A5002508453- Radiomics and Machine Learning in Medical Imaging
- COVID-19 diagnosis using AI
- Lung Cancer Diagnosis and Treatment
- Advanced MRI Techniques and Applications
- MRI in cancer diagnosis
- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Mathematical Biology Tumor Growth
- Neural Networks and Applications
- Advanced Neuroimaging Techniques and Applications
- Advances in Oncology and Radiotherapy
- Opinion Dynamics and Social Influence
- Distributed Control Multi-Agent Systems
- Genetics and Neurodevelopmental Disorders
- Personality Traits and Psychology
- Personality Disorders and Psychopathology
- AI in cancer detection
- Anomaly Detection Techniques and Applications
- Markov Chains and Monte Carlo Methods
- Dementia and Cognitive Impairment Research
- Stochastic Gradient Optimization Techniques
- History and advancements in chemistry
- Infant Development and Preterm Care
- Genomics and Rare Diseases
- Single-cell and spatial transcriptomics
Università della Svizzera italiana
2024-2025
Istituto Universitario di Studi Superiori di Pavia
2025
Istituto Nazionale di Fisica Nucleare, Sezione di Pavia
2021-2024
Istituto Nazionale di Fisica Nucleare
2023-2024
University of Pavia
2021-2024
Background Neuropsychiatric symptoms (NPSs) are a distressful aspect of dementia and the knowledge structural correlates NPSs is limited. We aimed to identify associations fronto-limbic circuit with specific in patients various types cognitive impairment. Methods Of 84 participants, 27 were diagnosed mild impairment (MCI), 41 Alzheimer’s disease (AD) 16 non-AD dementia. In all we assessed regional brain morphometry using region interest (ROI)-based analysis. The mean cortical thickness (CT)...
Abstract Background Body composition scores allow for quantifying the volume and physical properties of specific tissues. However, their manual calculation is time-consuming prone to human error. This study aims develop validate CompositIA , an automated, open-source pipeline body from thoraco-abdominal computed tomography (CT) scans. Methods A retrospective dataset 205 contrast-enhanced CT examinations was used training, while 54 scans a publicly available were independent testing. Two...
In this article, we extend a recently introduced kinetic model for consensus-based segmentation of images. particular, will interpret the set pixels 2D image as an interacting particle system that evolves in time view consensus-type process obtained by interactions between and external noise. Thanks to formulation model, derive large solution model. We show parameters defining task can be chosen from plurality loss functions characterize evaluation metrics.
Abstract Objectives This study aimed to assess whether pharmacokinetic parameters derived from DCE-MRI can stratify Programmed Death-Ligand 1 (PD-L1) expression in NSCLC. The secondary aim was identify a suitable model configuration for anisotropic temporally-spaced sequences, considering Tofts variants, population-averaged arterial input functions (AIF), and bolus arrival time (BAT) estimation methods. Materials methods From April 2021 May 2023, patients with locally advanced non-small cell...
Abstract Background The role of computed tomography (CT) in the diagnosis and characterization coronavirus disease 2019 (COVID-19) pneumonia has been widely recognized. We evaluated performance a software for quantitative analysis chest CT, LungQuant system, by comparing its results with independent visual evaluations group 14 clinical experts. aim this work is to evaluate ability automated tool extract information from lung relevant design support model. Methods segments both lungs lesions...
Abstract Background Radiomics is a quantitative approach that allows the extraction of mineable data from medical images. Despite growing clinical interest, radiomics studies are affected by variability stemming analysis choices. We aimed to investigate agreement between two open-source software for both contrast-enhanced computed tomography (CT) and magnetic resonance imaging (MRI) lung cancers preliminarily evaluate existence radiomic features stable techniques. Methods Contrast-enhanced...
This study aims at exploiting artificial intelligence (AI) for the identification, segmentation and quantification of COVID-19 pulmonary lesions. The limited data availability annotation quality are relevant factors in training AI-methods. We investigated effects using multiple datasets, heterogeneously populated annotated according to different criteria.We developed an automated analysis pipeline, LungQuant system, based on a cascade two U-nets. first one (U-net[Formula: see text]) is...
Abstract We propose a deep learning (DL) model and hyperparameter optimization strategy to reconstruct T 1 2 maps acquired with the magnetic resonance fingerprinting (MRF) methodology. applied two different MRF sequence routines acquire images of ex vivo rat brain phantoms using 7‐T preclinical scanner. Subsequently, DL was trained experimental data, completely excluding use any theoretical MRI signal simulator. The best combination parameters implemented by an automatic strategy, whose key...
In the retrospective-prospective multi-center "Blue Sky Radiomics" study (NCT04364776), we plan to test a pre-defined radiomic signature in series of stage III unresectable NSCLC patients undergoing chemoradiotherapy and maintenance immunotherapy. As necessary preliminary step, explore influence different image-acquisition parameters on features' reproducibility apply methods for harmonization. We identified primary lung tumor two computed tomography (CT) each patient, acquired before after...
: Research is lacking about the development of personality disorders (PDs) from adolescence to early adulthood. This study aimed characterize profile high-risk adolescents compared with full-blown PDs and other psychiatric identify clinical markers that constitute a risk profile.
Background: The advent of next-generation sequencing (NGS) techniques in clinical practice led to a significant advance gene discovery. We aimed describe diagnostic yields “dynamic” exome-based approach cohort patients with epilepsy associated neurodevelopmental disorders. Methods: conducted retrospective, observational study on 72 probands. All underwent first level 135 panel, second 297 genes for inconclusive cases, and finally, whole-exome negative cases. Diagnostic at each step...
Abstract Computed tomography (CT) scans are used to evaluate the severity of lung involvement in patients affected by COVID-19 pneumonia. Here, we present an improved version LungQuant automatic segmentation software ( v 2), which implements a cascade three deep neural networks (DNNs) segment lungs and lesions associated with The first network (BB-net) defines bounding box enclosing lungs, second one (U-net $$_1$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msub>...
Abstract Preterm infants cannot counteract excessive reactive oxygen species (ROS) production due to preterm birth, leading an excess of lipid peroxidation with malondialdehyde (MDA) production, capable contributing brain damage. Melatonin (ME), endogenous hormone, and its metabolites, act as a free radical scavenger against ROS. Unfortunately, preterms have impaired antioxidant system, resulting in the inability produce release ME. This prospective, multicenter, parallel groups, randomized,...
We proposed a DL method and an automatic hyperparameters optimization strategy to reconstruct T1 T2 maps acquired with two Magnetic Resonance Fingerprinting (MRF) sequences. The model was trained validated on preclinical MRF dataset tested independent test set. Through lower number of images k-space sampling percentage than the standard post-processing, DL-based deliver parametric similar accuracy as dictionary-based methodology.
Magnetic Resonance (MR) parameters mapping in muscle Imaging (mMRI) is predominantly performed using pattern recognition-based algorithms, which are characterised by high computational costs and scalability issues the context of multi-parametric mapping. Deep Learning (DL) has been demonstrated to be a robust efficient method for rapid MR However, its application mMRI domain investigate Neuromuscular Disorders (NMDs) not yet explored. In addition, data-driven DL models suffered...
Trajectories can be regarded as time-series of coordinates, typically arising from motile objects. Methods for trajectory classification are particularly important to detect different movement patterns, while methods regression compute motility metrics and forecasting. Recent advances in computer vision have facilitated the processing trajectories rendered images via artificial neural networks with 2d convolutional layers (CNNs). This approach leverages capability CNNs learn spatial...
Studying the spatiotemporal dynamics of cells in living organisms is a current frontier bioimaging. Intravital Microscopy (IVM) provides direct, long-term observation cell behavior animals, from tissue to sub-cellular resolution. Hence, IVM has become crucial for studying complex biological processes motion and across scales, such as immune response pathogens cancer. However, data are typically kept private repositories inaccessible scientific community, hampering large-scale analysis that...
In this article we extend a recently introduced kinetic model for consensus-based segmentation of images. particular, will interpret the set pixels 2D image as an interacting particle system which evolves in time view consensus-type process obtained by interactions between and external noise. Thanks to formulation derive large solution model. We show that choice parameters defining task can be chosen from plurality loss functions characterising evaluation metrics.
Lung Computed Tomography (CT) is an imaging technique useful to assess the severity of COVID-19 infection in symptomatic patients and monitor its evolution over time. CT can be analysed with support deep learning methods for both aforementioned tasks. We have developed a U-net based algorithm segment lesions. Unfortunately, public datasets populated huge amount labelled scans affected by are not available. In this work, we first review all currently available scans, presenting extensive...
We developed a Neural Network (NN) for the reconstruction of T 1 and 2 parametric maps obtained with Magnetic Resonance Fingerprinting (MRF) technique. The training phase was realized on experimental inputs, eliminating use simulated datasets theoretical models. set optimal hyperparameters NN supervised algorithm were established through an optimization procedure. model achieved similar performances to traditional method, but number MRF images required lower respect dictionary-based method....