individuare una possibile correlazione tra i potenziali fattori di rischio e la tossicità registrata è stato utilizzato il test esatto di Fisher e la sopravvivenza libera da malattia è stata calcolata mediante il metodo di Kaplan-Meier. Risultati: sono stati arruolati 71 pazienti. Il follow up medio è di 19 mesi (3-35 mesi). Nessun paziente ha dovuto interrompere il trattamento per la tossicità acuta. Il 14% dei pazienti (10 casi) ha presentato una tossicità acuta GI G ≥ 2 e il 15% (11 pazienti) ha riportato una tossicità acuta GU G2. Per quanto riguarda la tossicità tardiva GI e GU G ≥ 2, essa è stata documentata, rispettivamente, nel 14% dei casi (9 pazienti) e nell’11% (7 pazienti). Non è stata riscontrata nessuna tossicità, acuta o cronica, G4. Nessun fattore di rischio correlava con la tossicità. La sopravvivenza libera da malattia a 2 anni è del 94%. Conclusioni: il trattamento radioterapico ipofrazionato ad alte dosi con IMRT-IGRT appare essere sicuro ed efficace. Sono comunque necessari ulteriori studi per confermare questi dati ed i presupposti radiobiologici dell’ipofrazionamento del carcinoma prostatico.

Hypofractionated stereotactic radiotherapy after hyperbaric oxygenation for recurrent high-grade glioma: a pilot study

Purpose The presence of hypoxic cells in high-grade glioma (HGG) is one of the main reasons of local failure after radiotherapy (RT). The use of hyperbaric oxygen therapy (HBO) could help to overcome the problem of hypoxia in poorly oxygenated regions of the tumor. We performed a pilot study to evaluate the efficacy of hypofractionated image-guided helical TomoTherapy (HT) after HBO in the treatment of recurrent HGG (rHGG). Methods We enrolled 15 patients (aged >18 years) with diagnosis of rHGG. A total dose of 15-25 Gy was administered in daily 5-Gy fractions for 3-5 consecutive days after daily HBO. Each fraction was delivered up to maximum of 60 minutes after HBO. Results Median follow-up from HBO-RT was 28.6 (range: 5.3-56.8). No patient was lost to follow-up. Median progression-free survival (mPFS) for all patients was 3.2 months (95% CI: 1.34- 6.4 ), while 3-month, 6-month and 12 month PFS was 60% (95%CI: 31.8.4-79.7), 40% (95%CI: 16.5- 62.8) and10.0 (0.8-33.5) , respectively. Median overall survival (mOS) of HBO-RT was 11.7 months (95% CI: 7.3-29.3), while 3-month, 6-month and 12 month OS was 100% , 93.3% (61.3-99.0) and 46.7 % (21.2-68.8). No acute or late neurologic toxicity >grade 2 (CTCAE version 4.3) was observed in 86.66% of patients. Two patients developed G3 Radionecrosis. Conclusion HSRT combined to HBO seems effective and safe in the treatment of rHGG. One of advantages of HBO-RT is the reduced overall treatment time (3-5 consecutive days).

Exploring domain-informed and physics-guided learning in image-to-image translation

Image-to-image (i2i) translation networks can generate fake images beneficial for many applications in augmented reality, computer graphics, and robotics. However, they require large scale datasets and high contextual understanding to be trained correctly. In this thesis, we propose strategies for solving these problems, improving performances of i2i translation networks by using domain- or physics-related priors. The thesis is divided into two parts. In Part I, we exploit human abstraction capabilities to identify existing relationships in images, thus defining domains that can be leveraged to improve data usage efficiency. We use additional domain-related information to train networks on web-crawled data, hallucinate scenarios unseen during training, and perform few-shot learning. In Part II, we instead rely on physics priors. First, we combine realistic physics-based rendering with generative networks to boost outputs realism and controllability. Then, we exploit naive physical guidance to drive a manifold reorganization, which allowed generating continuous conditions such as timelapses.

Mind the gaps: cognitive-inspired AI for high-level visual sensemaking. Towards abstract concept image classification

The abundance of visual data and the push for robust AI are driving the need for automated visual sensemaking. Computer Vision (CV) faces growing demand for models that can discern not only what images "represent," but also what they "evoke." This is a demand for tools mimicking human perception at a high semantic level, categorizing images based on concepts like freedom, danger, or safety. However, automating this process is challenging due to entropy, scarcity, subjectivity, and ethical considerations. These challenges not only impact performance but also underscore the critical need for interoperability. This dissertation focuses on abstract concept-based (AC) image classification, guided by three technical principles: situated grounding, performance enhancement, and interpretability. We introduce ART-stract, a novel dataset of cultural images annotated with ACs, serving as the foundation for a series of experiments across four key domains: assessing the effectiveness of the end-to-end DL paradigm, exploring cognitive-inspired semantic intermediaries, incorporating cultural and commonsense aspects, and neuro-symbolic integration of sensory-perceptual data with cognitive-based knowledge. Our results demonstrate that integrating CV approaches with semantic technologies yields methods that surpass the current state of the art in AC image classification, outperforming the end-to-end deep vision paradigm. The results emphasize the role semantic technologies can play in developing both effective and interpretable systems, through the capturing, situating, and reasoning over knowledge related to visual data. Furthermore, this dissertation explores the complex interplay between technical and socio-technical factors. By merging technical expertise with an understanding of human and societal aspects, we advocate for responsible labeling and training practices in visual media. These insights and techniques not only advance efforts in CV and explainable artificial intelligence but also propel us toward an era of AI development that harmonizes technical prowess with deep awareness of its human and societal implications.