8th European Medical and Biological Engineering Conference

We present a new numerical method for the solution of the forward problem of electrical impedance tomography (EIT) with the shunt model. Given a mesh over the EIT region, we discretize directly the conditions on the current density in equilibrium, and solve the resulting system of linear equations for the amount of current flowing through each side of every element. Afterwards, the distribution of current density and potential are reconstructed. Results of simulations on both 2D and 3D models indicate that the new method gives comparable results to those of the traditional finite element method with linear elements.

Emerging technologies for physiological signals and data collection enable the monitoring of patient health and well-being in real-life settings. This requires novel methods and tools to compare the validity of this kind of information with that acquired in controlled environments using more costly and sophisticated technologies. In this paper, we describe a method and a MATLAB tool that relies on a standard sequence of statistical tests to compare features obtained using novel techniques with those acquired by means of benchmark procedures. After introducing the key steps of the proposed statistical analysis method, this paper describes its implementation in a MATLAB app, developed to support researchers in testing the extent to which a set of features, captured with a new methodology, can be considered a valid surrogate of that acquired employing gold standard techniques. An example of the application of the tool is provided in order to validate the method and illustrate the graphical user interface (GUI). The app development in MATLAB aims to improve its accessibility, foster its rapid adoption among the scientific community and its scalability into wider MATLAB tools.

This article demonstrates the feasibility of the convolutional neural network (CNN) and pulse transit time (PTT)-based approach in estimating the systolic blood pressure (SBP) and diastolic blood pressure (DBP). Electrocardiogram (ECG) and photoplethysmography (PPG) signals were employed to calculate the PTT, which is the time delay between the R-wave peak Rof ECG, and specific points of the PPG waveforms. Then, the Blood pressure (BP), which is inversely related to PTT was estimated. A total of 22 patients with available ECG, PPG and SBP data were selected from the Medical Information Mart for Intensive Care (MIMIC III) dataset to validate the proposed model. A window of five minutes of recoding was chosen for each patient. Duration of each cardiac cycle was around 0.6 s, centred at R-peaks and sampled at 125 Hz. A CNN-based model was developed with four convolutional layers. The results showed that the average root mean square error (RMSE) of 5.42 mmHg and 7.81 mmHg were achieved for SBP and DBP, respectively.

Digital biofeedback systems (DBS) which use inertial measurement units (IMUs) can support patients during home rehabilitation. Models which accurately segment IMU data for rehabilitation exercises are required to provide biofeedback but assessing accuracy in a clinical context is challenging due to technical and patient-related factors. In this paper, we propose a three-stage validation framework to overcome these challenges. We present the results of stage one and two segmentation accuracy assessment for our DBS for shoulder rehabilitation. The results demonstrate that most of the chosen exercises can be segmented to a high level of accuracy in an unseen, uninstructed dataset. Errors in segmenting and recommendations for improvement are presented, which must be addressed prior to the final stage of validation.

In low- and middle-income countries, private and public facilities tend to have highly variable characteristics, which might affect their performance in meeting reporting requirements mandated by ministries of health. There is conflicting evidence on which facility type performs better across various care dimensions, and only few studies exist to evaluate relative performance around nationally-mandated indicator reporting to Ministries of Health. In this study, we evaluated the relationship between facility ownership type and performance on HIV indicator data reporting, using the case of Kenya. We conducted Mann-Whitney U tests using HIV indicator data extracted from years 2011 to 2018 for all the counties in Kenya, from the District Health Information Software 2 (DHIS2). Results from the study reveal that public facilities have statistically significant better performance compared to private facilities, with an exception of year 2017 in reporting of indicators for HIV counselling and testing, and prevention of mother-to-child transmission programmatic areas.

Psychophysiology studies human autonomic nervous system’s responses related to hers/his mental activity and behavior. Psychophysiology usually compares physiology during stimulus presentation (or task solving) with physiology of the person’s baseline state. This paper investigates instructions to the participants for them to achieve an optimal relaxed baseline state during psychophysiological studies. It is focused on physiology of human skin - electrodermal activity (EDA). Four different conditions for achieving the optimal baseline were tested: i) relaxing with no special instructions, ii) watching a fixed dot on the screen, iii) watching a video of a calming fish-tank, and iv) playing a mentally non-demanding computer game. The results indicate that computer game was not an appropriate condition for reaching an optimal baseline. The EDA in other three conditions did not differ significantly. Our findings suggest that the content of the instructions for reaching the perfect baseline may not be very relevant.

Most of the recently available research and challenge data lack the meta-data containing any information about the patient position. This paper presents a tool for automatic rotation of CT data into a standardized (Head First Supine) patient position. The proposed method is based on the prediction of rotation angle with convolutional neural network, and it achieved nearly perfect results with an accuracy of 99.55 %. We provide implementations with easy to use example for both, Matlab and Python (PyTorch), which can be used, for example, for automatic rotation correction of VerSe2020 challenge data.

Benefiting and accessing high-performance computing resources can be quite difficult. Unlike domain scientists with a background in computational science, non-experts coming from, e.g., various medical fields, have almost no chance to run numerical simulations on large-scale systems. To provide easy access and a user-friendly interface to supercomputers, a web-based service portal, which under the hood takes care of authentication, authorization, job submission, and interaction with a simulation framework is presented. The service is exemplary developed around a simulation framework capable of efficiently running computational fluid dynamics simulations on high-performance computers. The framework uses a lattice-Boltzmann method to simulate and analyze respiratory flows. The implementation of such a web-portal allows to steer the simulation and represents a new diagnostic tool in the field of ear, nose, and throat treatment.

In this study an alternative way to measure the tidal volume is introduced and evaluated. While the gold standard is spirometry or body plethysmography, the introduced measurement system, which was based on optical encoder modules used three respiration induced changes in circumferences obtained of the upper body to determine tidal volume. Thus, no face mask or mouthpiece is required, and the respiratory measurement is more comfortable for the subjects. R2 > 0.95 and a mean error < 208 ml showed that the system can be used as an alternative and comfortable way to gain tidal volumes and would be suitable for surveillance reasons or for longer tidal volume measurements.

This paper describes the concept of laboratory process based on measuring cardiac output using pulmonary artery thermodilution (trans-right-heart thermodilution) method. The blood circulation is in the process simulated by simple water flow in the measuring chain driven by water pump. Measuring is performed by Swan-Ganz catheter used in medical practice connected to simple electronic circuit using Wheatstone bridge to obtain temperature value from thermistor. Analog values come to NI Engineering Laboratory Virtual Instrumentation Suite (NI ELVIS) connected to PC. There is also a program designed in Labview environment to calculate cardiac volume, which is an important indicator of the cardiovascular system state.

Linear tomosynthesis is a limited angle, cone-beam X-ray based imaging modality. In practice, typically 40–60 low dose projection images are acquired over 20°–80°. As the consequence of the limited total scan angle, the reconstruction problem is highly under determined. Therefore, the resolution of the reconstruction is highly anisotropic, it is poor in perpendicular direction to the plane of the detector. This paper proposes an anisotropic weighted TV-L0 regularized (maximum a posterior estimation based) reconstruction method to linear tomosynthesis modality. The numerical optimization problem of the reconstruction is solved by majorization - minimization optimization. This paper compares the proposed method with isotropic TV-L0 regularized one and other well-known approaches by reconstructions calculated from simulated projections.

Calcium electroporation is a local tumor ablative treatment, where supraphysiological concentration of calcium is used in combination with electroporation to eradicate tumors. It was demonstrated in preclinical in vitro and in vivo studies that calcium electroporation has similar effectiveness as electrochemotherapy with bleomycin in different subcutaneous tumors. In this study two murine tumor models were used to evaluate how different tumors respond to increasing concentrations of calcium solution alone or in combination with electrporation. For this purpose, two histologically, physiologically and immunologically different tumors B16F10 mouse melanoma and 4T1 mouse breast carcinoma tumor models were selected. We showed that both tumor models respond to calcium electroporation in a dose dependent manner. However, 4T1 tumor model responded better with tumor cures even without addition of electroporation, when 250 mM calcium solution was used. These results demonstrated variable antitumor effectiveness of calcium electroporation which is dependent on tumor properties.

Cryomicroscopy is one of the main techniques to visualize freezing and thawing events taking place during cryopreservation of cells, native and artificial tissues with the ultimate goal to provide cell- and tissue-specific cryogenic preservation. However, there is currently no unified software tool for the automated analysis of ice recrystallization kinetics for a variety of cryoprotective agents used in the cryobiological practice. In this regard, we focused on the particular aspect of image analysis in the course of ice recrystallization, i.e. the possibility of using a neural network for the segmentation of ice crystals during isothermal annealing. In the work, the U-Net deep neural network was used for segmentation of ice crystals on cryomicroscopic images. Using 100 images as training set, the resulting accuracy of ice crystal segmentation was about 74% on the test sample (30 images). The obtained results show the possibility of segmentation of ice crystals in cryomicroscopic images taking into account the overlapping of intensity levels of an object and background. Further improvement of the model through the use of an additional training data as well as augmentation techniques is required to more efficiently validate this approach.

Surgical medicine is one of the most radical approaches for the treatment of numerous types of diseases. Recently broad application has taken the direction of computational surgery that aims to improve the quality of treatment through the use of computer tools. The use of computational surgery in rhinoplasty is important due to the fact that the results of the intervention directly affect the geometry of the nasal cavity and, as a consequence, the aerodynamic parameters of the nose. In turn, these parameters determine the functional characteristics of the patient’s nasal cavity. In this paper, we have focused on modeling the deformation of anatomical structures using SOFA framework software library considering tetrahedron finite element modeling (FEM), hexahedron FEM, triangle FEM and mesh spring force fields. The simulation results indicate the high functionality of the SOFA framework for modeling the deformation of the airway in rhinosurgical interventions. These results could further be applied for modeling the deformation of the anatomical structure taking into account the change in the topology of a 3D model to simulate such surgical procedures as a cut.

Extensive research has been carried out to find associations between arterial stiffness markers and blood levels of low-density lipoprotein cholesterol (LDL-Chol), triglycerides (TG), total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-Chol) relating to assessment of cardiovascular diseases’ (CVD) risk factors and early detection of atherosclerosis. However, data on levels of atherogenic lipoproteins as apolipoprotein B (apoB), low-density lipoprotein(a) (Lp(a)) and high-density apolipoprotein A (apoA) related to arterial biomechanical properties are scarce.The aim of this study was to investigate whether stratification of the population according to serum levels of atherogenic lipoproteins (apoB, Lp(a), TG, TC, LDL-Chol) and high-density lipoproteins (HDL-Chol, apoA) are related to arterial biomechanical properties for early and non-invasive cardiovascular risk diagnosis.The investigation was conducted on 44 healthy subjects with familial hypercholesterolemia in their anamnesis. A SphygmoCor device for the aortic pulse wave velocity (PWVao) and augmentation index estimation, and an ultrasound device for carotid intima-media thickness, the radial artery, and the femoral artery wall thickness and lumen diameter measurements, were utilized.An association between the serum apoB values and the radial artery wall thickness (p = 0.03) and the radial artery wall thickness to lumen diameter ratio was found (p = 0.05), indicating that the apoB level may be related to the arterial remodeling process. Also, a dependence between TG levels and age-related PWVao (p = 0.02) was discovered, which needs further investigation.

Electrical stimulation of the deep parts of the brain is the standard answer for patients subject to drug-refractory movement disorders. Collective analysis of data collected during surgeries are crucial in order to provide more systematic planning assistance and understanding the physiological mechanisms of action. To that end, the process of normalizing anatomies captured with Magnetic Resonance imaging across patients is a key component. In this work, we present the optimization of a workflow designed to create group-specific anatomical templates: a group template is refined iteratively using the results of successive non-linear image registrations with refinement steps in the in the basal-ganglia area. All non-linear registrations were executed using the Advanced Normalization Tools (ANTs) and the quality of the normalization was measured using spacial overlap of anatomical structures manually delineated during the planning of the surgery. The parameters of the workflow evaluated were: the use of multiple modalities sequentially or together during each registration to the template, the number of iterations in the template creation and the fine settings of the non-linear registration tool. Using the T1 and white matter attenuated inverse recovery modalities (WAIR) together produced the best results, especially in the center of the brain. The optimal numbers of iterations of the template creation were higher than those from the literature and our previous works. Finally, the setting of the non-linear registration tool that improved results the most was the activation of the registration with the native voxel sizes of images, as opposed to down-sampled version of the images. The normalization process was optimized over our previous study and allowed to obtain the best possible anatomical normalization of this specific group of patient. It will be used to summarize and analyze peri-operative measurements during test stimulation. The aim is that the conclusions obtained from this analysis will be useful for assistance during the planning of new surgeries.

Augmented reality is an emerging technology allowing to add computer-generated perceptual information superimposed to a real-world object. Biofeedback based on electromyography converts the muscle activation levels into visual or auditory information. This information can be used to facilitate or inhibit muscle contraction and is considered a suitable treatment for a wide range of musculoskeletal disorders. However, current feedback techniques are simplistic and not intuitive for the patient which limits clinical effectiveness. This work describes the design and development of an Augmented Reality system which allows the visualization of a graphical information about muscle activity superimposed on the EMG detection system over the investigated muscle. The system integrates sensors for muscle activity detection (both bipolar and high-density sEMG), wearable acquisition systems for the conditioning and wireless transmission of surface EMG signals, and a processing/visualization system to provide information on muscle activation to the user through augmented reality. A proof of concept of the developed system has been conducted for the following application fields: rehabilitation, sport, and ergonomics. The system is expected to: i) provide a novel tool to assess muscle activity in different scenarios and 2) increase the effectiveness of EMG biofeedback.

Measurement of a voltage response caused by a current excitation signal sent to a biological system under study, i.e. measurement of electrical bioimpedance, is typically using analog or digital lock-in detection. Such a detector consists multiplying unit together with following low pass filter. Result of such a multiplication is typically acquired in-phase with excitation signal, and in quadrature with it. Later the actual bioimpedance vector can be calculated with its magnitude and phase. It can be viewed as performing correlation between excitation and response signals in two positions: with zero shift and 90° shift. It requires two identical channels, the exact identity of which is hard to achieve. Method for automatic single-channel detection is proposed, analyzed and tested in the lab for the acquisition of the electrical bioimpedance signal at the radial artery. The proposed method is useful when wearable low power acquisition of the electrical bioimpedance is required. It simplifies circuitry significantly, by thereby increasing reliability and lowering power consumption.

MR is the most sensitive clinical tool in the diagnosis and monitoring of multiple sclerosis (MS) alterations. Spinal cord (SC) evaluation has gained interest in this clinical scenario in the last 10 years but unlike in brain, there is a lack of algorithms assisting SC segmentation. Our goal was to investigate and develop an automatic MR cervical SC segmentation method that would enable seamless imaging biomarkers extraction related to SC atrophy and lesion infiltration. This algorithm was developed using a dataset based on real-world MR data of 121 MS patients. 96 cases were used as training data and the remaining 25 cases were retained as the testing data. MR sequences used consisted of 3D-T1 gradient echo MR axial images, acquired in a 3T system (SignaHD-USA), (TE/TR/FA:1.7–2.7 ms/5.6–8.2 ms/12°). Manual labeling ground-truth is performed under radiologist supervision. The architecture of the 2D convolutional neural network consisted of a hybrid residual attention aware segmentation method trained to extract the region of interest. The training was designed with a focal loss function based on the Tversky-index to address the issue of label imbalance in medical image segmentation and an automatic optimal learning rate finder. Our model provided an automated and accurate method achieving a DICE coefficient of 0.87. An automatic method for SC segmentation from MR was successfully implemented. It will have direct implications for accelerating the process for MS diagnosis, follow-up and extraction of imaging biomarkers.

Information obtained from an inertial sensor attached on the foot would be useful in gait rehabilitation, healthcare and so on, because it provide various evaluation parameters such as stride length, inclination angle and gait event timings. The purpose of this study was to develop an estimation method of average walking speed with the inertial sensor attached on the foot. In this paper, average walking speed was estimated by estimated stride length and time of movement state of each stride during walking. The proposed estimation method of average waking speed was evaluated in comparing to measured walking speed with a stopwatch in 10 m walking test with healthy subjects. Mean difference of walking speeds between measured and estimated walking speeds was −0.60 ± 4.04% (−0.007 ± 0.048 m/s), and mean absolute value of the difference was 3.05 ± 2.71% (0.035 ± 0.033 m/s). Correlation coefficient between them was 0.986. Most of average walking speed were estimated with good accuracy, which were almost between ±10% and between ±0.1 m/s. These results suggested that the average walking speed estimation with an inertial sensor would be useful. In order to improve estimation accuracy for a specific subject and fast speed walking, in which stationary state between strides is very short time, is expected.

Heart-rate variability (HRV) is a strong indicator of a healthy cardiovascular system. It is the physiological phenomenon defined by the variation of the duration between consecutive heartbeats. Consequently, for a proper and a reliable HRV analysis, it is essential to have an accurate estimation of the inter-beat intervals (IBI). In addition to accurate IBI detection, unobtrusive and low power consuming devices are important for long term HRV monitoring. In this study, we aim at evaluating the beat-to-beat detection accuracy of the ultra low power Senbiosys Photoplethysmography (PPG) sensor, the SB200. Eight male subjects ( $$37.25 \pm 10.67$$ 37.25 ± 10.67 years of age) participated in the study. The recordings include a finger PPG from the SB200, another finger PPG from the Shimmer3 optical development kit, and a reference ECG from the Shimmer3 ECG development kit. The study shows that the SB200 detects $$99.27\%$$ 99.27 % of the beats with IBI values of mean absolute error (MAE) 6.58 ms compared to the R-to-R interval (RRI) values derived from the ECG and an average current consumption of less than $$190\,\upmu \text {A}$$ 190 μ A . Moreover, reducing the LED power consumption of the SB200 by 1/2 and 1/4 does not affect the detection rate, maintaining its value at $$99.25\%$$ 99.25 % and $$99.22\%$$ 99.22 % , respectively. However, it does reduce the IBI estimation accuracy resulting in an MAE of 7.37 ms and 8.43 ms, respectively.

The target of the study was revealing of a potential relationship between the tiredness accompanying a hypokinetic monotonic load and breathing. The secondary target was to describe it with the respiration parameters. After that, there was a question whether some of the parameters used can be applicable not only to the tiredness indication but also to its prediction.Five volunteers participated in the experiment, who underwent measurements with simultaneous monitoring of the brain electric activity, respiration and changes in volumes of the thorax. For the whole time period of the experiment, their activities were recorded by a video camera. In the first part of the experiment, the probants carried out a monotonic task (Tracking Task), in which they were supposed to pursue as precisely as possible a small target on the monitor moving with random velocities in random directions. In the second part of the experiment, the probants only relaxed watching a movie. The two parts of the experiment were measured in two variants – active and tired (after 24-h sleep deprivation) probants. The data acquired were subsequently compared.The results achieved demonstrated that the presence, and to a considerable extent also intensity, of the tiredness, can be reliably indicated and evaluated through the mediation of monitoring the occurrence of respiration volume and frequency non-stationarities manifested in otherwise steady-state breathing regimens in the persons monitored. Based on the results achieved, it is also possible to state that the respiration cycle monitoring can also be used for the prediction of the tiredness.

The paper presents the thermal characterization of Fe-Mn oxide nanoparticles with high SAR value, when used in magnetic field hyperthermia conditions. The synthesis shows good reproducibility. Finally, the paper presents a possible use of the presented nanoparticles included in PLGA nanocarriers for biomedical application.

The aim of this work is to perform a survey of the BME educational programs offered in Europe today, compare them with the situation ten and twenty years ago and identify potential trends and approaches. The results demonstrate the exploding growth of the field with the number of graduate and postgraduate study programs offered, to approximately double every 10 years. According to our findings today across Europe 182 Universities offer 344 BME educational programs, of which 115 are Undergraduate offering BSc degrees, and 229 postgraduate programs, 175 of them offering an MSc degree and 54 PhD degrees. The Clinical engineering programs are still less than ten, but their number is increasing. In conclusion BME education is getting a leading role in engineering studies, almost everywhere in Europe. However, harmonisation of studies is necessary for the advancement of the BME/CE profession. This should come out from a wide acceptance, of a consensus-based agreement on a generic core curriculum, that will facilitate a worldwide opening of the BME job market, through mutual recognition of the competencies acquired.

This research is to hypothesize if C60 fullerene and cerium oxide nanoparticles (CeO2NPs) can be useful in cryobiological practice. To check the hypothesis solutions of dimethyl sulfoxide (DMSO) with and without nanoparticles were compared. The features of aqueous solutions crystallization are the subject of many cryobiological studies since the uncontrolled ice crystals formation is often the main cause of damage during cryopreservation of cells and tissues. The effect of nanoparticles on phase and glass transitions in aqueous DMSO solutions was studied using the method of low-temperature differential scanning calorimetry. The results of the study indicate that nanoparticles adding to the aqueous solution of DMSO leads to changes in low-temperature phase transitions. The research results showed that the confirmation of the hypothesis necessitates further comprehensive research.

Mental workload (MW) represents the brain resources an individual devotes to a task. The evaluation of MW is fundamental for Advanced driver-assistance systems (ADAS). To assess MW, non-invasive techniques are preferable to avoid interference with the driver. Infrared Thermal Imaging (fIRI) is highly suited given its contactless nature. The research reported aimed at developing a contactless physiological method to measure MW employing fIRI features obtained from human facial skin temperature modulations. The novelty of this study is that MW was evaluated employing functional Near Infrared Spectroscopy (fNIRS), a non-invasive optical technique that measures the hemodynamic oscillations related to cortical activations. Particularly, the Sample Entropy of the fNIRS signal was assumed as indicative of MW. A two-level (i.e. High MW vs. Low MW) Support Vector Machine (SVM) classifier with linear kernel was employed to predict the level of MW from fIRI features. A leave-one-subject-out cross-validation was implemented to test the generalization performances of the method. The classifier showed a cross-validated sensitivity of 77% and specificity of 69%. This study represents the first attempt to estimate MW evaluated by fNIRS from fIRI features.

Quantitative Phase Imaging (QPI) is a label-free microscopic technique, which provides images with high contrast, moreover, it provides quantitative cell mass measurements for each pixel. Segmentation of particular cells is an important step in the analysis of QPI image data. This paper presents a method for automatic cell segmentation in QPI images. The proposed method improves iterative thresholding, which is a very promising method, however, it is not able to segment densely clustered cells. Our improved iterative thresholding includes two additional steps – Laplacian of Gaussian image enhancement and distance transform-based splitting. The method was compared with original iterative thresholding and another method on two cell lines, where the proposed method successfully deals with a densely clustered type of cells and achieves significantly better results on both datasets.

Respiratory diseases are characterised by high prevalence among the European population. Medical aerosol inhalers are the most commonly used means of drug delivery into the human respiratory system. This paper focuses on characteristic waveforms that can be utilised during aerosol deposition studies to simulate conditions of rapid human inhalation. Additionally, an inhalatory waveform based on clinically recorded spirometry data is introduced. Experimental measurements are performed and simulation results mutually compared using the electro-mechanical lung simulator xPULM $$^{\text {TM}}$$ TM . The inhalatory waveforms are repeatably simulated with high fidelity in regards to the waveform shape with the lowest value of the Goodness of fit 0.89. Additionally, the measured values for all characteristic inhalatory parameters are simulated with low standard deviation $$\sigma $$ σ < 1. The differences between the required and measured waveform shapes are small, $$\varDelta PIF$$ Δ P I F < 3 L/min and do not influence the overall inhalatory volume. This opens a possibility of utilising the xPULM for medical aerosol inhalers testing.

Electrochemotherapy is a local tumor ablation technique with different effectiveness in different histological types of tumors. In the treatment of breast cancer metastases, the effectiveness of electrochemotherapy is among the lowest. Therefore, combined therapies are needed. The aim of our study was to combine electrochemotherapy with PARP inhibitor olaparib, which would inhibit the repair of bleomycin/cisplatin caused DNA damage and would most likely potentiate the effectiveness of electrochemotherapy. First we demonstrated the effectiveness of electrochemotherapy with bleomycin and cisplatin alone in human breast cancer cell line MCF7 in vitro. Further, we confirmed that effectiveness of electrochemotherapy could be potentiated by combining it with olaparib, but only when bleomycin was used as a cytotoxic drug.

Magnetoencephalography (MEG) is a neuroimaging technique for measuring magnetic signals in vicinity of the head. With various source localization algorithms we can estimate the excited areas inside cortex. Standard MEG devices use SQUID-based channels which, despite their very good signal-to-noise ratio (SNR) have numerous drawbacks. As an alternative to these, commercial optically pumped magnetometers (OPMs) have recently developed to such a degree, that they are suitable for measuring magnetic fields in MEG. In this work we present measurements of the brain auditory evoked fields (AEF) with a system of 15 OPM sensors, that can detect both radial and tangential components of the magnetic field. These results are compared to the results obtained with the SQUID system. However, the quantitative comparison of two MEG systems is not trivial. We present a method for comparing two MEG system, which operates on the principle of minimum norm estimate (MNE) source localization algorithm. We show, that performance of the MEG system consisting of a small number of OPM sensors is slightly worse, but still comparable to results of the complete standard squid system with 125 gradiometers.

This work deals with similarity measuring and indexing in biosignal time series. In particular, it focuses in an approach proposed by the authors in previous works, which consists in a transformed-based method, performed by means of a set of Haar wavelet basis functions that is reduced to an optimal subset through the Karhunen Loève transform (WKLT). This allows a biosignal to be efficiently described by the combination of a compact set of bases which coefficients are used in the similarity process. For similarity indexing, the method uses an iterative formulation that leads to a significant reduction of the computational complexity, namely in the required number of operations.Therefore, the goal of the present work is to demonstrate the efficiency of the referred approach, through a complexity analysis study. For this purpose, biosignals collected in intensive care units (MIMIC-II dataset) are employed in a set of experiments that compare the proposed approach with others, and which results confirm expectations.

The main assumption of the reservoir-wave model is that the blood pressure curve can be separated into two components: reservoir pressure and wave pressure. Also, the linear relationship between reservoir pressure with aortic volume and the wave pressure with blood flow velocity are assumed. This indicates that aortic characteristic impedance and arterial compliance should be constant during the whole cardiac cycle. Using impedance cardiography it is possible to measure changes in thoracic electrical impedance which reflects fluctuations of arterial blood volume and allows to estimate reservoir pressure.The aim of this study was to examine the relationship between blood pressure and impedance cardiography signal (ΔZ). For the preliminary, illustrative purposes, we examined data from three subjects. We tested if it is possible to assume linear characteristic impedance and linear compliance for signals measured in humans.We compared the theoretical pressure-volume loop computed using the model with the loop obtained from measured blood pressure and ΔZ. We found that the slope of the “real” loop is disturbed. The slope becomes linear only in late diastole. Motion artifacts and some disturbing waves distort ΔZ curves. Also, the occurrence of a backward pressure wave may disturb the expected similarity between the two curves. We also checked the slope between estimated wave pressure and blood flow velocity obtained through Doppler ultrasonography (characteristic impedance) and found a nonlinear relationship between them. It can be described by a power function (y = xn). We found that ΔZ curve does not fit well enough to blood pressure trace in the reservoir-wave model. Thus, we concluded that the assumption of using a constant value of compliance is an unjustified simplification. They should be treated as a blood pressure dependent value.

Measurements derived from the joint space segmentation are clinically pertinent to study knee osteoarthritis. Cone beam computed tomography (CBCT) is an emerging low dose imaging method with the potential to be used in a weight bearing position. With a CBCT prototype, we have tested 2 methods of reconstruction: iterative reconstruction (SART) and FDK reconstruction with different intensities and projections number. We used the segmentation of the joint space as a metric to assess the quality of reconstructions. For this aim, we calculated Jaccard (JAC) and Hausdorff indexes (AVD), and thickness measurements 2D and 3D (respectively JS.Th-2D and JS.Th-3D). We have found that the results were more consistent with the SART reconstruction than FDK reconstruction with more stable results whatever the intensities and the projections number. Indeed, JAC indexes were superior to 0.7 at 10 and 15 mA for SART reconstruction and unlike to FDK reconstruction, only 19% values were superiors to 0.7. Moreover, the referent value of JS.Th-3D being 5 mm, the JS.Th-3D results from the SART reconstruction were closer to this value ranging from 4.8 mm to 5.3 mm, whereas, the JS.Th-3D results from the FDK reconstruction were ranging from 4.4 mm to 4.7.

The use of a controlled cooling therapy may help in pain management and pain treatment, for example, of carpal tunnel syndrome. The article deals with possibilities and design of the control system and user interface for Peltier-based cooling module for braces. The experimental brace fixing the hand in the carpal region was designed to verify the control system of cooling forearm orthosis. This brace was fitted with 3 thermoelectric modules. The users can select their own cooling intervals, cooling intensity, or the set periods of required cooling with respect to the presumed battery lifespan. The verification of the system was performed on patients with carpal tunnel syndrome during 60-min measurement. The proposed wearable system has proven to be useful in the field of rehabilitation, but also reveals some limitations of using fans as tools for heat transfer, e.g. if it is necessary to cool larger areas, the energy consumption for cooling increases and portable batteries have limited usability and external sources of electricity must be assumed.

This case series study was designed to provide an overview of the efficacy of an active pelvis orthosis (APO) against the fall risk after multi-directional slippages in transfemoral amputees (TFA). To achieve this goal, we investigated the dynamic stability after antero-posterior (AP) and diagonal (D) slippages, as assessed by the Margin of Stability (MoS) in both the frontal and the sagittal planes.Results revealed that the detection algorithm can actually signal a lack of balance in about 400 ms, for both AP and D slippages, for three over five amputees. However, it was also noticed that its performance decreased with the severity of the amputee’s clinical status. Specifically, for most impaired amputees walking patterns are inherently less smooth thus the detection algorithm failed to detect abrupt modifications of gait patterns due to external perturbations.Results also demonstrated that the proposed assistive strategy can effectively promote balance recovery in the sagittal plane while subjects managed AP-slippages. On the other hand, the analysis of the stability in the frontal plane showed that the balance control does not systematically improve due to the APO assistance. Therefore, despite the APO could partially restrain the hip movement of the users in the frontal plane, balance control is mostly mediated by subjects’ abd/adduction muscle groups leading them to adopt a context- and subject-dependent counteractive strategy to manage the lack of balance.Concluding, the outcomes of this preliminary study are promising and suggest to further investigate subject-specific tuning of the control algorithm underlying the APO-mediated assistance.

Open issues and unmet needs in healthcare include the enhancement of the statistical power of the clinical outcomes along with the development of prediction models for effective disease management, the detection of prominent factors for disease progression and the identification of targeted therapies. In this work, we deploy a computational pipeline that uses data curation and ontology matching to curate and align heterogeneous data structures to enhance the statistical power of the outcomes. Then, we use federated learning to develop disease prediction models across harmonized cohort data that are stored in private cloud databases. A preliminary case study was conducted for the first time on three European cohorts on primary Sjögren’s Syndrome (pSS) yielding harmonized data with 90% average overlap along with a federated lymphomagenesis progression model with accuracy 0.848, sensitivity 0.833 and specificity 0.849.

Urinary tract infection (UTI) is very common and caused by various bacteria. The treatment of UTIs should be handled carefully. With the rise of neural networks, a possibility occurred to predict the outcome of consuming antibioticis for treating UTIs. This paper presents the development of expert system based on neural network for prediction of antibiotic sensitivity of two bacteria: Escherichia coli and Klebsiella pneumoniae. For the development of expert system based on neural network, total of 3226 samples were used: 486 samples of Klebsiella pneumoniae and 2740 samples for Escherichia coli. All samples were collected in one geographical area from hospitals and primary healthcare units. Feedforward neural network based on Bayesian regularization backpropagation training algorithm resulted in accuracy of 72.16% for prediction of antibiotic sensitivity of K. pneumoniae bacteria and 99.81% for prediction of antibiotic sensitivity of E. coli bacteria. The results of this study are promising since the usage of such expert systems in healthcare environment contributes to rational usage of antibiotics for treatment of infections and therefore contribute in fighting the antimicrobial resistance which is one of the rising challenges of healthcare nowadays.

We analyze the performance of recently introduced technique for decomposition of compound muscle action potentials (CMAPs), recorded by high-density surface electromyograms (HDEMG). This technique utilizes Convolution Kernel Compensation (CKC) method to estimate motor unit (MU) filters from HDEMG recordings of voluntary isometric muscle contractions. Afterwards, it applies these filters to the HDEMG recordings of elicited contractions and identifies MU spike trains. We then propose a novel method for segmentation of MU firings out of the identified spike trains and demonstrate on a synthetic HDEMG signals that at high MU synchronization levels this novel segmentation method significantly outperforms the spike segmentation introduced by CKC method. Namely, at 80% MU firing synchronization the area under receiver operating characteristic curve (AUC) increases from 0.96 ± 0.03 to 0.99 ± 0.01 when newly proposed segmentation is used instead of the previously introduced CKC-based segmentation. Thus, newly proposed segmentation supports significantly more accurate discrimination of true positive and false positive MU firings. At low MU synchronization levels (<60%) the newly proposed MU firing segmentation yields results that are comparable with the previously introduced CKC-based segmentation.

High-density surface electromyography(HDsEMG) decomposition has allowed us to study individual motor unit (MU) responses in great detail during voluntary or reflex muscle contraction. Being noninvasive in nature, the technique has been widely used for many different applications in both healthy and stroke impaired populations. Here for the first time, we demonstrate the use of a HDsEMG based MU decomposition technique to study the morphological changes in the recruited MU population after botulinum toxin (BT) injection in the biceps brachii muscle for spasticity management in chronic hemiparetic stroke survivors. Three stroke survivors were examined before and after intramuscular (biceps brachii) BT injections. The HDsEMG grid enabled simultaneous recordings over the entire muscle. We have recorded both force and the surface electromyogram (sEMG) during voluntary isometric contraction tasks. The HDsEMG was decomposed using the convolution kernel compensation (CKC) method. We report a 60% increase of the peak to peak amplitude of the motor unit action potential (MUAP) signals after the BT injection compared to pre-injection values. The overall generated muscle force and sEMG values decreased during this period compared to their pre-injection level. We discuss potential mechanisms that would result in the emergence of larger MUAPs in the weeks immediately following intramuscular BT.

Myocardial infarction is the leading cause of death worldwide. In this paper, we design domain-inspired neural network models to detect myocardial infarction. First, we study the contribution of various leads. This systematic analysis, first of its kind in the literature, indicates that out of 15 ECG leads, data from the v6, vz, and ii leads are critical to correctly identify myocardial infarction. Second, we use this finding and adapt the ConvNetQuake neural network model—originally designed to identify earthquakes—to attain state-of-the-art classification results for myocardial infarction, achieving 99.43% classification accuracy on a record-wise split, and 97.83% classification accuracy on a patient-wise split. These two results represent cardiologist-level performance level for myocardial infarction detection after feeding only 10 s of raw ECG data into our model. Third, we show that our multi-ECG-channel neural network achieves cardiologist-level performance without the need of any kind of manual feature extraction or data pre-processing.

Sleep spindles are bursts of brain activity during sleep. They occur during the NREM2 stage of sleep and appear as fluctuations in electric recordings, looking like yarn spindles. This increase of activity can be detected by complexity measures, the most popular of which are the entropy based estimations. In this paper, we use entropy to measure the brain activity during sleep spindle and non-spindle periods and discriminate them employing the machine learning technology. Two are the main outcomes of this work: a) we show that it is possible to achieve remarkable classification performance when detecting sleep spindles with entropy based measures and machine learning techniques, presenting classification accuracy of more than 95 $$\%$$ % and (b) we report that bubble entropy, a recently introduced definition of entropy, presented the lowest p-value of all examined features.

Many different urinary tract conditions are characterized by overlapping symptoms, such as strong urge to urinate, burning sensation, abnormal urine output, and fewer, which make the diagnosis and treatment difficult and time-consuming. In order to skip traditional time-consuming diagnostic methods, the direct treatment of patients with confirmed symptoms may be administered. An expert system that uses artificial intelligence methods can be used in order to solve problems of a specialized domain. In this paper, Artificial Neural Network for the classification of acute inflammation of urinary bladder and acute nephritis of renal pelvis origin is presented. Dataset of 120 samples has been used to determine whether a patient is suffering from both diseases, only from one disease or is healthy. A feedforward neural network was trained and yielded an accuracy of 95.83%. The sensitivity of the developed system is 94.44%, while specificity is 100%. This system allows the clinician to enter different symptoms observed or reported by a patient and understand the connection between them in order to more accurately provide diagnosis and prevent possible misdiagnosis.

Juvenile idiopathic arthritis (JIA) represents the most common rheumatologic disease in childhood, often characterized by temporomandibular joint (TMJ) disfunction (TMD). The gold standard to diagnose this pathology relies on magnetic resonance imaging. Alternatively, electromyographic (EMG) recordings could provide an early and immediate detection of TMD. Particularly, temporalis muscle is known to exhibit a greater EMG activity more frequently than the masseter in pathological subjects. Since muscular activity may influence the superficial circulation and, consequently, the skin temperature, the capabilities of functional thermal infrared imaging (fIRI) to detect TMD were also investigated. In this study, the feasibility of a multivariate data-driven approach based on General Linear Model to estimate the EMG ratio between masseter and temporalis (sEMG-M/T) from fIRI features was investigated. A leave-one-subject-out cross-validation was implemented to test the generalization capability of the model (r = 0.55; p = 1.72·10−6). Moreover, the output of the model was used to classify TMD and healthy controls. Since the two classes were unbalanced, a bootstrap procedure was applied. The performances of the classifier were investigated through Receiver Operating Characteristic analysis, which exhibited an area under the curve of 0.71. The results suggested that fIRI could be a relative cheap and simple to use tool for TMD assessment.

Instrumented versions of functional geriatric screening tests have been developed to improve clinical precision. Several different instrumented versions of the Sit-to-Stand (iSTS) test have been developed using a range of sensors such as accelerometers and cameras. An instrumented chair equipped with load cells and an ultrasound sensor was developed to detect phases of the STS (Sit to Stand). The chair was designed to be able to detect all the phases of the STS, including when the person was not in contact with the chair. Performance of the iSTS chair was compared between an RGB camera approach, and a data-fusion approach using the load-cell and ultrasound equipped chair. Ten adult subjects were tested performing the 5STS at two self-selected speeds. The accuracy of the load cell equipped chair was 70%, while the RGB camera achieved 76% accuracy. The ultrasound version of the chair and the fusion of the RGB and load cells technique both achieved significantly better accuracy at 86% and 89%, respectively. The new version of the instrumented chair obtained a high degree of accuracy in detecting the different phases of the STS and is suitable to detect STS phases without requiring additional sensors. Future work will test older subjects and aim to develop new parameters based on the phases of the STS as indicators of physical performance.

As novel tissue engineered constructs (TECs) are developed, current tissue banking practices need better control over ice formation and growth to prevent cryodamage to cells and a scaffold. Directional solidification demonstrates benefits in adhered cells and native tissues cryopreservation through controlled heat transfer. Therefore, this study aims to investigate the feasibility of using this technique for cryopreservation of cell-seeded electrospun fiber mats as model TECs. Fiber mats were produced using blend electrospinning of polycaprolactone (PCL, 200 mg/ml) and poly-L-lactic acid (PLA, 100 mg/ml) dissolved in 2,2,2-Trifluoroethanol. The fiber size and morphology was characterized using scanning electron microscopy. Specific heat measurements were conducted using differential scanning calorimetry. The square-shaped fiber mats were seeded under static conditions with HeLa cells and cultivated for 24 h. Samples were directionally frozen in a sandwich format either in 10% DMSO or culture medium with the sample movement at 30 μm/s through the predetermined temperature gradients along a 2.6 mm slit. After directional solidification, samples were gradually frozen at 1 K/min down to −80 ℃. Crystal shape was visualized using cryomicroscopic system. Before freezing and 24 h after thawing, cell viability was assessed using live-dead assay. Within randomly orientated PCL-PLA fibers, HeLa cells exhibited typical shape and attachment with higher than 90% viability prior to freezing. While up to 80% of HeLa cells were alive on fiber mats after freezing using DMSO with or without directional solidification step. The demonstrated controlled freezing may assist optimizing the freezing of more sensitive cells. The results suggest that directional freezing becomes a viable option for cryopreservation in tissue engineering applications.

This paper presents a machine learning method to detect and discriminate between Aortic Stenosis (AS) and Bicuspid Aortic Valve (BAV) based on heart sound analysis. Differentiation between the two heart conditions is clinically important, but complicated if relying merely on the conventional auscultation. A novel form of the Time Growing Neural Network (TGNN) is introduced for the classification purpose. The method is applied to a dataset comprised of 87 children referrals to a university hospital, from which 50 individuals are healthy (with and without innocent murmur), and the rest are abnormal with either AS (15 individuals) or BAV (22 individuals). The baseline for comparison is a Time-Delayed Neural Network (TDNN) with the same size of the feature vector and the temporal frame. We used our original validation methods, named A-Test, which provides valuable information about structural risk and also learning capacity of any supervised classification method. A-Test is an elaborated version of K-Fold validation method, in a rather profound way. Performance of the TGNN is superior comparing to the presented TDNN, with an accuracy of 85.8% against 81.5%. This method can be integrated with our intelligent phonocardiography to serve as an enhanced assessment tool in hands of nurses or practitioners at primary healthcare centers.

Sharing information between healthcare organizations is a must for the next-generation healthcare systems. Despite the huge advantages that have been lately associated with it, organizations are still reluctant to its adoption due to the related privacy issues. In this work we propose a distributed access control system that relies on Garbled Circuits to perform a XACML-like policy evaluation, where attributes required for the evaluation are spread across organizations, ensuring that no attributes of one organization can be learned by others. Since Garbled Circuits is used as underlying protocol, an analysis of the complexity of the proposed method is provided in terms of Non-XOR gates. Some examples of the cost of using this method with different policy sizes is also provided with an estimation of the achievable amount of evaluations per second. The obtained results are promising but further research is still needed to determine which real-world scenarios would be suited for the proposed system.

Improvements in the obstetrical and neonatal management have allowed children to survive. These enhancements have showed, anyway, a general increased incidence of healthcare-associated infections, one of the most influent causes of morbidity and mortality in neonatal intensive care units. The aim of this paper is to suggest corrective measures to reduce sentinel germs colonization and identify the relationships between bacteria colonization with the number of procedures and the length of hospital stay. The Lean Six Sigma methodology was used to tackle this issue using a tailored Define, Measure, Analyze, Improve, and Control problem-solving strategy. An increased number of procedures and an extended length of hospital stay demonstrated a statistically significant influence on newborns’ possibility to be infected by sentinel germs. These findings could guide the clinical staff to improve the management of neonates in neonatal intensive care units reducing the number of infected patients, their length of hospital stay and the costs for the hospital.

Introduction. Maxillary and mandibular cancer represents the sites where cystic and neoplastic conditions, of a benign or malignant nature, can occur. In general, for patients undergoing surgery to remove cancer of the oral cavity, the administration of oral antibiotics, Ceftriaxone and Cefazolin plus Clindamycin, can affect the Length of Stay (LOS).Objective. The objective of this study is to evaluate the efficacy of administering two antibiotics (Ceftriaxone and Cefazolin plus Clindamycin) through the statistical analysis of LOS in patients undergoing oral cancer surgery (mucosa/bone) at the Department of Maxillofacial Surgery of the University of Naples “Federico II”.Methods. The retrospective study was carried out to compare the efficacy of administering two antibiotics by analysing postoperative LOS in two different groups of patients by applying the Six Sigma (SS) methodology which follows the Define, Measure, Analyse, Improve, Control - cycle. Statistical tests were carried out during the Analyse stage to investigate the variables which could effectively influence the LOS of the two groups. A demographic study was conducted by applying the chi square test for each variable.Results. A statistically significant decrease in LOS was shown in patients not undergoing lymphadenectomy and in patients not exposed to surgical site infections treated with Ceftriaxone. A statistically significant difference in frequency (chi square test) was obtained based on the American Society of Anaesthesiologists score.Conclusions. This study showed the validity of the SS methodology for evaluating the efficacy of the administration of antibiotics in patients undergoing surgery for oral cancer.

Among skeletal deformities, Class III is the one that orthodontics gives more importance to timing compared to Class I or Class II, because growth modification is possible only before pubertal growth spurt. When class III malocclusion is diagnosed late it is especially difficult to treat with braces frequently requiring surgical intervention. In this study, we assessed the potential of a computational model for detecting Class III Malocclusion using the profile images of patients. The main purpose of our project is to develop this model into a mobile application for parents seeking a guidance on whether to reach out to an orthodontist at an early stage of developmental bone growth with a warning of Class III malocclusion risk.For detecting Class III malocclusion by discriminating the skeletal status from each other, we utilized several different angles from literature to mark the points on the profile, which are G, Prn, Sn, Ls, Li, Sm and Pg. In this study, a test dataset consisting of 60 profile images of patients is used to evaluate the performance of several different heuristic criteria. We devised three heuristic methods to evaluate the performance on our patient test data. In all three methods, if the calculated angles match Class III angle mean+-standard deviation intervals, patients are categorized as “Class III”. If the angles are outside the standard deviation intervals, we assigned the category as “not Class III”.The most successful method so far was the third method, categorizing the 49/60 images correctly (85% in Class I, 100% in Class II and 60% in Class III).

The human central nervous system integrates different sensory modalities with the visual information to produce a coherent mental representation of our own body, making us capable not only to process sensory events but also to plan and executes movements in the surrounding space. The basis of Mirror Therapy (MT) is the use of a mirror to create a visual reflection of an affected limb to create an illusion of movement of the paretic part of the limb. One of the uses of the MT is in motor recovery in post-stroke hemiparesis and even thought that it is valuable rehabilitation tool, its underlying neurophysiological manifestations and interaction with Motor Imagery (MI) are still unknown. Our study aim is to assess the effect of the MT by applying Forearm Bisection Test (FBT) and EEG measurement of the Event-related (de)synchronisation (ERD/S). Our results show that FBT scores were significantly higher in experimental Mirror Box (MB) group compared to control (CN) group (median 13.54 vs 0.00, respectively; p = 0.003). Furthermore, ΔERD/S (post-pre) differed significantly between the hemispheres in the MB group in the Mu, betalow and betahigh EEG bands, whilst did not differ in the CN group in the Mu and betalow bands. The results demonstrate improvement in ERD/S MI and an update of the body representation caused by MT. Moreover, findings suggest that the reflection of the moving hand in the mirror created an illusion of concomitant movement in the opposite hand that modulate the arm length representation which is detectable even during MI at EEG level. Our findings of neural basis and link of the MT and MI supports MT as favourable neurorehabilitation tool for motor recovery affecting not only the areas governing the moving hand but also the corresponding regions of the other hemisphere.

Purpose: To evaluate the lung compliance variation over the course of COVID-19 pneumonia, and to classify the patients into different types described as recruitable and non-recruitable, which lead to different ventilator support treatment. Method: Two ICU admitted COVID-19 patients, who were mechanically ventilated for more than 7 days, were included into this investigation. During a daily recruitment maneuver - a PEEP trial - they were monitored by Electrical Impedance Tomography (EIT). Deflation time constants were calculated offline from EIT data to determine the type of patient and to observe the transition of different types over the course of pneumonia. Result: The first patient was recruitable and had the tendency of transition to the other type. The second patient is non-recruitable. Both patients showed low lung compliance, but the first patient started in a better condition (higher compliance). Conclusion: EIT-based breath-by-breath time constant analysis can classify COVID-19 pneumonia into different classes of patients. The deterioration of lung mechanics can be monitored online by EIT which may help to find proper ventilation treatment.

Precise detection of electric current flowing in a conductive liquid sample by MRI (current density imaging technique – CDI) is often challenging due to required special hardware for application of electric pulses, need for sample reorientation during the experiment or for multiple current injections, and lastly, need for special pulse programs. In this study, another, much simpler approach of current detection, which based magnetic resonance spectroscopy (MRS), is considered. It is shown first theoretically and then by experiments on a phantom, that it is possible to detect currents, by change in $$ T_{2}^{*} $$ T 2 ∗ NMR relaxation time, with a comparable or even lower detection threshold as with CDI. In an image voxel $$ T_{2}^{*} $$ T 2 ∗ relaxation time changes due to magnetic field change induced by electric currents. The sensitivity of the method is proportional to the voxel linear dimension and to $$ T_{2}^{*} $$ T 2 ∗ relaxation time and the signal-to-noise ratio (SNR) of the signal from the voxel. While the proposed method cannot be used for an exact current determination and is sensitive only to the current component perpendicular to the static magnetic field its main advantages are simple implementation and good sensitivity that stems from an excellent SNR of proton MRS.

Unknown electric conductivities of human tissues is a common issue in medical engineering. Electrical impedance tomography (EIT) is an imaging modality that can be used to determine these conductivities in vivo from boundary measurements. In this paper, we demonstrate that local conductivity values of different skull segments can be solved from EIT measurements with the help of a box constraint. Based on our numerical results, the accuracy of the results depended on the locations of the current carrying electrodes and the signal to noise ratio of the measurements. Particularly, the conductivity values of the skull segments that located below the current carrying electrodes were reconstructed more accurately.

We have recently reported on the high practical utility of using electroporation of clinically relevant cells with sugars for their xeno-free cryopreservation in suspension. This paper extends our earlier approach to in situ electroporating attached cells for their robust cryopreservation on artificial scaffolds. Using CAD modelling, a two-electrode setup has been designed and in-house constructed allowing for simultaneous electroporation in multi-well cell culture plates. Blend electrospinning process has been optimized in order to manufacture porous fibrous mats made of polycaprolactone (PCL, 100 mg/ml) and polylactide (PLA, 50 mg/ml) with an average thickness of 100 µm. Chinese hamster ovary (CHO) cells were grown and directly electroporated in nanofibrous blend electrospun fiber mats. An electric pulse was applied in the presence of propidium iodide and CellTracker Green to determine viable permeabilized cell counts using fluorescence microscopy. Cell recovery was evaluated using metabolic MTS assay 24 h post-electroporation. Electric field intensity and distribution within a 3D reconstructed fiber mat was simulated and visualized using COMSOL software. The results demonstrate that with developed setup it is feasible to electroporate around 80% of attached cells with 80% viability after electroporation when electric field strength was ≥1.7 kV/cm. COMSOL simulations showed local increases of electric field at intersection points of numerous fibers which may in part contribute to the observed drop in cell viability post-electroporation. Future studies anticipate implementation of the developed approach in effective biopreservation of stem cells on electrospun fiber mats as a model of tissue-engineered constructs.

Atherosclerosis is a frequent disease in developed countries. Any vessels’ morphological impairments lead to blood flow changes. Even if vessels’ hemodynamic patterns extraction from X-ray fluoroscopic angiography has stimulated many scientists for 60 years, there is not a robust method to be implemented into angiography software to be used in daily clinics such as in the catheterism laboratories.A new algorithm is implemented into a graphical user interface for estimating blood velocity of the vessel’s segments selected from coronary monoplane angiography. The method is not invasive except from the medical imaging acquisition. The method includes the detection of the absolute vessel length, the blood transit time and hence, velocity estimation for a vessel segment from 20 patients.The validation is done using the measurements from the medical scientific literature. A further validation with a blood vessel phantom is considered.

Impedance cardiography (ICG) was discovered nearly half a century ago, being proposed as noninvasive monitoring method for estimation of several hemodynamics parameter. Despite of nearly 5 decades of clinical research and use there is still certain controversy about its performance when estimating Left Ventricular Ejection Time (LVET). This work present a comparison between using the different ICG subtype waveform and the ensemble averaged (EA) method to calculate the LVET. The ICG has been recorded from four volunteers, and the LVET parameter has been calculated using the two approaches. The result shows that each volunteer have different percentage of subtypes, and the mean relative error between the two approaches for estimation of LVET varied between 0.62 to 2.9% with an average mean absolute percentage error of 18,02% ranging between 13.82 to 18.42%.

Word embeddings have been widely used in Natural Language Processing as the input to neural networks. Such word embeddings can help in the understanding of the final objective and the keywords in a sentence. As such, in this work, we study the impact of different word embeddings trained with general and specific corpora using Joint Natural Language Understanding in a Spanish medication domain. We generate data using templates for training the model. The model is used for intent detection and slot-filling. We compare word2vec and fastText as word embeddings and ELMo and BERT as language models. We use three different corpora to train the embeddings: the training data generated for this scenario, the Spanish Wikipedia as general domain and the Spanish drug database as specialized data. The best result was obtained with word2vec continuous bag of words model learned with Spanish Wikipedia, obtaining a 71.77% F1-score for intent detection, an intent accuracy of 69.37% and a 74.36% F1-score for slot-filling.

The implementation of the training courses is remarkable in the field of education, in order to analyze and observe individual’s perception of the experience. Aim of this work is the combination of two methods, the Likert scale and the Analytic Hierarchy Process (AHP), to evaluate the quality of the training activities offered at Centre of Biotechnology of the National Hospital A.O.R.N. “A. Cardarelli” of Naples to improve its service. In particular, through the application of the AHP we get a hierarchy of needs but not information about user satisfaction, which is obtained with Likert scale. The synergistic application of both methods provided the necessary information to improve the service quality, allowing to identify the most valuable service features and, in case, improve them in order to meet users’ satisfaction. However, a continuous collection of users’ opinions is necessary.

The most frequently used method for hardness testing is the progressive loading test. This test measures surface hardness and requires direct measurement of the indentation’s depth. However, the diameter of the impression left by an indent can also be used to determine surface hardness. The authors designed a method and have custom-written a program for measuring and evaluating surface hardness using a digital microscope and digital microscopic image analysis. The proposed procedure and software are based on the calculation of an impression’s diameter left by an indent from digital microscopic images. The evaluation procedure and the software were tested on Ti and TiN samples of biocompatible surfaces. Measurements demonstrated that the calculated hardness results were largely similar in five different operations for measuring the sample. As a result, it seems that evaluation using surface hardness measured by three test operations appears to be sufficient. The new software and procedures proposed allows for the evaluation of surface hardness in a wide range of metal surfaces, while providing credible results. The advantage of this design is the possibility of using existing cheap systems, which can create an indenter imprint, but do not have the ability to directly determine the hardness. The disadvantage of our solution is the need to use a microscope, but the price of electronic microscopes is very low today and they are commonly available on the market.

Thermal ablative techniques have been used in kidney tumour treatments; these treatments use electromagnetic energy to increase the temperature of the target tissue in order to destroy it. Dielectric and thermal properties influence the deposition of electromagnetic energy and the heat distribution into the tissue, respectively. Accurate knowledge of dielectric and thermal properties permits accurate modelling of the therapeutic results. Extensive research has been conducted on dielectric property characterisation of tissue, while significantly less data are available in literature on thermal property characterisation. The aim of this study is to experimentally investigate the kidney thermal properties in ex vivo ovine models (n = 4) in a temperature range from room to ablative temperatures (95–96 °C). Results show changes in thermal properties at temperatures approaching the water transition to gas, i.e. above 95 °C.

Anthropomorphic phantoms are models of real or virtual parts of the body, organ or tissue, represented by tissue-equivalent materials that aim to provide a realistic and accurate representation of their anatomy and properties. The aim of this study is to evaluate experimentally the suitability of 3D printed materials in the production of both, physical breast phantoms and abnormalities, to be used in optimization tasks in breast imaging. For this purpose, we designed three computational breast models, composed of skin, duct tree, adipose compartments and lesions. Subsequently, they were printed by using two 3D printing technologies and different printing materials, which were previously studied in details. The physical phantoms were scanned at a mammography machine, which allows 2D and 3D mammography (tomosynthesis) modes. The images were evaluated from an experienced radiologist. The results showed that tomosynthesis images are characterized with better realism compared to 2D mammography images. Next step is improvement in the printing quality of tumour formations as well as quantitative evaluation of the obtained results.

In the case of the transplant surgery of ligaments, it is common practice to use tissues from other parts of the body with similar properties. Although they have similarities, there are some distinct properties, some of which may have serious effects on the success of the operation. These properties can be measured and compared to each other so that the surgeon can choose the best option.The goal of the present research is to validate the measured data of the tensile tests of tendons and the theoretical results on fibrous materials, with special regard to the rupture phase. The mechanical properties of the tendon vary significantly, so in the case of a static tensile test, the material’s rupture is not instantaneous, but consisting of multiple steps, as the different groups of fibres are tearing. The main objective is to model this behavior and to compare it with previous results.Our model is based on the results of in-vitro measurements and fibre bundle models describing the behavior of compound fibrous materials. Then by using finite element simulation, the realistic process of the rupture is approximated by a bilinear debonding process. The fibres’ mechanical properties can be set with distribution functions, representing the real variance of the tissue. In the method used, the tensile strength of the fibres is represented by contact stiffness, like static friction, practically separating the fibrous body examined into two parts. With the proper modelling of the fibrous structure’s stochastic behavior, it is possible to validate the previous in-vitro results. The method can be used in other areas, such as the examination of ligaments possessing similar properties, or in the description of the behavior of fibre-reinforced composites.

The detection of the characteristic points of the complex of the impedance cardiography (ICG) is a crucial step for the calculation of hemodynamical parameters such as left ventricular ejection time, stroke volume and cardiac output. Extracting the characteristic points from the dZ/dt ICG signal is usually affected by the variability of the ICG complex and assembling average is the method of choice to smooth out such variability. To avoid the use of assembling average that might filter out information relevant for the hemodynamic assessment requires extracting the characteristics points from the different subtypes of the ICG complex. Thus, as a first step to automatize the extraction parameters, the aim of this work is to detect automatically the kind of dZ/dt complex present in the ICG signal. To do so artificial neural networks have been designed with two different configurations for pattern matching (PRANN) and tested to identify the 6 different ICG complex subtypes. One of the configurations implements a 6-classes classifier and the other implemented the divide and conquer approach classifying in two stages. The data sets used in the training, validation and testing process of the PRANNs includes a matrix of 1 s windows of the ICG complexes from the 60 s long recordings of dZ/dt signal for each of the 4 healthy male volunteers. A total of 240 s. As a result, the divide and conquer approach improve the overall classification obtained with the one stage approach on +26% reaching and average classification ration of 82%.

Electroencephalogram (EEG)-based feature extraction for emotion recognition is a very challenging task. The vast majority of the feature extraction approaches are based on the frequency characteristics of the EEG signals which are extracted using, e.g., the traditional Fourier Transform approach. Lately, a new approach for processing signals in graph domains, namely Graph Signal Processing (GSP), has provided new means of feature extraction by exploiting the oscillations across the vertices of the graph domains instead of the time oriented oscillations of the signals. In this work we are investigating the effectiveness of a fused feature vector for emotional state recognition comprised both of the Graph Fourier Transform and the traditional Fourier Transform features over the DEAP dataset. Support Vector Machines classifier was used in order to classify the extracted features resulting in over 69% and 68% of classification rates for the arousal and valence dimensions, respectively. The fused feature vector outperformed the single GFT- and FT-based feature vectors revealing the need for feature vector approaches that combine graph- and time-based oscillatory features for emotion recognition from EEG.

The hand is a sophisticated enabler of many of our regular daily activities. When it is lost, e.g. due to disease or trauma, it has a significant effect on the amputee’s quality of life. Prosthesis users regularly report the desire for more sophisticated prosthesis technologies that provide sensory feedback to the body, and are more intuitive to use. This would lessen the requirement for visual feedback, for instance to determine if enough pressure has been applied to lift an object. Sensory feedback requires sensors that can respond to different stimulation in real-time. Graphene-based composites have many interesting electrical, mechanical and thermal properties, and have a conductivity that changes with applied pressure, movement (e.g. grasping) and temperature stimulation. We have developed a working proof of concept for a low-cost sensory feedback system using graphene-based composites and commercial-off-the-shelf technology. Prototype graphene- and graphite-based composite sensors, with electrical resistance $$\sim $$ ∼ 50M $$\Omega $$ Ω and $$\sim $$ ∼ 10–100 k $$\Omega $$ Ω respectively, were fabricated using a polyorganosiloxane matrix, and eight ligaments were mounted onto a reduced scale hand with four moving fingers each with three phalanges. Signals from pressure and movement stimuli show characteristic peaks at the start and end of the stimulation, which are not present in temperature stimulation. After stimulation, the signal from each sensor was digitized at $$\sim $$ ∼ 3 Hz and characterised by a bespoke processing, peak-finding and classifying algorithm running on a Raspberry Pi to provide real-time electrotactile stimulation $$\le $$ ≤ 10 mA to the body. When identifying between different stimuli combinations, the algorithm has an accuracy score $$\sim $$ ∼ 95%. In this paper, we outline the synthesis of prototype graphene- and graphite-polyorganosiloxane composite sensors, and discuss the classifiying algorithm used to discriminate between different combinations of stimuli. We present initial results from our upper-limb prosthesis demonstrator, and outline further developments such as introducing the magnitude of the stimulation into the classifying algorithm, and the direct scalable chemical synthesis of other graphene- and graphite-composite sensors.

One of the key issue of health technology management is to evaluate medical equipment lifecycle and to determine the necessity for the equipment replacement in time.Aim of the study: The main objective of this pilot study is to create an approach for assessment of medical equipments and their prioritizing in the decision-making process concerning their purchase for a medical facility within their replacement on the basis of factors effecting the length of their lifecycle.Methods: Methods of expert interviews and multicriteria decision-making were used for prioritization. On the basis of discussions with experts in respective fields of expertise (technical, clinical, user-operational), by using multicriteria decision methods (Fuller´s method), we determined the weights for individual criteria and areas necessary for calculation of the replacement priority index (PI).Results: Methodological approach was tested in a outpatient helthcare facility specialized in radiodiagnostics. 3 medical equipments were tested while all of them was X-Ray equipments: RTG GE Discovery; RTG DRGEM; Opera 500C, U rameno Polistat – M.Conclusion: This easy-to-apply methodology procedure may provide healthcare facilities with relevant information necessary for a decision-making process in the field of medical equipment replacement planning and thus to facilitate better investment planning.

Embedded Sensor Systems for Health (ESS-H) is a research profile where academia collaborates with healthcare organizations and industry with a focus to develop sensor systems for future healthcare. The overarching aim is that health monitoring should be possible to perform anytime, anywhere, using sensor systems for health monitoring and monitoring of humans.A system-wide holistic approach is used, including end-user involvement and close collaboration with companies. This way, user relevance and user acceptance, together with industrial interests, are assured throughout the system design and implementation. The research results have a high potential to become adapted, deployed, and commercialized through this approach.The work in ESS-H is focused within five subprojects:Microwave technology systems, where microwaves are used to measure human tissue, with the aim to detect tumors and strokes.Systems for prevention and monitoring of chronic diseases, where multiple physiological parameters are monitored, and the data is aggregated in order to diagnose and follow health trends. This also includes safe and secure communication, data aggregation and decision support.Vehicle and driver monitoring systems, where the driver environment detects the status of the driver, e.g., regarding alcohol level, attention, and sleepiness.Motion control and analysis, fall prevention, where motion parameters are captured and analyzed, e.g., in order to detect risk of falling or physical activity level.IT-platform for monitoring health at home, where a platform for reliable acquisition of physiological data as well as management and analysis of this data is provided.

Heart rate variability (HRV) is a widely used measure to assess emotional arousal and stress level. HRV is conventionally determined based on the beat-to-beat intervals (duration of heart cycles) calculated as the time difference between successive R-peaks in the ECG signal. However, the heart cycle begins with atrial depolarization, therefore, the onset of the P-wave (Ponset) is a physiologically more appropriate fiducial point to define heart cycles. This paper investigates how the result of HRV calculation changes if the duration of heart cycles is measured using the onset of P-waves instead of R-peaks. Measurements containing ECG signals recorded in Einthoven II lead and one measurement containing simultaneously recorded intracardiac electrograms and surface ECG signals were used. Our results suggest that HRV values characterized by the pNN0_20, pNN20_50, pNN50 parameters may differ by more than 5 percentage points (pp) depending on whether the onset of P-waves or R-peaks are used as fiducial points.

Endothelial progenitor cells (EPCs) play an important role in the pathophysiology of coronary artery disease (CAD). Secretome produced by human Umbilical Cord Blood-Mesenchymal Stem Cells (hUCB-MSCs) shown to have neovascularization and angiogenesis effect. However, its impact toward EPCs proliferation and migration is not yet elucidated. This study aims to determine the effect of hUCB-MSCs-derived secretome on EPCs proliferation and migration capability. EPCs were isolated from peripheral blood samples of one CAD patient and cultured in the Stemline II medium. Cultured EPCs were placed in 6-well plates until it reached confluence and incubated with hUCB-MSCs-derived secretome at a concentration of 2%, 10%, and 20% (v/v). EPCs proliferation was determined using the XTT assay. Migration of EPCs was evaluated using a Boyden chamber assay. Statistical analysis was done using ANOVA test to compare between groups and correlation between variables was obtained using Spearman correlation test. EPCs proliferation at secretome dose 10% and 20% were 1.5–1.8 times higher than the control group (p < 0.05) while EPCs migration was superior at all concentrations (5–15 times higher, p < 0.05) compared to the control group. hUCB-MSCs-derived secretome increase EPCs proliferation and migration in a dose-dependent manner. This study shows that hUCB-MSCs-derived secretome increase EPCs proliferation and migration, making it a potential regenerative therapy for CAD patients.

Rotational electrical impedance tomography provides novel possibilities for multimodal imaging. This could be especially useful in tissue engineering studies where non-destructive and label-free imaging is needed. In difference electrical impedance tomography, the change in conductivity distribution between two samples or states is reconstructed from boundary measurements. Typically, regularization is employed in the solution to tackle the ill-posedness of the problem. The amount of regularization is controlled by a hyperparameter value that is commonly found by subjective and time consuming heuristic selection. In order to find an automatized method that works with rotational data, three state-of-the-art methods for hyperparameter selection were investigated: BestRes, L-Curve and the averaged signal-to-noise ratio ( $$\overline{SNR}$$ SNR ¯ ) as noise performance metric. These were tested with conventional and rotational experimental data. The results show that $$\overline{SNR}$$ SNR ¯ was the only method that provided good image quality with rotational data.

The current health care paradigm clearly identifies prevention as a key element to an efficient disease management. In the context of cardiovascular disease, risk assessment models may be a valuable element in the support to clinical decision, contributing to that preventive care. Moreover, interpretability as well as personalization of risk assessment are decisive aspects to increase the physicians’ acceptance and consequently the respective application in the clinical practice.The proposed work aims to create an interpretable and personalized model based on the proper combination of a set of simple rules resembling the clinical reasoning. Three main steps can be identified in this approach: i) derivation of simple rules based on available risk factors (isolated/combined); ii) identification/selection of a subset of meaningful rules that have more likelihood to contribute to a correct risk classification; iii) ensemble scheme (e.g. voting technique) considering from this subset only the rules that are more adequate for each particular patient. The implementation of this model relied on supervised learning techniques.This methodology was validated with one real patients testing dataset provided by the Santa Cruz Hospital, Lisbon/Portugal, comprising of N = 460 of Acute Coronary Syndrome (ACS-NSTEMI) with an event rate of 7.2% (Death/MI; 30 days). The preliminary results are encouraging, reaching a geometric mean of Gmean = 0.86 for all patients, assuring simultaneously the clinical interpretability and the personalization of the model.

Due to ageing population, old age cognitive deficit is becoming an epidemic-like mass phenomenon. Some form of dementia occurs in 11% of men and 16% of women over the age of 71. Mental wellness is a major factor contributing to the quality of life, therefore, the early detection of deterioration is a very important but hard task. Improving detection would allow ageing at home and more cost effective care. As clinical tests are infrequent and expensive, methods applicable for regular home monitoring have to be developed. Estimation of mental wellness based on voluntary game playing has been investigated in the last years, serious computer games are especially suitable for that purpose. One of the main problems is that popular games are not well fitted for measurement, regular tests are not entertaining enough. The basic was applied in a research project, and it worked well, but some improvements are suggested based on the experience gathered. The most important one is that the combination of the game playing with special tests is proposed for better estimation capability. Consulting with end-users, the participation level and the feedback system is suggested to be improved. Basic considerations, challenges, potential solutions, presentation of preliminary analysis results are described in the paper. The work is performed in the FROm empoweriNg To Viable Living (FRONT-VL) project supported by the Celtic-Plus Programme.

Understanding the functional organization and execution mechanisms of the brain is one of the key challenges of neuroscience. Functional connectivity emerging from phase synchronization of neural oscillations of different brain regions provides a powerful tool for investigations. While the brain manifests highly dynamic activation patterns, most connectivity work is based on the assumption of signal stationarity. One of the underlying reasons is the problem of obtaining high temporal and spectral resolution at the same time. Dynamic brain connectivity seeks to uncover the dynamism of brain connectivity but the common sliding window methods provide poor temporal resolution, not detailed enough for studying fast cognitive tasks. This paper proposes the use of the Complete Ensemble Empirical Mode Decomposition followed by Hilbert transformation to extract instantaneous frequency and phase information, based on which the phase synchronization between EEG signals can be calculated and detected in every time step of the measurement. The paper demonstrates the suboptimal performance of the sliding window connectivity method and shows that the instantaneous phase based technique is superior to it, capable of tracking changes of connectivity graphs at millisecond steps and detecting the exact time of the activity changes within a ten millisecond margin. These results can open up new opportunities in investigating neurodegenerative diseases, brain plasticity after stroke and understanding the execution of cognitive tasks.

In this work, the effects of extracellular environment, in terms of both electroporation (EP) medium and extracellular matrix (ECM), on EP efficiency were evaluated in a 3D in vitro model composed of HCC1954 cells cultured on hyaluronic acid (HA) hydrogels enriched with self-assembling peptides carrying IKVAV motifs. The results from 3D cultures were compared to those derived from cell in suspension and adherent cultures. EP was carried out by using either RPMI (high conductivity medium) and electroporation buffer (low conductivity medium) and applying 8 rectangular voltage pulses at 700 V (electric field strength 1000 V/cm with plate electrode with 7 mm gap). Collectively, our data highlighted that cell organization and the presence of ECM modulate local electrical properties, thus affecting EP efficiency.

Studies and reviews show that the vast majority of students around the world use heavy and uncomfortable backpacks, which could negatively affect their skeletal-muscle development or at least generate a non-physiological functional overload and a change in the kinematics of gait. The purpose of this study is to investigate the role of the school backpack during the execution of the Walk test trying to identify if and how much it affects walking in terms of space-time parameters considering whether it might be correlated to potential spine disorders during the development age. A population-based sample of 98 students (60% female) aging from 10 to 12 years old was studied; gender, age, weight and lower limb length were recorded. Kinematic data were computed using a wearable inertial device for gait analysis: G-WALK System by BTS Bioengineering and analyzed using t-test and Machine Learning. Overall, concerning t-test between free walk and walk with backpack, it emerges a significant statistical difference on 9 out of 10 kinematic parameters, of which 6 with maximum statistical significance (p-value < 0.0001). Machine Learning analysis was carried out through Linear Discriminant Analysis, Naïve Bayes, AdaBoost and Random Forest algorithms considering as different classes: free walk and walk with backpack. Accuracy and ROC Area were considered as evaluation metrics. The best performances were reached with Linear Discriminant Analysis with an accuracy of 85.71% and a ROC Area of 0.92. Study results showed a drastic change on kinematic due to the backpack. These results should be taken in correct account to safeguard children’s health wearing backpack for prolonged periods.

Blood Pressure (BP) and Heart Rate (HR) provide information on clinical condition along 24 h. Both signals present circadian changes due to sympathetic/parasympathetic control system that influence the relationship between them. Moreover, also the gender could modify this relation, acting on both control systems. Some studies, using office measurements examined the BP/HR relation, highlighting a direct association between the two variables, linked to suspected coronary heart disease. Nevertheless, till now such relation has not been studied yet using ambulatory technique that is known to lead to additional prognostic information about cardiovascular risks. In order to examine in a more accurate way this relation, in this work we evaluate the influence of gender on the BP/HR relationship by using hour-to-hour 24 h ambulatory measurements. Data coming from 122 female and 50 male normotensive subjects were recorded using a Holter Blood Pressure Monitor and the parameters of the linear regression fitting BP/HR were calculated. Results confirmed those obtained in previous studies using punctual office measures in males and underlined a significant relation between Diastolic BP and HR during each hour of the day in females; a different trend in the BP/HR relation between genders was found only during night-time. Moreover, the circadian rhythm of BP/HR is similar in both genders but with different values of HR and BP at different times of the day.

The purpose of this study including the questionnaire survey is to perform the colorectal cancer screening behavior support program developed preliminary by the authors and to inspect the efficacy of the program. The subjects were male and female persons who lived in the Kinki area of Japan and were older than 40 years. The subjects were divided as the intervention group and control group. The program was applied to the intervention group. The questionnaire survey was carried out before and after the intervention for all subjects. However, the results of the intervention group are shown and discussed in the text.The ratios of men to women and the youngers (40–59) to the olders (≥60) were about one to one in the intervention group. Almost subjects lived not alone. The proportion of the persons whose screening history is “yes”, increased after the intervention. It is considered that the intervention is effective to raise awareness of screening test and lead to screening behavior. As a whole, women respect time and financial convenience for screening behavior more than men do. The olders (≥60) are concerned about their health, while the youngers (40–59) prioritize work and family. Such differences after the intervention. On the comparison before and after the intervention, the differences were found in “12. I think it is difficult to have the test at early time, because self-screening of colorectal cancer is impossible.”, “17. I don’t like to have detailed test after screening test in which abnormality is found, because I don’t want any medical treatment.” and “29. I have the test, because my family members often talk about the test.”. These results are considered to be caused by the intervention.The results of this study suggest that the screening behavior support program is useful as an educational the intervention for raising awareness on colorectal cancer screening behavior. Further improvement of the program is necessary for more effective the intervention.

The accurate measurement of blood pressure (BP) is essential for the diagnosis and management of several medical conditions. Intermittent non-invasive blood pressure (NIBP) is regularly used in different care units, as it is a readily available method to ascertain BP. This study aims to identify and interpret the current gaps and needs in clinical practice regarding NIBP monitoring.Based on a review of the literature on current BP monitoring protocols, a set of questions were prepared, and 12 interviews were conducted with health professionals. The interviews were followed by shadowing work in the hospital. The results indicate that there are some problems associated with NIBP that can have both human and financial negative consequences. It was found that the difficulty in establishing thresholds for defining clinically relevant changes in BP dynamics and the unreliability of NIBP readings in several clinical contexts, namely hypotension, are the most considerable problems in need of intervention.In conclusion to this work, a list of suggestions to future research is summarized, among which stand out the improvement of NIBP accuracy and NIBP adaptive sampling frequency.

Coronary artery disease is one of the most dominant factors of mortality worldwide. The stenting procedure has been adopted as an effective means of blood flow restoration in the diseased part of the artery. Producing improved clinical outcomes compared to Bare Metal Stents (BMSs), Drug-eluting stents (DESs) have been widely accepted in the field of cardiovascular intervention. In this study, through the BioCoStent in silico stent deployment platform, three different stents (two DESs and one BMS) are comparatively analyzed with respect to the prediction and assessment of their mechanical performance in realistic, patient specific arteries. The results demonstrate the nearly identical behavior for all stents, with minor differentiations. The differences in the results are attributed to the incorporation of the thin drug-eluting coating on the metal core of the stent, eventually inducing an insignificant increment in the post-deployment stresses. Furthermore, the strut peaks, the end rings of the scaffolds and the stenosed part of the arteries experience the highest stresses.

Currently available jaw tracking methods require large accessories mounted on a patient and thus are utilized in controlled environments, for short-time examinations only. In some cases, especially the evaluation of bruxism, a non-restrictive, 24-h jaw tracking method is needed. This study explores the possibility to use a permanent magnet and a 3-axial magnetometer to track mandible’s spatial position in relation to maxilla. An algorithm for determining the sensor’s coordinates from magnetic field values was developed and verified via analytical and finite element modelling of a magnetic field surrounding a magnet and by using a 3D positioning system. The trajectory of natural masticatory movement (10 × 7 × 5 mm) was replicated, and coordinates from sensor data were calculated with RMSE of 0.267 ± 0.023 mm. Estimation of the coordinates of cubic (a = 10 mm) trajectory resulted in RMSE of 0.325 ± 0.009 mm. Errors due to Earth’s magnetic field were shown to increase exponentially along magnet–sensor distance. Possible compensation techniques and sensor positioning possibilities were discussed. Despite the limited working range and large uncertainties in the periphery, the method is increasingly accurate and robust when approaching occlusion.

Sedentary occupation and lack of exercise lead to the gradual muscle wasting and thus shortening of the whole body and also the development of muscle imbalance which involves weakening of muscles, shortening of muscles and impaired spine statics and dynamics in a long-term sitting position. Prevention of muscle imbalance should be aimed at eliminating their causes. Myotonometric parameters were measured using MyotonPro device before and after a physiotherapeutic exercise aimed at strengthening and stretching the deep muscular system. Eight females (age: 56.9 ± 8.4 years) volunteered to participate in this study. Six muscles were selected for measurement on both sides of the body and symmetry index was calculated. Overall, symmetry improved in 119 cases out of 240. In 55 cases the right-left symmetry changed to left-right or vice versa. A graphical user interface has been designed for better understanding and visualization of results. The physiotherapist can choose an individual proband and compare the results before and after exercise. In the next exercise it can focus on individual muscle parts where the symmetry was not achieved.

Timeliness in the treatment of fracture of the femur, through surgery, is crucial in the elderly patient as it reduces the risk of mortality and disability. Here we propose a Lean Six Sigma (LSS) approach to reduce the preoperative length of stay for patients with femur fracture. Through the LSS, a tailored Diagnostic Therapeutic Assistance Path (DTAP) for these has been implemented and monitored over time. In particular, through the analysis, based on the application of the DMAIC cycle conducted on data extrapolated from the information system of the “San Giovanni di Dio e Ruggi d’Aragona” University Hospital of Salerno, the new DTAP was designed and implemented. After the introduction of the DTAP, a significant reduction in the average length of hospital stay was observed, with a preoperative length of stay within 48 h in 65% cases (compared to the previous 9%). In particular, the most significant reduction (over 55%) is obtained for patients aged over 65 years old. Such a result reflects not only the improvement in the care process but it is also compliant with the guidelines of the Italian Ministry of Health, as reported in the New Guarantee System for monitoring the quality of care.

To “put medical device to service” or “placing it on the market” in the EU, the manufacturer shall follow particular legislation. Significant number of start-ups fail to fulfill requirements of legislation for the developed products prior their termination. Many incredible ideas are left behind due to inappropriate approach to the project. Two main technical causes were recognized in this relation: the ignorance of legislation and improper implementation of the development. Nevertheless, fulfilling requirements of legislation to gain possibility of marketing the medical device does not ensure business success. Thus one should make medical device available, where marketing and sales jump in.

The development of non-invasive techniques to assess cardiovascular risks has grown rapidly. In this sense, a multi-lead electrocardiogram (ECG) provides useful information to diagnose myocardial infarction (MI), the leading cause of death worldwide. In this paper we used a support vector machine (SVM) to detect MI by exploiting temporal, morphological and nonlinear features extracted from 12-lead ECG recording from the PTB Diagnostic ECG database. Temporal features correspond to QT, ST-T and RR intervals, morphological features were extracted from P and T waves, and QRS complexes, and nonlinear features correspond to the sample entropy of QT, ST-T and RR intervals. A 10-fold Monte Carlo cross-validation was implemented by randomly splitting the data set into training (70%) and test (30%) sets with balanced classes. Sensitivity of 97.33%, specificity of 96.67%, and accuracy of 97.00% were obtained by jointly exploiting temporal, morphological and nonlinear features by the SVM. The inclusion of entropy favors the detection of healthy control cases because the information of signal regularity improves the specificity of classification.

Intracranial hemorrhages (ICHs) are life-threatening brain injures with a relatively high incidence. In this paper, the automatic algorithm for the detection and classification of ICHs, including localization, is present. The set of binary convolutional neural network-based classifiers with an innovatively designed cascade-parallel architecture is used. This automatic system may lead to a distinct decrease in the diagnostic process’s duration in acute cases. An average Jaccard coefficient of 53.7% is achieved on the data from the publicly available head CT dataset CQ500.

The amount of functional MRI (fMRI) data processed in research is growing, yet no practice or protocol to store them in a lossy format exists. Many researchers are struggling with limited storage space, and speed of common processing tools are often bound by storage speed. In this work, we present a lossy compression framework for fMRI data with user adjustable trade-off between compression ratio and root mean squared error (RMSE). Our goal is to demonstrate the usability of on-the-fly lossy compression for fMRI data. On one hand, the storage footprint and processing speeds both benefit from higher data compression rates achieved with lossy compression. On the other hand, data quality for functional analysis remains effectively the same. With this short demonstration we encourage the research community to develop a lossy data standard for fMRI data.

Diabetes is a complex pathology both for the affected patients and for the medical specialists who follow them. Furthermore, since diabetes is a pathology with a high prevalence and incidence, it is essential to intervene effectively in therapeutic actions through the application of common guidelines. Therefore, in order to improve the management of the diabetic patient, the aim of the work is to define a Diagnostic Therapeutic Assistance Pathway (PDTA). A questionnaire-based approach is adopted for data collection from 136 patients at the Clinical Dermatology Unit of the University Hospital “Federico II”. In most cases (64%) the diagnosis was made by the General Practitioner, 15% of patients obtained the diagnosis at the ASL and 12% at the Polyclinic of Naples AOU “Federico II” and the remaining part from the diabetologist specialist. The second access is generally carried out at the “Federico II” AOU (66%), followed by the ASL (17%), by a doctor specialized in diabetology (12%) while no patient has turned to the General Practitioner for the treatment of diabetes. The final visit is carried out at the “Federico II” AOU in almost cases. The data obtained follow the Italian guidelines: the patients get the diagnosis from the Family Doctor and then they are addressed either to ASL or to diabetologists specialists. For the subsequent visits, most of them prefer to turn to the “Federico II” AOU, especially when they have complications associated with the diseases as they are followed in a more careful and satisfying manner.

Biomedical technologies are the basis of a functioning health system, in particular, medical devices are essential for the prevention, diagnosis, treatment of diseases. However, while developed country hospitals are renewing their fleet of machines by divesting large quantities of biomedical equipment annually, there is a chronic lack of biomedical technology in developing countries to support clinical activities, which could be met by the reuse of used equipment, adapted to the new hospital environment. However, even if the donations of biomedical technologies are generally made with good intentions and not-profit making as in the case under study, obtained results are not what we expected also due to a not perfect communication between donors and recipients and a lack of culture about technology maintenance in the developing countries. At the moment, there is little documented evidence to support these statements. For this reason, the aim of this paper is to quantify the donated medical equipment that are out of service in two different hospitals in Benin. The information was collected on the type of communication existing between donors and beneficiaries and on the type of support that donors provide in terms of staff training, manuals and maintenance. It was observed that more than 50% of the donated equipment is not functional. In addition in more than 70% of the cases the donors do not support the beneficiaries nor training sessions and staff formation are provided. An in-depth assessments of beneficiary structures should be carried out and all donations must be accompanied by initial user training and monitoring by donors regarding the functionality of the system. Donors-beneficiaries communication results as a key elements in the management of health technologies in low-income countries.

This paper aims to summarize and discuss the privacy, security, interactions and safety challenges for the case study of a mobile Health application (mHealth app), for chronic obstructive respiratory diseases, namely AIRDOC - a Smart Mobile Application for Individualized Support and Monitoring of the Respiratory Function and Sounds. mHealth apps create opportunities for improving health outcomes of patients with chronic diseases. However, privacy and security features need to be much improved. Social/behavioural (interactional), technical and legal aspects need to converge to provide a more comprehensive privacy protection when patients interact with mHealth apps. By guaranteeing privacy, more secure, safer and better health results can be achieved, with patients more empowered and in control of their personal health data protection. An embedded security infrastructure, with GDPR compliance and controlling data access, usage and sharing functionalities, together with a continuous risk assessment and enhancement from user and interactions profiling feedback, can help provide useful and long-life trusted mHealth solutions.

The heart is a muscular organ, which pumps blood through the blood vessels of the circulatory system. The left ventricle is the main mechanical element of the human heart, as it receives low – pressure blood from the left atrium and “launches” the blood with high pressure through the aortic pump to the entire circulatory system. Computational modeling is widely used in cardiovascular research. The aim of this work is the computational study of deformation, stress and strain distribution of left ventricle’s wall due to the effect of blood hemodynamics. We employ the Navier-Stokes equations considering fluid-structure interaction using the finite element method. We build initially, the geometry of the left ventricle using computed tomography. The meshing, the creation of a framework and distribution of nodes for a more accurate solution of the given elements are examined through a mesh sensitivity analysis. We apply appropriate boundary conditions which simulate the pathophysiology of left ventricle. In this work the value of stresses are found within the range of values referring to a non-healthy left ventricle and especially in a ventricle after infarction.

Ageing adversely affects most physiological processes, including cardiovascular dynamics. Indeed, vascular ageing plays a major role in several cardiovascular diseases, currently deemed the main cause of death in the world. Vascular ageing produces measurable effects at cardiac level, on major vessels, but also at capillary level. While the effects on the first two components are well known, the dynamics of microcirculation still lacks a complete understanding. In this work, we analyzed the predictability of the pulse waveform at capillary level, using a linear model, aiming at relating the estimated prediction model with the subject age. The pulse waveform was described fitting each pulse with a sum of Gaussian curves, that showed good representation capabilities; each pulse was thus associated with a series of physiologically relevant parameters, derived from the Gaussian curves. For each subject, we estimated a set of auto-regressive models that represent the temporal sequence of each parameter. Then, a classifier was trained to discriminate auto-regressive models according to subject age.Results indicate that, in accordance with literature, the parameter that presents the higher discriminative power is the cardiac cycle duration, with an accuracy in assigning individuals to their age group of 90%, and an area under the ROC curve of 0.915. However, we observed also good performances by modelling the sequence of pulse amplitudes.

Amputation of a lower limb implies a re-organization in the strategies used to reach a stable walking pattern. The planar covariation law of elevation angles is a well-defined low-dimensional description of the kinematics of movement; according to this law, thigh, shank and foot elevation angle co-vary on a plane, defining a typical gait loop. However, a robust biomechanical interpretation of the outcomes of its related analysis is still missing. In this work, we tested the planar covariation law on a group of 14 trans-femoral amputees, comparing the results with the ones related to 12 healthy people. Moreover, by adopting a common covariance plane for all the subjects, we checked whether the projection of the original elevation angles on this plane is able to yield biomechanically meaningful information on the control of prosthetic gait. A common plane was able to describe the coordination of lower limb elevation angles in all subjects; on this plane, most of the differences among populations were identified on the time course of one principal component.

The economic cuts suffered by public health have in many cases led to the reduction of beds. In order to optimize the available resources, the length of stay (LOS) can be used as an efficiency parameter. The objective of this study is to predict the value of LOS using the clinical information that is generally supplied by a patient who is hospitalized following a fracture of the neck of the femur and to make a comparison with results obtained after the implementation of the new diagnostic-therapeutic-assistance pathway (DTAP). The analysis was conducted on data extrapolated from the information system of the University Hospital “San Giovanni di Dio and Ruggi d’Aragona” of Salerno (Italy). The results show promising outcome in the use of the proposed prediction models as a tool for determining an estimate of the LOS and support the decision making process and the management of hospital resources in advance. In addition, the comparison of between the two models can be used as an indicator to assess the efficiency of the implemented DTAP.

We assess the performance of previously introduced jDErpo differential evolution (DE) algorithm in the decomposition of Compound Muscle Action Potential (CMAP), recorded by High-Density Surface Electromyograms (HDEMG) and show that multiple runs of this stochastic algorithm results in the dispersion of Motor Unit Action Potential (MUAP) position estimates across different runs and in false exclusion of MUAPs in some of the runs. We then introduce a procedure that optimally combines the results of multiple runs and show that this additional optimization significantly improves the consistency and accuracy of the CMAP decomposition by jDErpo. On average, in noiseless signals, the proposed multi-run optimization correctly identified the firings of 31.6 ± 6.3, 33.4 ± 7.0 and 34.5 ± 10.3 MUs on simulated CMAPs with 136, 168 and 200 active MUs, respectively. When 20 dB Gaussian noise was added to the signals, the method identified 31.2 ± 20.4, 33.9 ± 27.5 and 35.0 ± 26.3 MUs, on average.

This article describes the application of nonlinear methods in the form of Multiscale Poincaré plots for the purpose of evaluating the variability of the horizontal amplitudes of nystagmus during bilateral stimulation on Bárány chair. This new approach can provide additional information for the needs of evaluating vestibular system disorders. The evaluated parameters were areas of the 95% confidence ellipses of amplitudes per- and post-rotational nystagmuses. 3D VOG glasses were used to assess the vestibulo-ocular reflex induced by the trapezoidal acceleration stimulation. A total of 10 volunteers without neurological disorders involved in the pilot measurement (5 males: mean age 23.4 years, 5 females: mean age 22.8 years). Jarque-Bera test did not show normal distribution of the data and results did not confirm the difference between ellipses surfaces per- and post-rotational nystagmus amplitudes, even for different coarse grained parameters Σs (s = 1–4). For further research, it would be advisable to focus on lower coarse grained Σs parameters.

The aim of the present study was to analyze the upper trapezius activity, and its relationship with muscular discomfort perceived, during prolonged smartphone texting. Seventeen healthy young subjects participated in the experiment; they were asked to use their own smartphone for texting (10 min), maintaining two different postures, sitting and standing. The muscular activity of the right and left upper trapezius was recorded, and the CR10 BorgScale was administered after each experimental section. To normalize the EMG signals, the maximum voluntary contraction was acquired at the beginning of the experiment. The median, the 10th percentile (P10) and the range (difference between 90th and 10th) of EMG RMS, the relative rest time (RRT), the correlation between P10 and RRT and between P10 and CR10 scale were calculated. The results showed no statistical difference between the postures, and the body side. The value of RMS parameters was around the 2% of MVC, showing a constant muscular activity throughout the experimental section. A significant negative correlation between P10 and RRT suggested that the subjects with greater P10 showed a lower rest period; moreover, the significant positive correlation between P10 and CR10 Borg scale, for both postures, suggested that the subjects with high P10 values perceived greater discomfort in neck and cervical zone. The results support the hypothesis that the prolonged use of smartphone for texting influences the upper trapezius activity and it is strictly linked with the absence of a period of muscular recovery, and with the perception of muscular discomfort: that may be a potential risk factor to develop neck pain and musculoskeletal cervical disorders.

In the current article we do report anthropometric measurements of 107 Bulgarian women performed by our team, that complement the representative anthropological investigation (Yordanov et al., 2006) of 2855 Bulgarian females aged 30–40 years. Based on these data, we present an approach for evaluating the mass-inertial parameters of the thigh, shank, upper and lower arm via using 3D geometrical modelling. We model the segments with a geometrical body, called right elliptic stadium solid, having cross-sections of elliptic stadium. The comparison of our model results with data reported in literature shows that the suggested modelling is successful since it uses 3D bodies closer to the actual shape of the foreseen segments. The results obtained could be used for the design of systems and devices aimed to enhance the rehabilitation, to help disabled persons and in criminology to evaluate body fall problems, ergonomics, sports, robotics, computer simulations and other fields of research.

In order to place a medical device of higher risk on the market or put it into service a manufacturer has to undergo a conformity assessment procedure at notified body. During the procedure deficiencies can be identified which has to be solved prior to issuing an EC certificate. With the statistical analysis of the conformity assessment procedures performed at SIQ as notified body in last 4 years we have found that knowledge of legal requirements and associated guidance, involvement of expert knowledge of specialists from different disciplines including medical doctors, thorough study of scientific literature prior design of product starts reduces number of deficiencies in the conformity assessment procedure. Number of deficiencies is correlated with the time needed from first audit of technical documentation to the EC certificate.

This paper is about electroporation using an endocardial catheter. It is focused on numerical modeling of the electroporation process inspired by in vivo experiment. Two electroporation models are compared; model with electric conductivity dependent and independent on the electric field intensity. Further, it also demonstrates the effect of the flowing blood on obtained results. The electroporation effect in the homogeneous tissue is symmetrical unlike the effect in the inhomogeneous tissue. Blood flowing around the application electrode cools down the tissue but takes also majority of the current. The treatment has to be either longer or use higher voltage. The temperature was 40.7 °C for the model with electric conductivity dependent on the electric field intensity vs 45 °C with independent electric conductivity. Comparison of the simulated and measured data is not part of this paper.

Use of heterogeneous biosensors in microfluidic channels for diagnostic and detection of diseases in early stages is a novel, relatively cheap and applicable solution for saving lives and health purposes. It has attracted great interest in experimental field and a lot of works are being done in order to reach better biosensors with reliable results and fast detection. A comprehensive numerical model of Langmuir-Hinshelwood mechanism would help speeding up the process of design and analysis of biosensors. For this means we have developed a code in Fortran to simulate flow inside a microfluidic channel combined with Langmuir-Hinshelwood reactions on the surface of the biosensor. Control-volume based finite-element method (CVFEM) with high order discretisation has been used to solve full Navier-Stokes equations with chemical reactions on the biosensor. Results has been validated with existing experimental data from literature. The influence of target concentration as well as the inlet velocity (flow rate replica) on the biosensor saturation time, moreover, investigated in this study. The results show that as the inlet velocity and concentration increases the time that concentration on the surface of biosensor reaches its asymptotic value decreases, although increase in inlet velocity does not have any effect on dissociation stage of binding cycle.

In recent years, many new decomposition methods for identification of motor unit (MU) firings from high-density surface electromyograms (HDEMG) have been developed, with recent attempts focused on the use of different neural networks (NN). In this study, we evaluated the need for HDEMG signal whitening in NN-based MU identification. For this purpose, we analyzed the learning efficiency of two different types of NN, namely dense NN and long short-term memory (LSTM) NN, on the same HDEMG signals, with and without spatial whitening applied to them. All the HDEMG signals used were simulated with advanced HDEMG simulator, providing a full control of MU firing patterns and MU characteristics in our test environment. Spatial whitening of HDEMG signals significantly improved the precision of MU identification, regardless of the type of NN tested. For dense NN, precision of identified MU increased from 32.2 ± 20.2% to 93.1 ± 8.7%, whereas miss rate decreased from 48.4 ± 23.9% to 12.0 ± 13.3% when whitening of HDEMG signals was employed. For LSTM NN the precision of MU identification increased from 59.7 ± 19.7% to 99.4 ± 2.0% whereas miss rate decreased from 43.1 ± 22.3% to 12.7 ± 9.7% with whitening.

We test the reuse of motor unit (MU) filters estimated by the Convolution Kernel Compensation (CKC) method from high-density surface electromyograms (HDEMGs) with different signal-to-noise ratios (SNRs). During the learning phase the MU filters are extracted from HDEMGs with four different SNRs, namely $$\infty $$ ∞ dB, 30 dB, 20 dB and 10 dB. The MU filters are then applied to HDEMG signals at the different SNRs, yielding the MU spike trains. We report mean precision and miss rate of MU firing identification. In order to test the sensitivity of MU filter learning to the length of the HDEMG signals, we repeated the experiment at 5 s and 15 s long learning sets of HDEMG signals.The number of identified MUs decreased from about 12 MUs, when using MU filters learned on 15 s long HDEMG signals with SNR of $$\infty $$ ∞ dB, to about 3 MUs, when using filters learned on HDEMG signals with SNR of 10 dB, no matter how much noise was present in the MU filter application phase. However, if there was no or little noise present in the MU filter learning phase then a decrease in precision and an increase in miss rate was observed when MU filter was applied to the HDEMG signals with a lot of noise. The opposite was true when large level of noise was present during the MU filter learning, but no or little noise was present in the MU filter application phase.

Big data and artificial intelligence-based researches in the health care arena have radically changed the sector with better preventive health care, early diagnosis of diseases, and advanced assistive technology along with numerous other areas. Health care facilities, academic research centers, and industries are collaborating in developed countries on such researches. Besides, developing and underdeveloped countries stay behind in this field of research due to infirm health and e-health infrastructure, insufficient technical manpower, low physicians to patient ratio, and other limitations. Our research focuses on developing an open-source and easy to use Machine Learning Software System that should uplift Big Data and data science researches focusing on health care in the developing and underdeveloped countries amid such obstacles. This pilot study is a part of that big project that helps to make sense about the working methodology and the expected outcomes by the end of the project. Apart from medical data analysis, it could serve as an efficient platform for storing patient data and we hope academicians, professionals, and physicians around the globe will be aided by such robust data analysis software, as it facilitates automated preprocessing of data, building and comparing different prediction models, cloud storage and data visualization techniques. This work visualizes most of the part of its concept to understand its facilities, although due to some restriction some techniques will be discussed only after completing this big project.

Optical projection tomography (OPT) is a powerful technique for imaging in developmental biology. It is similar to X-ray computed tomography where shadowgrams of rays transmitted through sample are recorded and, based on these shadowgrams, the internal structure of the sample is reconstructed. In OPT however, light is used instead of X-rays, which provides practically negligible effect to the sample in many cases. OPT can also be used in fluorescent mode, where emission of the excited fluorescent markers is imaged. The optical instrumentation, consequently, imposes blurring into sample details outside the focal plane of the imaging objective. To increase the quality of the tomographic reconstructions, we incorporated light beam model into the reconstruction process, both in transmission brightfield and in fluorescent emission modes. In this work, we quantitatively compare the performance the new models with that of conventional filtered backprojection. Based on our results, the incorporated light models and filtered backprojection perform close to each other. Noise-reduction improved the quantified measures in filtered backprojection case when 400 projection angles were used. We provide the related data and codes.

Pancreatic cancer is a highly lethal disease, projecting to be the second leading cause of cancer-associated deaths. It is considered as one of the most aggressive types of cancer, with one of the major problems reported being the lack of early detection. A patient is diagnosed with pancreatic cancer only in advanced stages, when the possibility of developing a metastasis is high. There is no standard procedure to diagnose high risk patients, since they remain asymptomatic in the cancer’s early stages. Based on the accumulated evidence revealing remarkable parallels in key biological signaling pathways that govern embryonic development and cancer, we sought to extract significant genes at the intersection of these two processes, aiming to identify new tumor markers for pancreatic cancer. Specifically, the aim of this work is to apply machine learning methods to identify biomolecular markers that are differentially expressed in pancreatic cancer patients and correlate them with markers from embryogenesis. After extracting such markers, we use them as predictors within different machine learning methods. Our work contributes a “25 gene signature” of biomolecular markers, which are involved in signaling pathways found in both embryogenesis and pancreatic carcinogenesis, obtained via feature extraction and feature selection methods. These markers are used in classifiers for pancreatic cancer classification and two machine learning models are tested, with good results. We finally justify the notion that our “25 gene signature” can play a classification role in discriminating patients with pancreatic cancer from healthy controls.

Patient Generated resources have been proved to be key for the future of the medicine. Thus, the acquisition of these resources must be as smooth and transparent as possible to ensure that patients collaborate on their collection. However, aligning usability with the restricted security and privacy requirements in healthcare information systems might not be an easy task. In this work we propose a two-step authentication and authorization workflow, that relaying on state-of-the-art standards, performs a seamless EHR access to store the data gathered during the patient daily follow-up. It combines a long-term login using strong authentication mechanisms with a just-in-time authentication and authorization that takes advantage of the identifying capabilities of some signals that are usually recorded by patients. A new OAuth grant type has been designed to that end. We finally discuss about the usability and security of the proposed approach and conclude that easing the data collection it is expected to increase patient’s implication with the follow-up.

Correct posture matters in everyday activities, without it one risks pathological deformities or fall injuries. Military professions consist of wide variety of specific duties, which can have specific impact on physical state. Therefore, it might be useful to monitor each military group for possible occupational disease. This study investigates stabilometric parameters in different groups of military professions with a hybrid transportable system designed and presented in previous study. The system comprises two Microsoft Kinect cameras and two Nintendo Wii Balance Boards, and it is used for evaluation of posture by measuring postural stability parameters, such as mutual position of spine vertebrae, center of gravity and center of pressure. Comparing different military professions, in this study particularly Airborne Troops, Castle Guards and Mechanized Infantry, statistically significant differences were found in center of pressure for anterior-posterior direction and in some of the mutual positions of defined anatomical points.

This article presents a preliminary evaluation of a novel language independent Speech-in-Noise test for adult screening in terms of Speech Reception Threshold (SRT) estimates and prediction of hearing sensitivity.The test is based on multiple-choice recognition of meaningless Vowel-Consonant-Vowel words and was administered to 26 normal hearing young adults and 58 unscreened adults who also underwent pure-tone audiometry. Receiver operating characteristics were built using the World Health Organization criteria for “slight/mild” and “moderate” hearing loss as gold standards and SRTs as test outcome. Both curves showed very good test performance in predicting success/failure in pure-tone audiometry (area under the curve: 0.79 for “slight/mild” and 0.83 for “moderate” hearing loss). A complete generalized linear model including SRT, age, and their interaction showed that the SRT and the interaction between SRT and age were significant predictors of pure-tone audiometry outcomes, whereas age alone was not a significant predictor of the degree of hearing loss. Moreover, preliminary results from test-retest data showed that the test was reliable in repeated measures (Spearman’s rank-order correlation coefficient = 0.72; Cohen’s kappa = 0.83 for “slight/mild” and 0.64 for “moderate” hearing loss). Further research is needed to fully assess test performance in a larger sample of participants, also including subjects with higher degrees of hearing loss (e.g. “severe” and “profound”).

Loss of vestibular feedback is considered a significant disabling factor among older adults. Studies show that prosthetic vestibular feedback in the forms of electro- and vibrotactile mechanisms expedite the treatment and recovery of vestibular feedback loss. We propose an alternative vibrotactile system in a hat that provides vestibular feedback to the skull based on signals from an accelerometer mounted on the hat. This work presents a proof of concept of our proposed wearable system and evaluates the balance performance of two healthy volunteers before and after wearing the hat. The balance performance evaluations were based on the Sensory Organization Test (SOT) protocol. Quantitative comparisons of balance performance show that our proposed system changes balance performance after wearing the hat for a short training session. The proposed system is easy to use and comfortable.

This paper presents a method for automatic pulmonary crackle detection based on the Hilbert energy envelope (HEE). Automatic detection of crackles in lung sounds offers a non-invasive way of monitoring or diagnosing cardiopulmonary diseases. The algorithm is divided into four main steps: (a) preprocessing, (b) estimation of HEE, (c) thresholding, and (d) applying time width conditions based on crackle two-cycle deflection and initial deflection width. Its performance is tested using a publicly available lung sound dataset of fine and coarse crackles and evaluated by the sensitivity (95.7%), positive predictive value (89.5%), and F-score (91.7%) for crackle detection. The good detection performance indicates the potential of the HEE-based algorithm as an automatic method for crackle detection in lung sound recordings.

Gait is an essential movement and has been shown to be a relevant measure for differentiating gait pathologies and neurological conditions in humans as well as in animals. Inertial measurement units have been suggested as a promising tool for gait analysis. Gait analysis performed in pre-clinical animal models can improve the conversional reliability of preclinical research. Large animal models can confirm and augment results achieved in rodents prior to adaptation to humans. Because pigs are of similar body size to humans and their brains are more like humans than rodent brain, pigs are a more direct assessment of dosing in a preclinical model. Pig gait analysis is used to characterise the pathologies of motor control and to evaluate the effectiveness of treatments performed previously in clinical settings. Nowadays, there is no information on the reliability of large animal model gait signals, namely pig gait signals. This paper presents the pilot analysis of gait angular velocity and acceleration provided by inertial sensors placed on the front shoulders and tests them for intra-individual reliability. An intra-class correlation was employed to analyse inertial sensor signals from three healthy pigs. Most of tested pigs performed with good reliability for roll and pitch angular velocity, and vertical and medio-lateral acceleration. Therefore, we can recommend these signals for the basis in of continuous signal analysis.

Placenta previa (PP) and Placenta Accreta Spectrum (PAS) are obstetric pathologies whose early detection is fundamental for an appropriate patient management. In this paper, ultrasonography (US) is performed on 53 patients and from the images a texture analysis feature extraction is performed through PyRadiomics. The US images were acquired with 3 different resampling resolutions: 1 × 1, 2 × 2 and 3 × 3. The features extracted from the images at each resolution were used to investigate which one is the best to make the correct diagnosis by employing machine learning techniques. Knime analytics platform was employed to implement decision tree, k nearest neighbor and naïve Bayes. Synthetic minority oversampling technique was used to balance the dataset and some evaluation metrics were computed after a leave one out cross-validation. Averaging all the metrics among all the algorithms, 1 × 1 resolution achieved the best mean accuracy (75.97%), sensitivity (83.33%), specificity (68.50%) and Area Under the Curve Receiver Operating Characteristics (0.81). Moreover, k nearest neighbor was the algorithm with the highest metrics (greater than 80%). Despite using also artificial data to balance the dataset (less than 30% of total analysed sample), this study provides researchers with the idea that employing a 1 × 1 resolution could be the best option when analysing images with machine learning algorithms on texture analysis features US-derived.

The electroencephalography (EEG) source imaging problem is very sensitive to the electrical modelling of the skull of the patient under examination. Unfortunately, the currently available EEG devices and their embedded software do not take this into account; instead, it is common to use a literature-based skull conductivity parameter. In this paper, we propose a statistical method based on the Bayesian approximation error approach to compensate for source imaging errors due to the unknown skull conductivity and, simultaneously, to compute a low-order estimate for the actual skull conductivity value. By using simulated EEG data that corresponds to focal source activity, we demonstrate the potential of the method to reconstruct the underlying focal sources and low-order errors induced by the unknown skull conductivity. Subsequently, the estimated errors are used to approximate the skull conductivity. The results indicate clear improvements in the source localization accuracy and feasible skull conductivity estimates.

Tongue cancers are among the most frequent malignancies in the population and their influence can be affected by many risk factors. Patients undergoing tongue surgery face different complications and can experience a long length of hospital stay (LOS). The aim of this paper is to compare two pharmacological therapies in order to understand which one decreases the LOS. At the University hospital of Naples “Federico II” two antibiotics were employed: Cefazolin plus Clindamycin and Ceftriaxone. Six Sigma methodology was employed to analyse two group of patients treated with these two different antibiotics: 55 patients treated with the antibiotic Cefazolin plus Clindamycin and 66 patients with the antibiotic Ceftriaxone. This is the first time that this methodology is used in order to compare two antibiotics in the oncology field. The results obtained show clearly and with a statistical evidence that patients treated with Ceftriaxone experienced a lower LOS (−28.6% in terms of percentage between medians). Reducing the LOS for patients means limiting the number of complications and, therefore, reducing the hospitalization costs. It would be valuable for both hospital and patients: the former would save money that they could invest in other important care activities; the latter would experience a higher quality of care with fewer complications.

Stroke is a leading cause of death and remains a major healthcare burden worldwide. Effective rehabilitation strategy is required to improve motor impairment and functional status of stroke survivors. The imagination of movement is one of the methods that can be used in the therapy of stroke survivors at home to gain full recovery. This paper describes the spectral analysis and 2D topography of EEG signals obtained during actual and imagined hand twisting for stroke rehabilitation. The EEG signals were recorded from thirty-two channels, processed and filtered to remove the unwanted signals. The signal features were then extracted using power spectral density and analysed through EEG 2D topography. The results showed that monitoring the status of brain region during actual and imagined twisting could be performed using eight electrodes. The EEG topography revealed a suitable frequency range to monitor the status of the brain activation area for both cases.

Video-based surgical tool detection is an important yet challenging problem for developing context-aware systems (CASs) in the operating theatre. In this paper, we address tool presence detection in laparoscopic videos using a combination of convolutional neural network (CNN) and Long-short term memory (LSTM) network. Firstly, a pre-trained CNN model was fine-tuned to learn visual features from laparoscopic images. Since the data is sparsely labelled, an LSTM network was then employed to learn temporal dependencies from short sequences of adjacent frames. Several experiments have been conducted with the Cholec80 dataset to validate the proposed framework and investigate the effect of the video clip length on tool prediction performance. Results demonstrate the advantage of employing temporal information to the tool detection task and show the most notable improvement (mAP of 91.1%) is achieved when sequences of previous and next frames were employed.

The swallowing process involves the coordinated activation of several muscles to ensure the transfer of nutrients from the mouth to the stomach. A proper segmentation of swallowing into its constituent phases is relevant to obtain a quantitative biomechanical and electrophysiological description of this sensorimotor task.The aim of the study was to design a non-invasive measurement framework integrating electromyographic and acceleration measurements to detect the swallowing onset and event-related muscular symmetry indexes during the oropharyngeal phase. Therefore, the experimental protocol included: surface electromyography (sEMG), accelerometry and Fiberoptic Endoscopic Examination of Swallowing (FEES) as a clinical gold standard. A comparative study on five healthy subjects was performed in order to evaluate the results of the accelerometer-based segmentation with respect to those obtained through the gold standard.Results showed that the accelerometer-based method consistently underestimated the swallowing onset (204 ± 192 ms, mean and standard deviation). Despite this bias towards the onset estimation, sEMG symmetry indexes computed from the accelerometer- and FEES-based onset exhibited comparable values.These preliminary results suggest that the observed underestimation is not relevant in order to study symmetry differences in swallowing muscular activation. Thus, the acceleration measurements can provide a possible non-invasive alternative to the FEES-based segmentation for the extraction of event-related symmetry indexes during the oropharyngeal phase of swallowing.

In strategic plan 2014–2019 the Tanzanian Ministry of Health and Social Welfare identified “Human Resource for Health” to be the key component for “delivery of quality health and social welfare services, with the ultimate goal of having effective health services in a dispensary at every village, a health center at every ward and a district hospital at every district”. Another five year development plan from 2016 announced the construction of 67 new health facilities, in part improving existing district hospitals. In 2019 the government gave notice to end the public private partnership (PPP) with faith based (district) hospitals (at least 13% of Tanzanian hospitals). - District hospitals are supposed to have a minimal total staff 200 persons with a low number of (biomedical) technologists. The lack of human resources for health, particularly doctors, is nothing new. Therefore, district hospitals hardly can find the recommended number of doctors: 75% of the Tanzanian population lives in rural areas, 26% of doctors serve in rural areas. This situation mirrors in respect to clinical and hospital engineers: in 2017 the Tanzanian minister of education reported a shortage of 7000 biomedical engineers. A substantial relief is hardly to be expected in the near future: the gap between the governmental efforts and reality appears evident. In respect to biomedical engineering, extensive knowledge transfer, supervision and training to increase the provider’s skills will become more important than financially supporting the installment of technologically most advanced medical equipment in an inadequate infrastructure of district hospitals.

Deep brain stimulation (DBS) is an intervention used for several neurological conditions such as Parkinson’s disease. To evaluate the clinical response in relation to anatomical location, electric field simulation using the finite element method is commonly used. The models presented in different studies are varying in complexity and this study aims to evaluate the effect of including anisotropy in the tissue model using homogenous tissue with varying level of anisotropy both parallel and perpendicular to the DBS lead. As a benchmark, data from one patient was included and simulations was performed in zona incerta (Zi) and the internal capsule (IC). The parameters investigated were impedance, volume within the 0.2 V/mm isosurface, radial and longitudinal expansion as well as visual representation of the isosurface. The investigations show that both the impedance and volume are increasing with increasing anisotropy together with the electric field isosurface in the principal direction of the anisotropy. When comparing different stimulation modes, current control (CC) stimulation had a steeper increase with increasing anisotropy for all parameters compared to voltage control (VC) stimulation. This could be due to a joint effect of the anisotropy and the increasing impedance. The result from the patient simulations are in the anisotropy range where simulations from the homogenous models starts to have a higher slope for all parameters. This indicates that including anisotropy in computer models will be of importance in areas of high anisotropy.

In this paper a novel head-foot wheelchair steering system based on force sensor arrays (FSAs) for people diagnosed with dyskinetic cerebral palsy (DCP) is introduced. The user applies pressure on FSAs placed on the head, and foot supports of the electrically powered wheelchair (EPW), based on his/her intention to accelerate, brake, steer right or left. The microcontroller-based electronic system acquires and translates the mean voltage generated by the applied force into wheelchair control signals. In such a system, FSAs are integrated into the head support of the wheelchair using support materials and textiles for the comfort of the user, having an effect on the sensor readings. This work aims to explore the effect of integrating FSAs into the head support of the wheelchair using support material and textiles. Four different sensor integration approaches were examined and compared to baseline readings of a non-integrated FSA. It was found that when a maximum force of 80 N is applied to a single sensing element (sensel) the support material decreases the mean voltage by approximately 70%, and the sensor integration into textile has shown increases between 5.2 and 14.9% compared to the support material. Furthermore, when force is applied over multiple sensels, the support material accounts for a reduce in the mean voltage ranging from approximately 3 to 32%. The addition of textile has exhibited peak decrease in the mean voltage up to roughly 41% for the tested integrations.

In this study we used the bipolar surface electromyography to investigate whether a passive exoskeleton reduces the degree of activity of shoulder muscles. Twelve young healthy volunteers participated in the study. Subjects were asked to hold four different static postures: (P1) shoulder abducted at 90°, elbow flexed at 90°, elbow pronated at 90°; (P2) shoulder flexed at 90°, elbow flexed at 90°, elbow pronated at 90°; (P3) shoulder flexed at 90°, elbow pronated at 90°; (P4) shoulder abducted at 90°, elbow pronated at 90°. Subjects maintained each posture for 20 seconds five consecutive times, with a rest time in-between of 20 seconds. Surface EMG signals were collected from anterior, medial and posterior deltoids and upper trapezius muscles.Our main statistical results showed a significant (p < 0.05) attenuation effect of exoskeleton on the RMS amplitude computed for all muscles evaluated, though not for all postures. For the anterior, medial deltoids and upper trapezius a lower level of activity was observed in all postures with than without exoskeleton, while for posterior deltoid only for P2-P3 and P1-P4 respectively.These findings suggest the passive exoskeleton evaluated in this study attenuates the shoulder muscles’ effort during static work-related tasks, with implications on the prevention of musculoskeletal disorders.

Among all possible postures of the human body, NASA selected eight of principle importance for space exploration. The current article aims to determine the mass-inertial characteristics of the human body of the average Bulgarian male in one of these positions—the so-called relaxed, or weightless position. We determine the corresponding characteristics of the centre of mass and principal moments of inertia using a 16-segmental biomechanical mathematical model that is generated within the SolidWorks environment. We verify the model by comparing the analytical results for each of the body segments with the results obtained using the computer model. The geometric data needed for the construction of the 3D model are taken to be in correspondence with experimentally available anthropometric data for about 2500 Bulgarian men. One their basis one determines the characteristics of the average Bulgarian men. Then, using the CAD realization of the model the inertial parameters of this “average” male in different body positions can be determined. The comparison made between our model results described in this article and the data reported in the literature, where available, gives us confidence that the suggested model can be used to calculate the properties in question at any postures of the body of interest. In principle, our approach can be also used to calculate the corresponding mass inertial data for any individual provided that the anthropometric set of parameters for that individual are measured. The model we used is suitable when one needs such parameters in problems appearing not only in space exploration with the participation of male astronauts but also in rehabilitation, sport, criminology, robotics, etc.

The possibility of using simple and effective models to estimate the patient’s length of stay in intensive care units is decisive to support the clinical professional decisions. These models can help professionals in the stratification process and, particularly, in the identification of the necessary intervention plan to improve the patient’s health condition. In clinical practice specific prognostic scores are available and validated in the cardiovascular context. These risk tools address the primary prevention domain, as well as the secondary prevention domain, usually involving long-term (years) and short-term (months) prediction periods, respectively.The aim of this study is to investigate the capacity of available prognosis risk tools, in particular SCORE (primary tool) and GRACE (secondary tool), to estimate the length of stay in a cardiac intensive care unit. For validation purposes a dataset collected by the Centro Hospitalar Universitário de Coimbra was used, consisting of approximately 1400 patients that have been admitted into the cardiology intensive care unit. The obtained results suggested that SCORE and GRACE models are not sufficiently accurate to estimate the actual length of stay. Moreover, GRACE presents better results than SCORE, which can be justified by the employed risk factors, more specific for short-term prediction periods.

Breast reconstruction is fundamental and urgent for patients in order to avoid future psychological and physical issues. That’s why immediate breast reconstruction has been requested increasingly in the last years. In this study two prosthesis with different structures and properties were compared according the aesthetic appearance (BREAST-Q© was employed) and five complications (seroma, hematoma, infections, dehiscence and red breast syndrome). The overall population was composed by 56 patients: 24 received a Tutomesh prosthesis and 32 received a Surgimend prosthesis. The DMAIC (define, measure, analyse, improve and control) cycle was implemented as a problem-solving strategy of the Six Sigma to compare the prostheses. While statistically significant difference between the two groups wasn’t found according to the overall BREAST-Q© (p-value = 0.674), the number of complications of the two groups resulted statistically different (p-value of chi-square test less than 0.001). Although it is not possible to understand from this study the reasons of the differences between the complications, this research proved that Surgimend and Tutomesh prostheses can be both implanted safely for immediate breast reconstruction since the higher costs of Surgimend could be neutralized with its lower hospitalization compared to Tutomesh.

Caries and periodontal disease are considered a primary cause of tooth loss and extraction. Implant dentistry has improved the rehabilitation of edentulous patients providing a 10-year success rates of over 97%. However, the insertion torque, the superficial characteristics of the implants, and the heat generated during implant site preparation could represents a critical factor for early implant failure. Hence, monitoring the temperature during the insertion could be fundamental to predict the probability of success of the prothesis. Although several studies investigated the thermal effects of drilling and fixture placement, a comparative study between the thermal outcome of the insertion of implants with different diameter is missing. The objective of the study was to compare thermal changes, evaluated through infrared thermal imaging, induced by the insertion of narrow (3.0 mm x 10.0 mm) and regular (4.5 mm x 10.0 mm) implants in an animal bone model (swine ribs). An increase of the bone temperature was found for both narrow and regular implants. Moreover, a higher thermal effect was found for the narrow with respect to regular implants (p < 0.05), but always lower than the temperature limits of the bone necrosis. Although preliminary, these results confirmed that narrow implants are thermally and clinically safe.

Carotid artery disease is considered as the pathological disease of carotid arteries and is considered as a principal cause of stroke. Therefore, early diagnosis of carotid artery disease is of high clinical importance. This study aims to present an overall methodology for the accurate identification of the inner wall, outer wall and the atherosclerotic plaques (calcified and non-calcified) of the carotid arteries. The proposed methodology is based on a level set based approach, which is fully adapted in each computed tomography acquisition protocol. Briefly, the methodology includes the following steps: (i) the estimation of intensity membership functions for the inner wall, the outer wall and CP, (ii) the carotid artery centerline extraction, (iii) the inner wall, outer wall and calcified plaques segmentation, (iv) the noncalcified plaques segmentation and finally (v) the 3D models construction. The segmentation accuracy of the proposed methodology has been validated against manual expert’s annotations in 4 patients, and more specifically in 300 computed tomography angiography slices for the inner wall segmentation and in 30 slices as far as the atherosclerotic plaques is concerned. The utilized evaluation metrics were the Dice coefficient and the Hausdorff Distance and our very first results are promising for the accurate and automated segmentation of carotid arteries.

The transmembrane voltage induced by reversible electroporation protocols in cell suspensions and adhered cells is reported in this work for five cancerous cell lines and one non-cancerous cell line. Therefore, cell-specific morphology was evaluated. Realistic three-dimensional models of adhered cells were reconstructed from sequences of confocal microscopy images. The induced transmembrane voltage (ITV) was analytically and numerically determined for cell suspensions and adhered cells respectively. Cells at a tissue level show a very irregular shape, and therefore the models of adhered cell conferred a closer representation to the scenario that may be found in vivo, and the calculation of ITV in these realistic models may be more accurate. ITV values resulted to be much higher for cell suspensions in comparison to the values numerically determined for adhered cells. Analytical and numerical values are in good agreement only for cells showing a spheroidal shape when adhered. In general, the ITV numerically determined for all adhered cells, ranged between 200 mV and 800 mV. Specifically, the ITV resulted to be dependent on the cell type and cell morphology. In addition, the pathophysiological state of cells might influence the ITV values since they resulted to be much higher in non-transformed cells than in cancerous cells. It has been reported that ITV should not be unnecessarily high in order to avoid intracellular signaling to the cytoskeleton of non-target cells and influence a metastatic process. Hence, calculation of certain ITV values is encouraged based on prior determination of cell-specific reversible electroporation thresholds.

Lung cancer is one of the most common causes of cancer-related death worldwide. It is usually detected by CT or MRI and removed through thoracoscopic surgery. However, during the surgery, the lung collapses and a new determination of the position of the pulmonary nodule is necessary which is particularly challenging in the case of minimally invasive surgeries when palpation is not possible.In this contribution, ultra-wideband (UWB) radio technology is proposed for the localization of lung cancer. This was investigated through numerical simulations mimicking the frequencies range between 0.5 and 5 GHz and a nodule depth of 1 to 6 cm.A confocal map was reconstructed by positioning a monostatic antenna on a 5 $$\times $$ × 5 grid distribution on top of the lung tissue. The results show that the cancer localization was possible in the frequency between 0.5 and 1 GHz and nodules depth between 4 and 6 cm, while at lower depths artifacts appeared and at higher frequencies the electromagnetic attenuation given by the lung tissue was too high to detect the pulmonary nodule.

Spine disease is a growing problem in modern society and has been debilitating for every age-group. Researches have shown that more than 266 million people are facing degenerative spine disease and low back pain. CT scanning is a fast, painless, non-invasive diagnostic imaging modality that provides high spatial accuracy in obtaining the 3D structure of the vertebral. However, in real-life scenario, the clinic CT image might not cover the whole spine and the field of view might be hard to determine. Henceforth, this project aims to create and validate an automatic method that can detect, locate, and classify each vertebra from the partial field of view using deep learning. We used Mask R-CNN, a deep neural network aimed to solve the instance segmentation problem in machine learning or computer vision, and produce features such as bounding boxes, classes, and masks to identify each vertebra. This auto-detection method was validated on an open source dataset which has been used on Computational Spine Imaging (CSI 2014). The dataset was chosen by a radiologist with an eight years involvement with thoracic and lumbar spine column scans, and the data of twenty patients were collected using standard CT scanning protocol. The accuracy of the vertebra mask on 210 test images has been increased up to 99.9% DICE Coefficient in Mask R-CNN compare with 69.2% Dice Coefficient in another Deep-learning-based semantic segmentation framework U-Net.

In Chapter 76, one of the chapter author’s surname has been changed from “Federica Amintrano” to “Federica Amitrano”.