Histopathological image analysis is the gold standard for cancer diagnosis but is time-consuming, requires a high level of expertise, and is subject to inter-observer variability. The advancement of Digital Pathology enables the application of deep learning models for automating and enhancing diagnostic accuracy. In particular, Convolutional Neural Networks (CNNs) have emerged as a powerful tool for identifying morphological features in histopathological images, achieving accuracy comparable to that of medical experts in specific tasks such as tissue classification. Within the framework of this study, an interpretable CNN model is developed and evaluated for classifying gastric histopathological images as benign or malignant using the GasHisSDB dataset. The Monte Carlo Dropout method is applied to estimate the uncertainty of the model's predictions, while the Gradient-weighted Class Activation Mapping (Grad-CAM) method is combined with quantitative image feature analysis towards enabling the spatial localization of image regions that influence the model's predictions. The proposed approach achieved an AUC score of 98.6% while maintaining low computational cost and architectural simplicity. The generated interpretations yielded useful insights into spatially important image regions and their associated nuclear morphological characteristics.

The COVID-19 pandemic intensified the demand for rapid, accessible diagnostic methods. Machine learning models using cough audio recordings have shown potential for remote COVID-19 detection but often exhibit performance disparities across demographic and clinical subgroups. This study investigates fairness-aware machine learning models for COVID-19 diagnosis using crowdsourced data from the COVID-19 Sounds dataset. A Random Forest and a VGGish-based deep neural network classifier were developed. To mitigate bias, four different methods were applied and comparatively evaluated, namely correlation remover as a pre-processing approach, exponentiated gradient and adversarial debiasing as in-processing approaches, and threshold optimizer as a post-processing approach. Evaluation across sensitive attributes including gender, age, recording device's operating system, and the intersection of age and gender revealed that fairness could be substantially improved with minimal loss in predictive accuracy. The best equalized odds ratio values were$0.956,0.998$, and 0.88 with respect to gender, age, and recording device's operating system, respectively. Similarly, for the combination of gender and age the corresponding metric was 0.804. These findings support the feasibility of fair and reliable audio-based diagnostic systems, emphasizing the importance of integrating fairness into clinical machine learning pipelines.

In this study, we present preliminary results from the evaluation of an automated real-time stress detection approach during cognitively demanding human-computer interaction. The proposed method relies on heart rate variability (HRV) analysis derived from photoplethysmography (PPG) data collected via a wireless wearable sensor. A personalized HRV threshold is established during a Stroop task of escalating difficulty and subsequently used for binary classification between calm and stressed states. The stress-detection approach is evaluated through an experimental protocol employing a novel dynamic biofeedback serious game (SG) and continuous post-game self-annotations of perceived stress levels. Fourteen individuals participated in the study, with three excluded due to sensor-related issues. The collected data were analyzed in two directions-absolute values and ordinal-centric measuresexamining associations between the detected state and userreported annotations. Statistically significant differences between calm and stressed game segments were observed for the mean ($\mathbf{p}=0.002$), median ($\mathbf{p}=0.003$), and trapezoidal interval$(\mathbf{p}=\mathbf{0. 0 0 2})$of the annotation values. Ordinal analysis further confirmed this relationship, with positive Spearman rank correlations for the same features ($\mathbf{p}=\mathbf{0. 0 0 2}$) and consistent directionality in ten of eleven participants. These preliminary findings underscore the potential of the proposed thresholding approach and motivate further refinement of rule-based stress detection systems.

Carotid atherosclerosis represents a major risk factor for ischemic stroke, requiring accurate risk stratification for effective clinical intervention. While deep learning models demonstrate excellent performance in medical image analysis, their lack of uncertainty quantification limits deployment in clinical environments. This study investigates the use of two uncertainty estimation techniques, Monte Carlo Dropout (MCD) and Deep Ensembles (DE), in cardiovascular risk prediction from B-mode carotid ultrasound images. Using the CUBS dataset for training and two datasets (ATTIKON and BUSI) for external evaluation under distributional shift and out-of-distribution (OOD) conditions, the model's performance, calibration, and robustness are assessed. Results demonstrate that MCD provides superior generalization and OOD detection capabilities ($\text{AUC}=0.9589$), while DE excels in graceful degradation through rejecting high uncertainty samples, achieving 16.49 % accuracy improvement on low uncertainty samples. These findings highlight the complementary strengths of both methods and underscore the critical importance of uncertainty aware AI in clinical decision support systems.

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder characterized by significant phenotypic heterogeneity. Emerging evidence is associating ASD with disruptions in the gut microbiome pointing towards the gut-brain axis as a major contributor to ASD pathophysiology and offering microbial biomarkers with potential as early predictors. In this study supervised machine learning (ML) models trained and evaluated in a nested crossvalidation (NCV) framework are leveraged to classify ASD based on gut microbial profiles from nine different cohorts$(\mathrm{n}=929)$. To address the challenges posed by high dimensionality and biological variability, compositional data augmentation techniques - Aitchison Mixup, Compositional CutMix, and Feature Dropout - were integrated into the modeling pipeline. Among the evaluated ML models, XGBoost with Feature Dropout achieved the best performance (Accuracy: 73.3%, AUC:$\mathbf{8 2. 3 \%}$, F1-score:$\mathbf{7 3. 3 \%}$). To interpret the ML model's predictions, SHAP values and information gain were employed, highlighting key microbial species such as Enterobacter kobei, Dialister hominis, Leyella stercorea, and Comamonas kerstersii. These results reinforce the potential of the gut microbiome as a promising source of ASD screening biomarkers and highlight the utility of interpretable ML models to extract biologically meaningful patterns in complex microbiome datasets.

Reliable risk assessment for carotid atheromatous disease requires integrating diverse clinical and imaging information in a transparent and interpretable manner. This study investigates state-of-the-art large vision-language models (LVLMs) for multimodal carotid plaque assessment by integrating ultrasound imaging with structured clinical, demographic, laboratory, and biomarker data. A framework that simulates realistic diagnostic scenarios through interview-style question sequences is proposed, comparing open-source LVLMs including general-purpose and medically tuned models. Zero-shot experiments reveal that while most LVLMs accurately identify imaging modality and anatomy, all perform poorly in risk stratification. To address this, LLaVa-NeXT-Vicuna is adapted using low-rank adaptation (LoRA), achieving substantial improvements in stroke risk stratification. Integrating multimodal tabular data as text further enhances specificity and balanced accuracy, yielding competitive performance compared to prior CNN baselines. Our findings highlight both promise and limitations of LVLMs in ultrasound-based cardiovascular risk prediction, underscoring the importance of multimodal integration, domain adaptation, and model calibration for clinical translation.

Serious Games (SGs) are nowadays shifting focus to include procedural content generation (PCG) in the development process as a means of offering personalized and enhanced player experience. However, the development of a framework to assess the impact of PCG techniques when integrated into SGs remains particularly challenging. This study proposes a methodology for automated evaluation of PCG integration in SGs, incorporating deep reinforcement learning (DRL) game testing agents. To validate the proposed framework, a previously introduced SG featuring card game mechanics and incorporating three different versions of PCG for non-player character (NPC) creation, has been deployed. Version 1 features random NPC creation while versions 2 and 3 utilize a genetic algorithm approach. These versions are used to test the impact of different dynamic SG environments on the proposed framework's agents. The obtained results highlight the superiority of the DRL game testing agents trained on Versions 2 and 3 over those trained on Version 1 in terms of win rate (i.e. number of wins per played games) and training time. More specifically, within the execution of a test emulating regular gameplay, both Versions 2 and 3 peaked at a 97% win rate and achieved statistically significant higher (p = 0.009) win rates compared to those achieved in Version 1 that peaked at 94%. Overall results advocate towards the proposed framework's capability to produce meaningful data for the evaluation of procedurally generated content in SGs.

Abstract The authors propose a method for breast dynamic contrast enhanced‐magnetic resonance imaging classification by combining radiomic texture analysis with a hybrid adaptive neuro‐fuzzy inference system (ANFIS)‐particle swarm optimization (PSO) classifier. The fast discrete curvelet transform is utilized as a decomposition scheme in multiple scales. The mean and entropy features extracted from the produced scheme are used as texture descriptors. Principal component analysis (PCA) involves reduction of the dimensionality of the initial feature set. The transformed feature vector is subsequently introduced to a hybrid ANFIS‐PSO classifier. The average overall classification power of the proposed hybrid ANFIS‐PSO classifier is comparatively assessed to that obtained using several classifiers (ANFIS, linear discriminant analysis, Naïve Bayes, artificial neural networks, random forest and support vector machine) by using the 70 training‐30 testing data ratio. The comparison performed highlights the superiority of the proposed methodology, thus underlying the potential of ANFIS‐PSO for the breast cancer diagnosis with a classification accuracy of 94%.

Technological advancements in glucose sensing, insulin pumps, and closed-loop glucose control algorithms open new opportunities towards the realization of Artificial Pancreas (AP). However, the effective management of meal disturbances in these systems still remains a challenge. Meal detection algorithms eliminate the need for meal announcements and enable the shift to more automated and reliable AP systems. The aim of the present study is to develop and evaluate a personalized approach for the detection of meal disturbances in patients with Type 1 Diabetes Mellitus (T1DM). Long Short Term Memory Neural Networks (LSTM)'s inherent ability to efficiently handle sequential data is leveraged within an ensemble learning strategy towards the development of different versions of ensemble models. The models receive as input sequences of Continuous Glucose Monitoring (CGM) measurements (glucose profiles) of a 120-min duration and classify them as positive or negative for the onset of an ingested meal. In silico evaluation is performed using the UVA-PADOVA T1DM Simulator. All ensembles achieve acceptable discriminative performance (mean c-statistic: 75.12%-79.52%) and are able to detect meals in a timely manner (mean detection time: 7.08-12.84 min). Statistical analysis demonstrates the superiority of the simple averaging combination scheme over the other schemes in terms of the c-statistic.

Cardiovascular Disease (CVD) is an important cause of disability and death among individuals with Diabetes Mellitus (DM). International clinical guidelines for the management of Type 2 DM (T2DM) are founded on primary and secondary prevention and favor the evaluation of CVD-related risk factors towards appropriate treatment initiation. CVD risk prediction models can provide valuable tools for optimizing the frequency of medical visits and performing timely preventive and therapeutic interventions against CVD events. The integration of explainability modalities in these models can enhance human understanding on the reasoning process, maximize transparency and embellish trust towards the models' adoption in clinical practice. The aim of the present study is to develop and evaluate an explainable personalized risk prediction model for the fatal or non-fatal CVD incidence in T2DM individuals. An explainable approach based on the eXtreme Gradient Boosting (XGBoost) and the Tree SHAP (SHapley Additive exPlanations) method is deployed for the calculation of the 5-year CVD risk and the generation of individual explanations on the model's decisions. Data from the 5-year follow up of 560 patients with T2DM are used for development and evaluation purposes. The obtained results (AUC=71.13%) indicate the potential of the proposed approach to handle the unbalanced nature of the used dataset, while providing clinically meaningful insights about the model's decision process.

Implantable medical devices are attracting high scientific interest, as they can significantly improve the quality of life of people with chronic diseases. Retinal implants can partially restore vision to the blind with a functional optic nerve. This paper presents the design of biocompatible patch antennas for retinal implants operating at the MedRadio band (401-406MHz). Biocompatibility is ensured by encapsulating the antenna within a biocompatible material. First, a parametric microstrip patch antenna model is considered and the initial antenna design is modified for encapsulation in three different biocompatible materials (Silastic MDX4-4210 medical-grade elastomer, Zirconia and PEEK). Antenna design takes place inside a homogeneous spherical eyeball model in order to take into consideration mismatching caused by surrounding tissues. An automated Quasi-Newton method is used to refine the antenna design in order to achieve the desired tuning properties. The resulting three antenna designs are comparatively assessed in terms of size, tuning properties, safety and radiation performance for telemetry in the MedRadio band (401-406 MHz). Further validation and safety simulations are carried out for antennas implanted in an anatomical eyeball model consisting of four different tissues (sclera, cornea, vitreous humor and lens) in order to explore potential detuning of the antenna caused by its actual operation environment. Simulations for design optimization are performed using the Ansoft HFSS suite, based on the finite element (FE) method, while validation and safety performance studies are carried out using the SEMCAD X suite, based on the finite difference time domain (FDTD) method.

In this paper, a procedural content generation (PCG) technique that was incorporated in a novel serious game for obstructive sleep apnea (OSA) is presented. The technique is based on a genetic algorithm and aims to enhance user engagement and deliver educational material tailored to user needs. The genetic algorithm monitors user choices and game progress by means of two fitness functions that dictate suitable candidates to produce offspring in each generation. An initial validation in terms of user experience was conducted, by deploying three different versions of the serious game. Version A and B incorporated the genetic algorithm, along with mechanisms for automated game difficulty adjustment. Version A was designed to be difficult and frustrating while version B to adjust difficulty smoothly. Version C did not display any adaptive properties. 42 participants were recruited and split in two groups to play two versions of the game, A and C or B and C, without any prior knowledge of differences between them. After each session, two modules of the Game Experience Questionnaire (GEQ) were applied. The obtained results reveal statistically significant differences regarding user perception in terms of competence, challenge and negative experience for versions A and C respectively, and competence and negative experience for versions B and C respectively. Version B achieved better GEQ scores than version C while version A resulted in worse GEQ scores than version C.

This paper reviews state-of-the-art research solutions across the spectrum of medical imaging informatics, discusses clinical translation, and provides future directions for advancing clinical practice. More specifically, it summarizes advances in medical imaging acquisition technologies for different modalities, highlighting the necessity for efficient medical data management strategies in the context of AI in big healthcare data analytics. It then provides a synopsis of contemporary and emerging algorithmic methods for disease classification and organ/ tissue segmentation, focusing on AI and deep learning architectures that have already become the de facto approach. The clinical benefits of in-silico modelling advances linked with evolving 3D reconstruction and visualization applications are further documented. Concluding, integrative analytics approaches driven by associate research branches highlighted in this study promise to revolutionize imaging informatics as known today across the healthcare continuum for both radiology and digital pathology applications. The latter, is projected to enable informed, more accurate diagnosis, timely prognosis, and effective treatment planning, underpinning precision medicine.

The goal of this work is to study mechanisms of visual attention to assist visual perception for patients suffering from age-related macular degeneration (AMD) or retinitis pigmentosa (RP) through artificial retina devices. We present a method to predict where humans look; we extend a visual saliency model by incorporating additional features and use this model to obtain saliency maps. These are thresholded at different scales to estimate the points of an image upon which the human eye fixates as well as the exact sequence of these fixations. The sequence of fixations extracted is further used to identify the part of the image that will mostly attract visual attention. Contrary to most existing approaches our method can indicate specific coordinates for the fixation points rather than generic areas that may attract visual attention and is thus more appropriate to imitate human fixations. Our method performs marginally better than the well-known method for saliency prediction we compare against (≈76% accuracy) and very satisfactorily in terms of estimating the sequence of fixations upon any given image (up to 98% accuracy).

In this paper, preliminary results of the evaluation of a serious game promoting nutrition literacy (NL) and food literacy (FL) are presented. The serious game's effectiveness was evaluated in terms of educational value and user experience through a two-part evaluation strategy. In the first part, a quasi-experimental study was designed to assess the serious game's educational value compared to an alternative intervention based on the study of text-based material. Appropriate questionnaires were delivered prior to, immediately after, and one week after the intervention. In the second part of the evaluation strategy, the user experience was measured by means of the Game Experience Questionnaire (GEQ). Nineteen and 29 participants enrolled in the first and second part of the evaluation, respectively. The results of the study showed that both serious game and control intervention enhance user's NL and FL skills (p-value = 0.002, 0.025 respectively). Comparison between the two groups did not yield significant results (p-value = 0.25). Increased levels of competence, immersion, flow and positive affect were declared in the GEQ demonstrating the attractiveness of the serious game. Moreover, the study revealed an important association between the level of game interaction, as measured by the number of mouse clicks per second, and the user experience. Intermediate levels of mouse interaction indicate lower user engagement.

The papers in this special section focus on neuro-informatics which is considered one of the most attractive research fields for scientists, engineers, practitioners and physicians due to its profound importance in healthcare and in our lives. Human curiosity, the BRAIN project in USA with a very large funding budget, and the exponential evolution of computational informatics and nanotech during the last two decades have inspired and motivated many researchers around the globe to contribute with their research to the brain. The papers are associated with the IEEE BIBE-2017 Conference.

Abstract A method based on the combination of Local Binary Pattern operator and radial lengths is presented aiming at the identification of Architectural Distortions (ADs) in mammograms. Local Binary Pattern operator, a number of its variants, and radial lengths are combined together producing a high‐dimensional feature space. A process, based on the combination of Principal Component Analysis and ttest, is used to effectively transform feature space and reveal the most descriptive features. The classification step is performed using a Support Vector Machine classifier. Open access databases (Mammographic Image Analysis Society and Digital Database for Screening Mammography) are used through an exhaustive evaluation framework that aims at eliminating both mammogram selection bias and limited subtlety variation, thus enabling a fair and complete comparison procedure. Furthermore, in order to provide a test bed for future comparisons, a dataset is constructed from all the available AD Regions Of Interest in Digital Database for Screening Mammography (163 AD vs 375 Regions Of Interest from specific normal cases) and is used to further evaluate the performance of the proposed method. The method performed flawlessly and classified correctly all cases.

Abstract The aim of the present study is to comparatively assess the performance of different machine learning and statistical techniques with regard to their ability to estimate the risk of developing type 2 diabetes mellitus (Case 1) and cardiovascular disease complications (Case 2). This is the first work investigating the application of ensembles of artificial neural networks (EANN) towards producing the 5‐year risk of developing type 2 diabetes mellitus and cardiovascular disease as a long‐term diabetes complication. The performance of the proposed models has been comparatively assessed with the performance obtained by applying logistic regression, Bayesian‐based approaches, and decision trees. The models' discrimination and calibration have been evaluated using the classification accuracy (ACC), the area under the curve (AUC) criterion, and the Hosmer–Lemeshow goodness of fit test. The obtained results demonstrate the superiority of the proposed models (EANN) over the other models. In Case 1, EANN with different topologies has achieved high discrimination and good calibration performance (ACC = 80.20%, AUC = 0.849, p value = .886). In Case 2, EANN based on bagging has resulted in good discrimination and calibration performance (ACC = 92.86%, AUC = 0.739, p value = .755).

Alarming epidemiological features of Alzheimer's disease impose curative treatment rather than symptomatic relief. Drug repurposing, that is reappraisal of a substance's indications against other diseases, offers time, cost and efficiency benefits in drug development, especially when in silico techniques are used. In this study, we have used gene signatures, where up- and down-regulated gene lists summarize a cell's gene expression perturbation from a drug or disease. To cope with the inherent biological and computational noise, we used an integrative approach on five disease-related microarray data sets of hippocampal origin with three different methods of evaluating differential gene expression and four drug repurposing tools. We found a list of 27 potential anti-Alzheimer agents that were additionally processed with regard to molecular similarity, pathway/ontology enrichment and network analysis. Protein kinase C, histone deacetylase, glycogen synthase kinase 3 and arginase inhibitors appear consistently in the resultant drug list and may exert their pharmacologic action in an epidermal growth factor receptor-mediated subpathway of Alzheimer's disease.

The papers in this special section were presented at three well-known conferences organized in 2014: EAI Mobihealth, IEEE HealthCom, and IEEE Biomedical and Health Informatics. EAI Mobihealth is an annually organized conference, which started in 2010, to address the demands of the rapidly evolving disciplines of wireless communications, mobile computing, and sensing technologies in healthcare. The IEEE-Healthcom is held every year since 1999 in different countries in Asia, Europe, and in America. It aims at bringing together interested parties working in the field of healthcare to exchange ideas, discuss innovative and emerging solutions, and develop collaborations. The IEEE Biomedical Health Informatics Conference started in 2013 and is organized every year providing the forum to showcase enabling technologies of computing, devices, imaging, sensors, and systems that optimize the acquisition, transmission, processing, storage, retrieval, visualization, and analysis of medical data. The aim of this special section is to present an overview of recent advances in sensing technologies, monitoring of patients, security and privacy of data transfer, provision of collaborative environments, data gathering and analysis from various sources, and predictive models, which all finally target the best strategy for patient monitoring and treatment.

The aim of this special section is to provide an overview of the emerging biomedical informatics technologies and their application in research and clinical environments. Recent developments in biomedical informatics have created methods, techniques and tools, which are based on the analysis of heterogeneous data, data mining, decision support systems, multiscale modeling, etc. The distance from the development of such systems and the real clinical environments is still long enough, and only some of them have been used in a clinical scale.

A computer-aided diagnosis (CADx ) framework for the diagnosis of clustered microcalcifications (MCs) has already been developed, which is based on the analysis of MCs' morphologies,the shape of the cluster they form and the texture of the surrounding tissue. In this study, we investigate the diagnostic information that the relative location of the cluster inside the breast may provide. Breast probabilistic maps are generated and adopted in the CADx pipeline, expecting to empower its diagnostic procedure. We propose a flowchart combining alternative classification algorithms and the aforementioned probabilistic maps in order to provide a final risk for malignancy for new considered mammograms. For the evaluation performance, a large dataset of mammograms provided from the Digital Database of Screening Mammography (DDSM) has been used. The obtained results indicate that the proposed modifications lead to the enhancement of the diagnostic process, as the classification results are further improved. Additionally, a straightforward comparison between the CADx pipeline and the radiologists who assessed the same mammograms, reveal that the CADx pipeline performs toward the right direction, as the sensitivity remains at high levels, while improving both the accuracy, from 51.4% to 69%, and the specificity, from 16.6% to 54.7%.

Valid characterization of carotid atherosclerosis (CA) is a crucial public health issue, which would limit the major risks held by CA for both patient safety and state economies. This paper investigated the unexplored potential of kinematic features in assisting the diagnostic decision for CA in the framework of a computer-aided diagnosis (CAD) tool. To this end, 15 CAD schemes were designed and were fed with a wide variety of kinematic features of the atherosclerotic plaque and the arterial wall adjacent to the plaque for 56 patients from two different hospitals. The CAD schemes were benchmarked in terms of their ability to discriminate between symptomatic and asymptomatic patients and the combination of the Fisher discriminant ratio, as a feature-selection strategy, and support vector machines, in the classification module, was revealed as the optimal motion-based CAD tool. The particular CAD tool was evaluated with several cross-validation strategies and yielded higher than 88% classification accuracy; the texture-based CAD performance in the same dataset was 80%. The incorporation of kinematic features of the arterial wall in CAD seems to have a particularly favorable impact on the performance of image-data-driven diagnosis for CA, which remains to be further elucidated in future prospective studies on large datasets.

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) remains an empirical, yet highly effective, surgical treatment for advanced Parkinson's disease (PD). DBS outcome depends on accurate stimulation of the STN sensorimotor area which is a trial-and-error procedure taking place during and after surgery. Pathologically enhanced beta-band (13-35 Hz) oscillatory activity across the cortico-basal ganglia pathways is a prominent neurophysiological phenomenon associated with PD. We hypothesized that weighing together beta-band frequency peaks from simultaneous microelectrode recordings in "off-state" PD patients could map the individual neuroanatomical variability and serve as a biomarker for the location of the STN sensorimotor neurons. We validated our hypothesis with 9 and 11 patients that, respectively, responded well and poorly to bilateral DBS, after at least two years of follow up. We categorized "good" and "poor" DBS responders based on their clinical assessment alongside a > 40% and <30% change, respectively, in "off" unified PD rating scale motor scores. Good (poor) DBS responders had, in average, 1 mm (3.5 mm) vertical distance between the maximum beta-peak weighted across the parallel microelectrodes and the center of the stimulation area. The distances were statistically different in the two groups ( p = 0.0025 ). Our biomarker could provide personalized intra- and postoperative support in stimulating the STN sensorimotor area associated with optimal long-term clinical benefits.

Implant devices are used to measure biological parameters and transmit their results to remote off-body devices. As implants are characterized by strict requirements on size, reliability, and power consumption, applying the concept of cooperative communications to wireless body area networks offers several benefits. In this paper, we aim to minimize the power consumption of the implant device by utilizing on-body wearable devices, while providing the necessary reliability in terms of outage probability and bit error rate. Taking into account realistic power considerations and wireless propagation environments based on the IEEE P802.l5 channel model, an exact theoretical analysis is conducted for evaluating several communication scenarios with respect to the position of the wearable device and the motion of the human body. The derived closed-form expressions are employed toward minimizing the required transmission power, subject to a minimum quality-of-service requirement. In this way, the complexity and power consumption are transferred from the implant device to the on-body relay, which is an efficient approach since they can be easily replaced, in contrast to the in-body implants.

Breast microcalcifications are one of the most important mammographic findings related to the existence of the breast cancer. Radiologists usually characterize microcalcifications based on their morphologies, the distribution within the cluster they form, the shape of the cluster and its relative location inside the breast. In this study, we focus on the latter factor and we study its effect on the probability of malignancy. The main purpose of our study is to generate probabilistic breast cancer atlases for clusters of microcalcifications in order to visualize the influence of cluster location on cancer probability. We propose a framework for the generation of such atlases, including segmentation of important breast landmarks and projection of different clusters of microcalcifications on a reference breast shape. The generation of the atlases is implemented using mammograms from the Digital Database of Screening Mammography. The obtained probabilistic atlases reveal specific areas in the breast of higher occurrence of clusters and higher risk of malignancy.

In this work, we study the factors that influence the efficacy of a proposed Computer Aided Diagnosis (CADx) framework for the diagnosis of clustered microcalcifications (MCs) using a large dataset of mammograms containing cases of varying breast density and findings' subtlety. The reported results indicate that the proposed framework performs towards the right direction, as it appears high classification performance (Az=0.909) for specific subsets of cases, while outperforming at the same time the performance of the radiologists who evaluated the same cases. The effect of the initial enhancement of mammograms in the CADxpipeline is then investigated, by applying three different image enhancement techniques on several subsets of mammograms. We observed that for the considered subsets of dense mammograms, a wavelet-based enhancement algorithm outperformed the rest and provided superior classification performance (Az=0.849). We indicate therefore that the density of the breast determines the need of different computational algorithms for the analysis of a mammogram and as a result the a priori knowledge of this factor may be exploited for the optimization of the diagnostic process.

The SIFEM project targets the development of an infrastructure in order to semantically link open source tools and libraries with existing data as well as new knowledge towards the multi-scale finite element modelling of the inner-ear. The SIFEM system is designed based on an open architecture schema that consists of a set of tools and subsystems in order to develop robust multi-scale models. The project mainly delivers: (i) tools for finite elements modelling, (ii) cochlea reconstruction tool and (iii) 3D inner ear models visualization tool. The main scientific results contribute to the knowledge of alterations associated to diverse cochlear disorders and could lead, in long-term, to personalized healthcare. The overview of the SIFEM platform and its architecture is presented in this paper.

The purpose of this study is to present a hybrid approach based on the combined use of a genetic algorithm (GA) and a nearest neighbours classifier for the selection of the critical clinical features which are strongly related with the incidence of fatal and non fatal Cardiovascular Disease (CVD) in patients with Type 2 Diabetes Mellitus (T2DM). For the development and the evaluation of the proposed algorithm, data from the medical records of 560 patients with T2DM are used. The best subsets of features proposed by the implemented algorithm include the most common risk factors, such as age at diagnosis, duration of diagnosed diabetes, glycosylated haemoglobin (HbA1c), cholesterol concentration, and smoking habit, but also factors related to the presence of other diabetes complications and the use of antihypertensive and diabetes treatment drugs (i.e. proteinuria, calcium antagonists, b-blockers, diguanides and insulin). The obtained results demonstrate that the best performance was achieved when the weighted k-nearest neighbours classifier was applied to the CVD dataset with the best subset of features selected by the GA, which resulted in high levels of accuracy (0.96), sensitivity (0.80) and specificity (0.98).

Smart cards in Health Information Services (HIS) are considered to have great potential to improve the delivery of healthcare services and reduce healthcare costs. On the other hand, HIS smart cards also introduce new challenges and limitations that require further analysis before a full-scale implementation. In this paper, smart cards in HIS are presented, along with current implementations and the benefits and the limitations that arise from their use.

In this study, we numerically assess detuning issues for an ingestible antenna which is designed to operate in the Medical Device Radiocommunications Service (MedRadio, 401-406 MHz), as it travels along the gastrointestinal (GI) tract. For this purpose, we evaluate the antenna resonance performance within four canonical single-tissue models of the human esophagus, stomach, small and large intestine. The antenna is further placed at different locations within the aforementioned tissue models in order to assess detuning issues related to its relative positioning within each of them. Inherent detuning issues are observed and discussed in the four different simplified tissue models considering three specific locations of the antenna in each model, resulting in twelve different scenarios. The resonance, radiation and safety performance of the ingestible antenna is, finally, evaluated.

The incorporation of wavelet-based multiscale image decomposition in motion-estimation schemes has been shown to have a favourable impact on accuracy in tracking motion of the carotid artery wall from B-mode ultrasound image sequences. In this work, in an attempt to further enhance accuracy, we investigate the effects of different parameters of multiscale image decomposition. To this end, we optimize multiscale weighted least-squares optical flow (MWLSOF), a previously presented multiscale motion estimator, in terms of (a) the type of wavelet transform (WT) (discrete (DWT) and stationary (SWT) WTs), (b) the WT function and (c) the total number of levels of image decomposition. The optimization is performed in the context of an in silico data framework, consisting of simulated ultrasound image sequences of the carotid artery. We propose SWT, a high-order coiflet function (ex. coif5) and one level of multiscale image decomposition as the optimal parameterization for MWLSOF to achieve maximum accuracy in the particular application. Finally, we demonstrate the usefulness of an accurate motion estimator in real data experiments, by applying the optimized MWLSOF to real image data of patients with carotid atherosclerosis.

The present paper aims at the design, the development and the evaluation of a personalized glucose-insulin metabolism model for patients with Type 1 Diabetes Mellitus (T1DM). The personalized model is based on the combined use of Compartmental Models (CMs) and a Self Organizing Map (SOM). The model receives information related to previous glucose levels, subcutaneous insulin infusion rates and the time and amount of carbohydrates ingested. Previous glucose measurements along with the outputs of the CMs which simulate the sc insulin kinetics and the glucose absorption from the gut into the blood, respectively, are fed into the SOM which simulates glucose kinetics in order for the latter to provide with future glucose profile. The personalized model is evaluated using data from the medical records of 12 patients with T1DM for the time being on insulin pumps and CGMS. The obtained results demonstrate the ability of the proposed model to capture the metabolic behavior of a patient with T1DM and to handle intra- and inter-patient variability.

These tutorials/keynote speeches discuss the following: the effects of nicotine exposure on the complexity and the genetic patterns of dopamine neurons in VTA; from 6-Ps medicine to cardiovascular health informatics; computer-aided interpretation of vascular images towards valid diagnosis and risk stratification of atherosclerosis; turning data into predictions of gene and protein function; from reading to writing (and rewriting) the code of life: the future of biology - scientific, ethical, legal, civil and social issues.

This paper studies the application of fuzzy clustering with viewpoints in order to cluster cell samples according to their gene expression profile. This method combines fuzzy clustering with external domain knowledge represented by the so-called viewpoints. The viewpoints that we employ are obtained from previously available expression data. The method was compared to the clustering algorithms of k-means, fuzzy c-means, affinity propagation, as well as a method of clustering microarray data that is based on prior biological knowledge, and has shown comparable/improved results over them.

Extracellular recordings in the area of the subthalamic nucleus (STN) of Parkinson's disease patients undergoing deep brain stimulation comprise fast events, Action Potentials and slower events, known as Local Field Potentials (LFP). The LFP is believed to represent the synchronized input into the observed area, as opposed to the spike data, which represents the output. We have shown before that there is an input-output relationship between these two components in the STN. In the present paper, we extend these observations by using LFP-driven Volterra models and the Laguerre expansion technique to estimate nonlinear dynamic models which are able to predict the recorded spiking activity. To this end, we rigorously examine the optimal model order. The improved performance of the second-order Volterra models indicates that there is a nonlinear relationship between the LFP and the spiking activity. To obtain a more compact and readily interpretable model, the most significant dynamic components of the identified Volterra models are extracted using principal dynamic mode analysis.

Motion of the carotid artery wall is important for the quantification of arterial elasticity and contractility and can be estimated with a number of techniques. In this paper, a framework for quantitative evaluation of motion analysis techniques from B-mode ultrasound images is introduced. Six synthetic sequences were produced using 1) a real image corrupted by Gaussian and speckle noise of 25 and 15 dB, and 2) the ultrasound simulation package Field II. In both cases, a mathematical model was used, which simulated the motion of the arterial wall layers and the surrounding tissue, in the radial and longitudinal directions. The performance of four techniques, namely optical flow (OF (HS)), weighted least-squares optical flow (OF (LK(WLS))), block matching (BM), and affine block motion model (ABMM), was investigated in the context of this framework. The average warping indices were lowest for OF (LK(WLS)) (1.75 pixels), slightly higher for ABMM (2.01 pixels), and highest for BM (6.57 pixels) and OF (HS) (11.57 pixels). Due to its superior performance, OF (LK(WLS)) was used to quantify motion of selected regions of the arterial wall in real ultrasound image sequences of the carotid artery. Preliminary results indicate that OF (LK(WLS)) is promising, because it efficiently quantified radial, longitudinal, and shear strains in healthy adults and diseased subjects.

In this paper, we discuss the use of a nonlinear cascade model to predict the subthalamic nucleus spike activity from the local field potentials recorded in the motor area of the nucleus of Parkinson's disease patients undergoing deep brain stimulation. We use a segment of appropriately selected and processed data recorded from five nuclei to acquire the information of the spike timing and rhythm of a single neuron and estimate the model parameters. We then use the rest of each recording to assess the model's accuracy in predicting spike timing, rhythm, and interspike intervals. We show that the cumulative distribution function (CDF) of the predicted spikes remains inside the 95% confidence interval of the CDF of the recorded spikes. By training the model appropriately, we prove its ability to provide quite accurate predictions for multiple-neuron recordings as well, and we establish its validity as a simple yet biologically plausible model of the intranuclear spike activity recorded from Parkinson's disease patients.

In this paper, a multiresolution approach is suggested for texture classification of atherosclerotic tissue from B-mode ultrasound. Four decomposition schemes, namely, the discrete wavelet transform, the stationary wavelet transform, wavelet packets (WP), and Gabor transform (GT), as well as several basis functions, were investigated in terms of their ability to discriminate between symptomatic and asymptomatic cases. The mean and standard deviation of the detail subimages produced for each decomposition scheme were used as texture features. Feature selection included 1) ranking the features in terms of their divergence values and 2) appropriately thresholding by a nonlinear correlation coefficient. The selected features were subsequently input into two classifiers using support vector machines (SVM) and probabilistic neural networks. WP analysis and the coiflet 1 produced the highest overall classification performance (90% for diastole and 75% for systole) using SVM. This might reflect WP's ability to reveal differences in different frequency bands, and therefore, characterize efficiently the atheromatous tissue. An interesting finding was that the dominant texture features exhibited horizontal directionality, suggesting that texture analysis may be affected by biomechanical factors (plaque strains).

BACKGROUND: Obesity is a multifactorial trait, which comprises an independent risk factor for cardiovascular disease (CVD). The aim of the current work is to study the complex etiology beneath obesity and identify genetic variations and/or factors related to nutrition that contribute to its variability. To this end, a set of more than 2300 white subjects who participated in a nutrigenetics study was used. For each subject a total of 63 factors describing genetic variants related to CVD (24 in total), gender, and nutrition (38 in total), e.g. average daily intake in calories and cholesterol, were measured. Each subject was categorized according to body mass index (BMI) as normal (BMI ≤ 25) or overweight (BMI > 25). Two artificial neural network (ANN) based methods were designed and used towards the analysis of the available data. These corresponded to i) a multi-layer feed-forward ANN combined with a parameter decreasing method (PDM-ANN), and ii) a multi-layer feed-forward ANN trained by a hybrid method (GA-ANN) which combines genetic algorithms and the popular back-propagation training algorithm. RESULTS: PDM-ANN and GA-ANN were comparatively assessed in terms of their ability to identify the most important factors among the initial 63 variables describing genetic variations, nutrition and gender, able to classify a subject into one of the BMI related classes: normal and overweight. The methods were designed and evaluated using appropriate training and testing sets provided by 3-fold Cross Validation (3-CV) resampling. Classification accuracy, sensitivity, specificity and area under receiver operating characteristics curve were utilized to evaluate the resulted predictive ANN models. The most parsimonious set of factors was obtained by the GA-ANN method and included gender, six genetic variations and 18 nutrition-related variables. The corresponding predictive model was characterized by a mean accuracy equal of 61.46% in the 3-CV testing sets. CONCLUSIONS: The ANN based methods revealed factors that interactively contribute to obesity trait and provided predictive models with a promising generalization ability. In general, results showed that ANNs and their hybrids can provide useful tools for the study of complex traits in the context of nutrigenetics.

SMARTDIAB is a platform designed to support the monitoring, management, and treatment of patients with type 1 diabetes mellitus (T1DM), by combining state-of-the-art approaches in the fields of database (DB) technologies, communications, simulation algorithms, and data mining. SMARTDIAB consists mainly of two units: 1) the patient unit (PU); and 2) the patient management unit (PMU), which communicate with each other for data exchange. The PMU can be accessed by the PU through the internet using devices, such as PCs/laptops with direct internet access or mobile phones via a Wi-Fi/General Packet Radio Service access network. The PU consists of an insulin pump for subcutaneous insulin infusion to the patient and a continuous glucose measurement system. The aforementioned devices running a user-friendly application gather patient's related information and transmit it to the PMU. The PMU consists of a diabetes data management system (DDMS), a decision support system (DSS) that provides risk assessment for long-term diabetes complications, and an insulin infusion advisory system (IIAS), which reside on a Web server. The DDMS can be accessed from both medical personnel and patients, with appropriate security access rights and front-end interfaces. The DDMS, apart from being used for data storage/retrieval, provides also advanced tools for the intelligent processing of the patient's data, supporting the physician in decision making, regarding the patient's treatment. The IIAS is used to close the loop between the insulin pump and the continuous glucose monitoring system, by providing the pump with the appropriate insulin infusion rate in order to keep the patient's glucose levels within predefined limits. The pilot version of the SMARTDIAB has already been implemented, while the platform's evaluation in clinical environment is being in progress.

The 15 papers in this special section on new and emerging technologies in bioinformatics and bioengineering have been grouped into the following categories: 1) computational intelligence and data mining in support of decision making in biomedicine; 2) new applications and wireless communication issues in body sensor networks; 3) human-computer interaction; and 4) medical imaging and physiological systems modeling.

UNLABELLED: The uniqueness of a measured molecular mass or peptide sequence plays a very important role in the fields of protein identification and peptide/protein-biomarker investigation. We present a publicly available web application that offers information concerning the uniqueness of one or more molecular masses and one or more peptide sequences in the human proteome. When a sequence is found to be unique in humans, the application is able to search across all species querying whether this sequence is unique, not only in humans but also in other species found in the Swiss-Prot Database. The application is also able to search for unique protein fragments derived computationally from enzymatic digestion driven by certain enzymes. Furthermore, the application can list all the unique masses and peptides of a given protein. Through this application, researchers are able to find unique tags, either on a molecular mass level or on a sequence level. These unique tags are remarkably important in research related to protein identification or biomarker discovery and measurements. AVAILABILITY: UniMaP web-application is available at http://bioserver-1.bioacademy.gr/Bioserver/UniMaP/

The wavelet entropy (WE) of rest electroencephalogram (EEG) and of event-related potentials (ERP) carries information about the degree of order or disorder associated with a multi-frequency brain electrophysiological activity. In the present study, WE, relative WE and WE change were estimated for the EEG and ERP signals recorded during a working memory task, from dyslectic children and healthy subjects. The analysis of the two groups (controls vs dyslectics) revealed differentiations mainly in relative WE and WE change that takes into account the variability of rest EEG. These findings indicate that the WE can be employed as a quantitative measure for monitoring EEG and ERP activities and may provide a useful tool in analyzing electrophysiological signals associated with dyslexia.

Recent studies in the field of molecular imaging have demonstrated the need for PET probes capable of imaging very weak activity distributions. Over this range of applications the sensitivity and the energy resolution of a PET system can be critical, as it can possibly affect the minimum amount of activity which can be reliably detected. Clinical PET systems, as opposed to small animal systems, are less sensitive and, therefore, imaging of very low activity sources could be challenging. Moreover the presence of LSO crystal detectors can further raise the minimum detection threshold due to the intrinsic radioactivity of the176Lu contained in the LSO compound. Our aim is to examine the feasibility of using an LSO-based clinical PET scanner for imaging activity distributions of 4nCi/mm3or less. In this study the parameter of minimum detectable activity (MDA) has been used for the quantification of the detection threshold of a system. A series of acquisitions was simulated using the Monte Carlo simulation software package of GATE. Existing validated GATE models of the clinical Siemens Biograph 6 PET/CT and pre-clinical microPET Focus 220 scanners have been applied to quantify and compare the effect of the LSO background and the energy window on the MDA performance of the systems. Four square regions, each with a unique signal-to-background activity concentration ratio (SBR), were scanned simultaneously. The intrinsic LSO background spectrum and the total energy spectrum, as well as their relative positions and intensities were estimated. The simulated data were histogrammed on various time frames, which were later reconstructed using a filtered back-projection algorithm. Detectability in every image was quantified using a modified Currie equation to associate an MDA value with a specific region and frame length. In the case of Biograph an MDA of 4nCi/mm3can be reliably detected for frame lengths longer than 5 min and in regions where the SBR was higher than 4. When higher contrast regions are imaged, detection can be achieved even for frame lengths down to 1 min. The previous analysis was repeated by using a GATE model of a hypothetical BGO-based clinical PET scanner. The results between the two scanner models were compared with each other as well as with those of a previous MDA study on a pre-clinical microPET Focus 220 scanner.

Deep brain stimulation (DBS) is considered a surgical treatment alternative for Parkinson disease (PD) patients with intractable tremor or for those patients who are affected by long-term complications of levodopa therapy such as motor fluctuations and severe dyskinesias. However, the susceptible accuracy of placement of the DBS electrode inside the brain nucleus determines the therapeutic efficacy of the method. Unlike normal cases, untreated Parkinsonian states in basal ganglia structures produce oscillations at various frequencies, the most prominent of which is a synchronization in the beta frequency band recorded in the subthalamic nucleus (STN). The actual frequency range and the strength of the beta peak vary among patients. We propose that an on-line spectral analysis of the population activity, as evidenced by microelectrode recordings (MERs) could form a neurophysiological biomarker for confirmation of the on-target placement of the electrode within the STN.

The fractal dimension (FD) is a natural measure of the irregularity of a curve. In this study the performances of two FD-based methodologies are compared in terms of their ability to detect the onset of epileptic seizures in scalp EEG. The FD algorithms used is Katzpsilas, which has been broadly utilized in the EEG analysis literature, and the k-nearest neighbor (k-NN), which is applied in this study in a time series sense for the first time. 244.9 hours of EEG recordings, including 16 seizures in 3 patients, were analyzed. Both approaches achieved 100% sensitivity with a false positive rate of 0.85 FP/h for the k-NN algorithm and 1 FP/h for Katzpsilas algorithm. The corresponding detection delays were 6.5 s and 10.5 s on the average, respectively. The k-NN algorithm seems to outperform Katzpsilas algorithm. Results are satisfactory in comparison to other methodologies applied on scalp EEG and proposed in the literature.

A population level model of the basal ganglia has been shown to reliably reproduce the local field potential (LFP) activity recorded from subthalamic nucleus (STN) during typical microelectrode recording sessions. The purpose of the present work is to investigate optimization methods that can be used to fit that model to actual recorded LFPs. For that, we utilize data derived from seven parkinsonian subjects prior to the permanent implantation of the deep brain stimulation (DBS) electrode. For the fitting, five optimization methods are used, combined with two methods for estimating the error between the actual recorded and the model predicted LFP signals in the frequency domain. The procedures are focused on re-generating the characteristic beta peak of the ST LFP. The results indicate that the model is able to reproduce the beta peak in various frequencies in the range of both low and high beta, while at the same time, the values of the critical parameters bringing the model in that area of behavior reveal the crucial role of the synaptic strengths in Parkinson’s disease pathophysiology.

Identification of gene-gene and gene-environment interactions that contribute in the onset of a multi-factorial disease supports the prevention of diseases like the Cardiovascular Disease (CVD). Body Mass Index (BMI), a measure of human obesity, is an independent risk factor of CVD. Furthermore, it is known that a subjectpsilas BMI is affected both by his/her lifestyle, e.g. nutrition, and genetic profile. Aim of the paper is to predict a subjectpsilas onset of obesity using lifestyle and genetic information. The prediction is performed by a computational intelligence based system using a Parameter Decreasing Method (PDM) combined with an Artificial Neural Network (ANN). The system uses an initial set of 63 input variables corresponding to sex, average nutrition intake measurements, and genetic variations to identify the 32 most important ones that affect BMI. The selected variables are the ones to interact with each other towards the complex trait of BMI, which is used as a 2-class output variable (BMI les 25 vs. BMIges25) in the ANN. The system achieved a mean accuracy of the system evaluated by a 3-cross validation resampling technique equal to 77.89%.

Fold recognition based on sequence-derived features is a complex classification problem and usually sequence-derived features are exploited using proper machine learning techniques. Here we adress the task of fold recognition on a protein similarity network (PSN) basis. We construct a protein sequence similarity network (PSeSN) using a set of 125 sequence-derived features for an available set of 311 proteins. PSeSN is optimized by using a Genetic Algorithm (GA) to select the features that construct a PSeSN which is as similar as possible with the corresponding protein structure similarity network (PStSN). A random walk based algorithm is then utilized to recognize the fold of a query protein sequence by calculating its affinities to sequences-vertices both in the initial and the optimized PSeSN. Total accuracy (TA) measurements obtained using 10-fold cross validation show that the use of 48 out of 125 sequence-derived features (optimized PSeSN) yielded better results (mean TA: 0.35 in testing sets) than the initial PSeSN (mean TA: 0.316 in testing sets).

UNLABELLED: The identification of unknown amino acid sequences of peptides as well as protein identification is of great significance in proteomics. Here, we present a publicly available web application that facilitates a high resolution mapping of measured molecular masses to peptides and proteins, irrespectively of the enzyme/digestion method used. Furthermore, multi-filtering may be applied in terms of measured mass tolerance, molecular mass and isoelectric point range as well as pattern matching to refine the results. This approach serves complementary to the existing solutions for protein identification and gives insights in novel peptides discovery and protein identification at the cases where the identification scores from the other approaches may be below significance threshold. Peptide Finder has been proven useful in proteomics procedures with experimental data from MALDI-TOF. AVAILABILITY: Peptide Finder web-application is available at http://bioserver-1.bioacademy.gr/Bioserver/PeptideFinder/.

SPECT images suffer from low contrast as a result of photons scatter. The standard method for excluding scatter component in pixelized scintillators is the application of an energy window around the central photopeak channel of each crystal cell, but small angle scattered photons still appear in the photopeak window and they are included in the reconstructed images. In this work we have assessed three subtraction techniques that use a different approach in order to calculate the scatter component and subtract it from the photopeak image. The dual energy window subtraction technique (DEWST) the convolution subtraction technique (CST) and a deconvolution technique (DT). All these techniques are compared to the standard method. The experimental results showed the superiority of DT, as a scatter correction technique.

Computer aided diagnosis is based on classification of medical data by intelligent classifiers. Especially for medical purposes, the classification must be very efficient, as diagnosis demands a high rate of reliability. Under most circumstances, single classifiers, such as neural networks, support vector machines and decision trees, exhibit worse performance than ensemble combinations of them such as bagging and boosting. In order to further enhance performance, we propose here a combination of these combination methods in a multi-level combination scheme. After experimentation by using four medical diagnosis problems, the proposed approach seems to be efficient in decreasing the error, compared to the best combining method standalone.

A computer aided diagnosis (CAD) system for the characterization of hepatic tissue from computed tomography (CT) images is presented. Regions of interest (ROI's) corresponding to four types of hepatic tissue are drawn by an experienced radiologist on abdominal non-enhanced CT images. For each ROI, five sets of texture features are extracted and combined to provide input to the CAD system. If the dimensionality of a feature set is greater than a predefined threshold, appropriate feature selection based on a genetic algorithm (GA) is applied. Classification of the ROI is then carried out using an ensemble of classifiers consisting of two neural network (NN) and three statistical classifiers. The final decision of the CAD system is based on the application of a voting scheme across the outputs of the primary classifiers of the ensemble. A classification performance of the order of 90.63% was finally achieved.

In this paper, a computer-aided diagnostic (CAD) system for the classification of hepatic lesions from computed tomography (CT) images is presented. Regions of interest (ROIs) taken from nonenhanced CT images of normal liver, hepatic cysts, hemangiomas, and hepatocellular carcinomas have been used as input to the system. The proposed system consists of two modules: the feature extraction and the classification modules. The feature extraction module calculates the average gray level and 48 texture characteristics, which are derived from the spatial gray-level co-occurrence matrices, obtained from the ROIs. The classifier module consists of three sequentially placed feed-forward neural networks (NNs). The first NN classifies into normal or pathological liver regions. The pathological liver regions are characterized by the second NN as cyst or "other disease." The third NN classifies "other disease" into hemangioma or hepatocellular carcinoma. Three feature selection techniques have been applied to each individual NN: the sequential forward selection, the sequential floating forward selection, and a genetic algorithm for feature selection. The comparative study of the above dimensionality reduction methods shows that genetic algorithms result in lower dimension feature vectors and improved classification performance.

The concept of in silica radiation oncology is clarified in this paper. A brief literature review points out the principal domains in which experimental, mathematical, and three-dimensional (3-D) computer simulation models of tumor growth and response to radiation therapy have been developed. Two paradigms of 3-D simulation models developed by our research group are concisely presented. The first one refers to the in vitro development and radiation response of a tumor spheroid whereas the second one refers to the fractionated radiation response of a clinical tumor in vivo based on the patient's imaging data. In each case, a description of the salient points of the corresponding algorithms and the visualization techniques used takes place. Specific applications of the models to experimental and clinical cases are described and the behavior of the models is two- and three-dimensionally visualized by using virtual reality techniques. Good qualitative agreement with experimental and clinical observations strengthens the applicability of the models to real situations. A protocol for further testing and adaptation is outlined. Therefore, an advanced integrated patient specific decision support and spatio-temporal treatment planning system is expected to emerge after the completion of the necessary experimental tests and clinical evaluation.

A simplified three-dimensional Monte Carlo simulation model of in vitro tumor growth and response to fractionated radiotherapeutic schemes is presented in this paper. The paper aims at both the optimization of radiotherapy and the provision of insight into the biological mechanisms involved in tumor development. The basics of the modeling philosophy of Duechting have been adopted and substantially extended. The main processes taken into account by the model are the transitions between the cell cycle phases, the diffusion of oxygen and glucose, and the cell survival probabilities following irradiation. Specific algorithms satisfactorily describing tumor expansion and shrinkage have been applied, whereas a novel approach to the modeling of the tumor response to irradiation has been proposed and implemented. High-performance computing systems in conjunction with Web technologies have coped with the particularly high computer memory and processing demands. A visualization system based on the MATLAB software package and the virtual-reality modeling language has been employed. Its utilization has led to a spectacular representation of both the external surface and the internal structure of the developing tumor. The simulation model has been applied to the special case of small cell lung carcinoma in vitro irradiated according to both the standard and accelerated fractionation schemes. A good qualitative agreement with laboratory experience has been observed in all cases. Accordingly, the hypothesis that advanced simulation models for the in silico testing of tumor irradiation schemes could substantially enhance the radiotherapy optimization process is further strengthened. Currently, our group is investigating extensions of the presented algorithms so that efficient descriptions of the corresponding clinical (in vivo) cases are achieved.

Retinal image registration is commonly required in order to combine the complementary information in different retinal modalities. In this paper, a new automatic scheme to register retinal images is presented and is currently tested in a clinical environment. The scheme considers the suitability and efficiency of different image transformation models and function optimization techniques, following an initial preprocessing stage. Three different transformation models--affine, bilinear and projective--as well as three optimization techniques--downhill simplex method, simulated annealing and genetic algorithms--are investigated and compared in terms of accuracy and efficiency. The registration of 26 pairs of Fluoroscein Angiography and Indocyanine Green Chorioangiography images with the corresponding Red-Free retinal images, showed the superiority of combining genetic algorithms with the affine and bilinear transformation models. A comparative study of the proposed automatic registration scheme against the manual method, commonly used in the clinical practice, is finally presented showing the advantage of the proposed automatic scheme in terms of accuracy and consistency.