Using artificial intelligence (AI) offers opportunities to analyze medical images and to support early cancer detection. For instance, neural networks, in different implementation, can be used to analyze data image parts (i.e., voxels), defining a trained network useful for lung cancer nodules detection. We present our experience in designing and testing a transformerbased deep learning architecture, aiming to detect pulmonary cancer nodule candidates using 3D Computed Tomography (CT) images. The module also includes a preprocessing pipeline based on dynamic sampling of voxels extracted from images, to support data filtering and results explainability. The proposed architecture has been implemented, trained, and tested using the LUNA16 publicly available dataset. Experimental results proved both high effectiveness and competitive performance metrics across standard evaluations. Trained module can be used on a large CT dataset aiming to support clinicians in lung cancer early detection as well as to support in followup for lung cancer patients treatments. This work represent, indeed, results for preliminary applications in a research project (Advancing Lung Cancer Screening: Artificial Intelligence, Multimodal Imaging and Cutting-Edge Technologies for Early Detection and Characterization), conducted in collaboration with San Raffaele Hospital (Italy), Campus Biomedico University (Italy) and University Hospital of Salerno.

Sarcopenia is a clinical condition marked by loss of muscle mass and strength, leading to reduced mobility and quality of life. Accurate identification and quantification of muscle mass are essential for the timely diagnosis and treatment of sarcopenia. To calculate muscle volumes and sarcopenia indexes, segmentation techniques are required on CT images, helping to adjust treatments for chronic diseases. However, evaluating muscle volumes and thus determining sarcopenia indexes currently is highly dependent on manual image segmentation by human operators. We propose a deep learning architecture for automatic muscle mass segmentation, integrating DeepLabv3+ and U-Net3+ models for 2D and 3D segmentation, respectively. These models have been tested on available datasets and combined using an ensemble learning approach to enhance predictive accuracy. This proposed architecture can be integrated into clinical workflows for the assessment of sarcopenia, increasing the reliability and efficiency of image-based diagnoses.

There is increasing evidence that many molecular processes exhibit differences with age and sex. Such differences produce also differences in the insurgence and progression of many complex diseases. For instance, demographic data on the insurgence of comorbidities of mellitus diabetes, on the lethality of COVID-19, and on some cancers shows differences between sex and age groups. Therefore, the growing interest in such areas requires the management of related data as well as the development of algorithms and tools for the analysis. The availability of omics data annotated with metadata related to age and sex is mandatory for building the analysis pipeline. The number of databases containing data related to age and sex is henceforth growing. We here show some databases and tools storing such data. Finally, future research directions are highlighted.

Healthcare access (HA) is a multi-dimensional concept that includes health services availability and accessibility for the populations. These services should be determined by population healthcare needs, especially for vulnerable populations. Digital healthcare service became more important to facilitate the access to medical care by vulnerable people and by citizens in general. Digital health and technologies have provided many online e-services to address social and healthcare services.In this contribution, we propose the implementation of an innovative platform to support and manage the access to health and social services for vulnerable people. Two different use cases have been proposed to demonstrate the application of this platform to different healthcare contexts. The results shows the benefits of using the platform in terms of request management times and reduction of hospitalization.

The quality and traceability of agricoltural and food products (indicated as agri-food) represents an important task for industries to focus on environments and wellness targets. In the context of milk and vegetable production processes, it is no possible to monitor and control animals behaviour, environmental conditions, and overall quality affecting these productions. Accurate and explainable predictions of quantities, as well as food properties qualities, is relevant for marketing and planning action in agri-food companies thus to in obtaining more efficient higher-quality productions and contribute to citizens wellness.We here report examples and experiences of machine learning algorithms application to evaluate and predict quantity and frequency of production in an large south of Italy farm. Data are extracted from a tracking system storing all production phases, i.e.: (i) from fruits plants to storage, cold maintaining and transportation, and (ii) cows management, fresh milk analysis and packaging. The here proposed experience contributes to evaluate and predict quantity and frequency of production, aiming to support farms in product planning and production phases.

Magnetic Resonance Images (MRI) allow to extract meaningful structural information. Machine learning and neural network based algorithms are used to analyze such images, to extract features and to identify anomalies related to diseases. To perform anomaly detection tasks in MR images of the human brain, we propose the use of the Variational AutoEncoder (VAE) method. A VAE is a deep-learning method able to compress and reconstruct the original image through well-defined functions aiming to extract only significant features that are used to identify abnormal pattern. In this contribution, we present a tool based on VAE method for the identification and annotation of brain lesions in MRI aiming to support physicians in the detection of anomalies. Moreover, a MongoDB database is also used to store the data and manage the annotations.

Brain network analysis represents an effective tool useful to explore the connectivity patterns to uncover related features and phenomena concerning different brain functions and diseases, such as the epilepsy. According to the clinical-psychological and neurological studies, signal patterns for network analysis mainly include data coming from functional magnetic resonance imaging (fMRI), positron emission tomography (PET), magnetoencephalogram (MEG), and electroencephalogram (EEG). High spatial-resolution fMRI and high temporal-resolution EEG are non-invasive and common techniques. EEG is quite low cost and, in many cases, EEG signal can provide important physiological information about the activity of the human brain which can be used to identify and detect epilepsy. In this paper we propose a framework for graph analysis of EEG functional connectivity networks, using EEG of epileptic subjects to build complex networks. Our aim is to provide a method to identify epilepsy in the early stages supporting physician in performing appropriate measures to prevent accidental consequences and to ensure health patient.

Medical image fusion allows to combine multiple images from single or multiple imaging modalities to improve the imaging quality. It supports physicians in performing more accurate clinical decisions and analysis based on medical images. Hybrid imaging with FDG PET/CT is became a standard in the day-to-day clinical practice for management of cancer because it combines the subtle biologic changes detected by FDG PET with the anatomic information offered by CT scans.In this context, we consider hybrid imaging with FDG PET/CT to propose a methodology aiming to use the fusion CT and PET images to evaluate drug administration in patients undergoing treatment.

Quantitative analysis of PET images is a clinical common practice. It is used to estimate the input function of 18F-FDG tracer in order to study a physiological process and to evaluate the response to a treatment. It allows the evaluation of coronary artery pathologies, as well as metabolic syndrome associated to cardiovascular diseases. We propose a method for analyzing the dynamic PET cardiac images aiming to assess the progress of glucose metabolism on large vessels as the aorta one. The aim is to study the relation among drug dosage with metabolic syndrome. Indeed, the aim is to correlate the glucose metabolism values (specifically MRGlu - Glucose Metabolic Rate) quantified in the aorta and in the left ventricle, by using PET dynamic images. The measures are presented and proposed for clinical drug validations.

Patient medical records contain several types of data, such as images, signals, or textual data. The integration of such data on a single system provides the possibility to select the clinical data of interest and then to choose the information extraction operation to be performed on such data. Formulating diagnoses of complex diseases is a challenging task, which is often the key to the precise identification of the correct therapies. Hence, a uniforming environment for data clinical staging, in which physicians can perform data annotations and images manipulation could be of great help in order to convey relevant information forming the clinical summary of a patient with great precision and detail. In this work we present a semi-automatic tool for clinical data annotation aiming to be a decision support system.

Parotid, sub-mandibular and several glands are salivary glands and they are responsible of saliva secretion. Moreover, they can present diseases related to neoplastic or inflammatory lesions. The parotid is the biggest salivary gland and it is very important because it is often the site of neoplastic formations. The identification and the diagnosis of salivary disorders can be performed by physiological inspection correlated with the analysis of clinical data and images. In this context, diagnostic imaging is able to support early detection, helping physicians in clinical examination and diagnosis. Here, we present a semi-automatic tool for the annotation and integration of clinical images aiming to be a decision support system and to train automatic lesion classifiers.

The salivary glands are exocrine glands aiming to make, modify and secrete saliva into the oral cavity. These glands can present diseases related to inflammatory or neoplastic lesions. Generally, diagnosis of salivary gland disorders is performed by physiological inspection supported by clinical data and images analysis. We focus on a methodology to annotate and integrate clinical information and bioimages to support physicians in disease examination and diagnosis.

Physical therapy and rehabilitation therapy aim to support patients in dealing with the consequences of their and physical impairments in daily activities. The recent developments in biomedical sensors combined with the wireless network infrastructure will deeply transform healthcare systems and help physicians in designing better and more precise therapies with faster feedback from patients in terms of health-related measured data. Furthermore, these new systems will enable distributed healthcare services for remote patients who may live far away from health structures or who may have movement impairment. Efficiently monitoring or acquiring data from a large number of patients will cause improvements during rehabilitation and help in early diagnoses together with reducing the costs in the healthcare system with more effective prevention. We present a programmable rehabilitation device which can be useful to evaluate patients' performances in a set of physiotherapy exercises designed to evaluate subjects by neurophysiological impairments which slow down some types of movements. The tool is able to support the definition of rehabilitation exercises involving upper limbs and hands. The presented device assesses the responsiveness and movement capacity of patients undergoing physiotherapy aiming to test and measure the mobility, strength and functional ability of the hand during prone supination exercises.

Epidemiological models are relevant to study and analyze clinical as well as environmental and behavioural data, useful to support health studies. The target is to perform epidemiological analysis producing fast and reliable data access useful to guide prevention and curing processes. This is currently true in pandemic emergency as the current Covid-19 context. Epidemiological models should support in the early identification of pandemic phenomena and in making available data set for studying more accurate drug-based strategy for vaccines or virus containment.In this contribution we present an epidemiology database which integrates different types of clinical data to support research, follow-up and patient monitoring. The idea starts from an hospital databases cooperation integration where virus available data have been integrated to support statistical based studies. Starting from an available database containing 5 years data of infection related viruses (such as HPC, hepatitis) and patient anonymous data, the proposed system provide an integrated data access able to (i) extracting data filtered by means of clinical hypothesis based on patient profiles, environment and drugs and (ii) allowing to build large scale geographical data mappings in order to study correlations among chronic infection diseases and their relations with upcoming pandemic phenomena. Even if the application is in its infancy, the application is relevant with high very important applications.

Prevention is essential to counteract the onset of cancer. The analysis of clinical and environmental data as well as their integration are basic topics to prevent chronic diseases, especially neoplasms, and to identify the correlations between cancer and environmental factors. The proposed contribution aims to acquire, analyze and integrate clinical and geographical data to evaluate their possible correlations. A GIS application is here reported to correlate TSH (Thyroid-Stimulating Hormone) values with environmental data as case study.

Wellness can be related to the prevention of diseases by means of ensuring the quality of products and is an important challenge in food industry. To this end, food traceability has become a priority in the industry in order let the final users to verify which production phases the food went through. Furthermore it gives domain experts the opportunity to trace defect or issues in the production workflow when problems arise. The proposed information system aims to track the production process of milk and vegetable products. This system is useful both for producers than consumers, giving them a complete tool for food traceability. It provides data management and processing in order to check each production step, making traceability a simpler and more efficient process. Information about raw materials, nutritional facts and activities is readily available and guarantees a transparent and secure supply chain.

Electroencephalography (EEG) is a technique for the acquisition of electrical brain signals. In recent years the increase of information acquired from signal analysis has generated a large amount of data; therefore, the development of tools for analysis has become necessary. In this paper, the mEEG prototype for EEG data managing is presented. It offers a user-friendly communication solution to exchange data between physicians and biomedical engineers. Features can be used for: (i) perform a fast diagnoses; (ii) show reports about clinical information; (iii) store and retrieve neurological data.

The analysis of bioimages and their correlated clinical patient information allows to investigate specific diseases and define the corresponding medical protocols. To perform a correct diagnosis and apply a precise therapy, bioimages must be collected and studied together with others relevant data as well as laboratory results, medical annotations and patient history. Today, the management of these data is performed by single systems inside hospital departments that often do not provide dedicated data integration platforms among different departments as well as different health structures to exchange of relevant clinical information. Also, images cannot be annotated or enriched by physicians to trace temporal studies for patients or even among patients with similar diseases. In this contribution, we report the results of a research project called GIDAC (standing for Gestione Integrata DAti Clinici) that aims to define a general purpose framework for the bioimages management and annotations as well as clinical data view and integration in a simple-to-use information system. The proposed framework does not substitute any existing clinical information system but is able in gathering and integrating data by using a XML-based module. The novelty also consists in allowing annotations on DICOM images by means of simple user-interface to take trace of changes intra images as well as comparisons among patients. This system supports oncologists in the management of DICOM images from different devices (e.g., ecograph or PACS) to extract relevant information necessary to query (annotate) images and study similar clinical cases.

Quantitative analysis of PET images is a useful as well as essential practice to perform an objective measurement of a physiological process. It allows to study diseases, evaluating treatment response and comparing patients data by quantify images. The analysis consists in estimating the quantity of radionuclide tracer uptaken by tissues. We focus on quantitative analysis of dynamic PET studies to evaluate the diseases of coronary artery and myocardium perfusion. We report experiences on quantitative cardiac PET analysis by using a commercial and largely used software to evaluate viable myocardium through Patlak method. We report also results obtained on PET images provided by clinical departments of the Magna Graecia University Medical School of Catanzaro.

Health promotion represents the process to empower the citizens to improve their health lifestyle and to achieve higher levels of wellness. The health models focus on helping people to prevent illnesses through their behavior, and on looking at ways in which a person can pursue better health or ideal health. We report on a project aiming to propose a new model for wellness improvement, consisting in actions to be performed to encourage individuals to become aware of their wellness and develop healthier habits.