Machine learning and integrative multi-omics network analysis for survival prediction in acute myeloid leukemia.

BACKGROUND: Acute myeloid leukemia (AML) is the most common malignant myeloid disorder in adults and the fifth most common malignancy in children, necessitating advanced technologies for outcome prediction. METHOD: This study aims to enhance prognostic capabilities in AML by integrating multi-omics data, especially gene expression and methylation, through network-based feature selection methodologies. By employing artificial intelligence and network analysis, we are exploring different methods to build a machine learning model for predicting AML patient survival. We evaluate the effectiveness of combining omics data, identify the most informative method for network integration and compare the performance with standard feature selection methods. RESULTS: Our findings demonstrate that integrating gene expression and methylation data significantly improves prediction accuracy compared to single omics data. Among network integration methods, our study identifies the best approach that improves informative feature selection for predicting patient outcomes in AML. Comparative analyses demonstrate the superior performance of the proposed network-based methods over standard techniques. CONCLUSIONS: This research presents an innovative and robust methodology for building a survival prediction model tailored to AML patients. By leveraging multilayer network analysis for feature selection, our approach contributes to improving the understanding and prognostic capabilities in AML and laying the foundation for more effective personalized therapeutic interventions in the future.

Assessment of Singularities in the EEG During A-Phases of Sleep Based on Wavelet Decomposition.

Electroencephalography (EEG) signals convey information related to different processes that take place in the brain. From the EEG fluctuations during sleep, it is possible to establish the sleep stages and identify short events, commonly related to a specific physiological process or pathology. Some of these short events (called A-phases) present an organization and build up the concept of the Cyclic Alternating Pattern (CAP) phenomenon. In general, the A-phases abruptly modify the EEG fluctuations, and a singular behavior could occur. With the aim to quantify the abrupt changes during A-phases, in this work the wavelet analysis is considered to compute Hölder exponents, which measure the singularity strength. We considered time windows of 2s outside and 5s inside A-phases onset (or offset). A total number of 5121 A-phases from 9 healthy participants and 10 patients with periodic leg movements were analyzed. Within an A-phase the Hölder numerical value tends to be 0.6, which implies a less abrupt singularity. Whereas outside of A-phases, it is observed that the Hölder value is approximately equal to 0.3, which implies stronger singularities, i.e., a more evident discontinuity in the signal behavior. In addition, it seems that the number of singularities increases inside of A-phases. The numerical results suggest that the EEG naturally conveys singularities modified by the A-phase occurrence, and this information could help to conceptualize the CAP phenomenon from a new perspective based on the sharpness of the EEG instead of the oscillatory way.

Network analysis with biological data of cancer patients: A scoping review.

BACKGROUND & OBJECTIVE: Network Analysis (NA) is a mathematical method that allows exploring relations between units and representing them as a graph. Although NA was initially related to social sciences, the past two decades was introduced in Bioinformatics. The recent growth of the networks' use in biological data analysis reveals the need to further investigate this area. In this work, we attempt to identify the use of NA with biological data, and specifically: (a) what types of data are used and whether they are integrated or not, (b) what is the purpose of this analysis, predictive or descriptive, and (c) the outcome of such analyses, specifically in cancer diseases. METHODS & MATERIALS: The literature review was conducted on two databases, PubMed & IEEE, and was restricted to journal articles of the last decade (January 2010 - December 2019). At a first level, all articles were screened by title and abstract, and at a second level the screening was conducted by reading the full text article, following the predefined inclusion & exclusion criteria leading to 131 articles of interest. A table was created with the information of interest and was used for the classification of the articles. The articles were initially classified to analysis studies and studies that propose a new algorithm or methodology. Each one of these categories was further screened by the following clustering criteria: (a) data used, (b) study purpose, (c) study outcome. Specifically for the studies proposing a new algorithm, the novelty presented in each one was detected. RESULTS: & Conclusions: In the past five years researchers are focusing on creating new algorithms and methodologies to enhance this field. The articles' classification revealed that only 25% of the analyses are integrating multi-omics data, although 50% of the new algorithms developed follow this integrative direction. Moreover, only 20% of the analyses and 10% of the newly developed methodologies have a predictive purpose. Regarding the result of the works reviewed, 75% of the studies focus on identifying, prognostic or not, gene signatures. Concluding, this review revealed the need for deploying predictive and multi-omics integrative algorithms and methodologies that can be used to enhance cancer diagnosis, prognosis and treatment.

A data-driven approach to build a predictive model of cancer patients' disease outcome by utilizing co-expression networks.

BACKGROUND: Next Generation Sequencing (NGS) technologies have revolutionized genomics data research over the last decades by facilitating high-throughput sequencing of genetic material such as RNA Sequencing (RNAseq). A significant challenge is to explore innovative methods for further exploitation of these large-scale datasets. The approach described in this paper utilizes the results of RNAseq analysis to identify biomarkers related to the disease and deploy a disease outcome predictive model. METHOD: Chronic Lymphocytic Leukemia (CLL) was used as an example in the implementation of this approach. The approach proposed follows this methodology: (1) Analysis of RNAseq raw data, (2) Construction of a gene correlation network, (3) Identification of modules and hub genes in this network, which constitute the features for the classification algorithm, (4) Deployment of an efficient predictive model, with the use of state-of-the-art machine learning techniques and the association of the indicators with the clinical information. RESULTS: The features/hub genes finally selected were 25 in total and were used as the input to the classifiers. The models, then, were validated leading to very satisfactory results, with the best performing of them achieving 95% cross-validation and 93,75% external validation accuracy. CONCLUSIONS: Concluding, this exploratory data-driven approach attempts to make use of big genomic data by summarizing them in a way that is more understandable and facilitates their use by other techniques, such as Machine Learning. This method manages to extract a gene set that can predict the disease progression. The validation results of the proposed data-driven predictive models are very promising and constitute a significant contribution to medical research and personalized medicine.

Nutrition Adherence in Critically Ill Patients; How is nutritional intake within the 1st week of hospitalization affecting the patient's Outcome?

Nutritional requirements vary during a patient's stay in the Intensive Care Unit (ICU) and their calculation can be relatively complex. During ICU stay nutrition requirements are rarely met, especially during the initial days of the hospitalization. Studies have shown that poor nutrition is associated with adverse patient outcome. This study examines for correlation between poor nutrition (calories, proteins, lipids and micronutrients) during the 1st week of ICU stay and adverse patient outcome. Nutritional adherence effect is examined on groups of patients, such as patients with high BMI that receive low nutrition and critically ill males. Regarding the latter analysis, an accuracy rate of 76.4% was achieved when classifying the critically ill males towards their outcome. The results of this work could contribute to the development of smart alarms in the ICU.

Prediction of the Adherence to a Home-Based Cardiac Rehabilitation Program.

The incidence and prevalence of cardiovascular diseases (CVD) is increasing which is partly due to an increase in unhealthy lifestyles, including lack of physical activity. Therefore, following a cardiovascular event, patients are encouraged to participate in a supervised exercise-based cardiac rehabilitation (CR) program. However, uptake rates of these programs are low and compliance to adequate volumes of physical activity after the completion of such programs are even lower. An approach that has been proposed towards the increase of patient adherence to exercise, is the incorporation of technology-enabled solutions which are applied at patient's homes. However, different factors may affect patient engagement with such alternative solutions. In this work, we use diverse types of data, including baseline characteristics of the patient (i.e. physiological, behavioral, demographical data) as well as usage data of a tele-rehabilitation solution during a 4-week familiarization period, in order to predict the compliance of patients with CVD to a technology-supported physical activity intervention after completion of a supervised exercise program. Patients were clustered based on their use of a technology intervention during a previously conducted study. Following a feature selection approach, a support vector machine was trained to classify patients as adherent or non-adherent to the intervention. The performance of the classifier was assessed by means of the receiving operator curve (ROC). Bio-psycho-social baseline variables predicted adherence with a ROC of 0.86, but adding usage data of the platform during a 4-week familiarization period increased the ROC up to 0.94. Furthermore, the high sensitivity values (83.8% and 95.5% respectively) support the strength of the models to identify those patients with CVD that will be adherent to a technology-enabled, home-based CR program.

Pre-diabetic patients with severe obstructive sleep apnea: novel parameters of hypoxia during sleep correlate with insulin resistance.

Obstructive sleep apnea (OSA) is an increasingly prevalent sleep disorder with estimated prevalence of moderate to severe sleep apnea ranging between 6-13% in the adult population. Nocturnal intermittent hypoxia has been associated with an increased risk of developing type 2 diabetes and insulin resistance (IR). The existing indices of hypoxia, used in polysomnography (PSG), cannot express accurately the influence of the mild desaturations precisely during sleep. In the present study, 73 subjects without other comorbidities were examined by PSG. Hypoxia parameters were analyzed, in the intervals with low values of SpO2 signal. The thresholds were set at 94% and 92% and the average value (M) of the SpO2 signal, in areas below thresholds, were calculated. The desaturations were analyzed together with their duration within the recording in terms of SpO2 signal parameters. Blood samples were taken for biochemical analysis. A total of 50 subjects were diagnosed with OSAS with mean AHI of 56.11±27.70/h and 23 subjects had no sleep apnea with mean AHI of 3.47±0.6/h. The amount of desaturations was significantly correlated with insulin levels (r=0.301, p=0.034) and the percentage of desaturation events (Dev) that were longer than 6 points (2 sec) (r=0.301, p=0.034). In addition, mean duration of desaturations was significantly correlated with IR as expressed by HOMA index (r=0.289, p=0.047), as well as with total duration of desaturation of SpO2 (r=0.322, p =0.025) and percentage of Dev that were long than 6 points (2 sec) (r=0.292, p=0.044). A strong correlation was also revealed between total duration of desaturations and fasting glucose (r=0.887, p=0.000). Results suggest that hypoxia parameters derived from SpO2 signal analysis, are strongly correlated with IR and fasting glucose levels, implying a role of hypoxia in the pathogenesis of diabetes.

Exploring the impact of sleep and stress on daily physical activity of cardiac patients: a preliminary study.

BACKGROUND: Current approaches to cardiac rehabilitation services tailoring are often based on patient demographics or readiness for behavior change. However, the success of interventions acceptance and improved adherence to recommendations could be much higher when considering and adapting to a patient's lifestyle, such as sleep and stress. AIMS: We aimed to analyze the potential associations between patient sleep and stress and daily moderate-intensity activity in patients with cardiovascular disease and to gain experience on the methods to collect and analyze a combination of qualitative and quantitative data. METHODS: Patients with cardiovascular disease enrolled for an outpatient cardiac rehabilitation program were assessed at the study baseline regarding sociodemographic, clinical profile, and perceived level of stress. To collect daily physical activity and sleep data, all participants had two-week long diaries. Collected data was analyzed through correlation analysis, linear regression, and one-way ANOVA analysis. RESULTS: The mean age of the participants (n =11) was 67.3 ± 9.6 years old. The patients were mainly male (82 %), married (91 %), and having at least one comorbid disease (64 %). The results of the analysis revealed that the night sleep duration is associated with moderate-intensity physical activity [F(1,6) =7.417, p =0.034]. Stress was not associated with patients' moderate-intensity daily physical activity. CONCLUSION: The outcomes of the study can support the development of e-health and home-based interventions design and strategies to promote adherence to physical activity. Tailoring an intervention to a daily behavioral pattern of a patient, such as sleep, can support the planning of the physical activity in a form to be easier accepted by the patient. This finding emphasizes the need for further investigation of the association with a larger population sample and the use of objective physical activity and sleep-related measure instruments. HIPPOKRATIA 2019, 23(1): 15-20.

Linear and nonlinear features of fetal heart rate on the assessment of fetal development in the course of pregnancy and the impact of fetal gender.

OBJECTIVE: This work aims to investigate the impact of gestational age and fetal gender on fetal heart rate (FHR) tracings. APPROACH: Different linear and nonlinear parameters indicating correlation or complexity were used to study the influence of fetal age and gender on FHR tracings. The signals were recorded from 99 normal pregnant women in a singleton pregnancy at gestational ages from 28 to 40 weeks, before the onset of labor. There were 56 female fetuses and 43 male. MAIN RESULTS: Analysis of FHR shows that the means as well as measures of irregularity of FHR, such as approximate entropy and algorithmic complexity, decrease as gestation progresses. There were also indications that mutual information and multiscale entropy were lower in male fetuses in early pregnancy. SIGNIFICANCE: Fetal age and gender seem to influence FHR tracings. Taking this into consideration would improve the interpretation of FHR monitoring.

Detecting fetal heart sounds by means of Fractal Dimension analysis in the Wavelet domain.

Phonocardiography is a low-cost technique for the detection of fetal heart sounds (FHS) that can extend clinical auscultation in mobile and home care setups. The work presented here examines the transferability of a Wavelet Transform (WT)-based method that combines also Fractal Dimension (FD) analysis, previously proposed as WT-FD for the cases of lung and bowel sound analysis [4], to the extraction of FHSs. The WT-FD method has been evaluated with 12 simulated FHS signals and has shown promising results in terms of accuracy and performance (89%) in identifying the location of heartbeat, even in cases of signals with additive noise up to (6dB). This robustness paves the way for WT-FD testing in real FHSs, recorded under clinical setting, clearly contributing to better evaluation of the fetal heart functionality.

Restrictions in the T-cell repertoire of chronic lymphocytic leukemia: high-throughput immunoprofiling supports selection by shared antigenic elements.

Immunoglobulin (IG) gene repertoire restrictions strongly support antigen selection in the pathogenesis of chronic lymphocytic leukemia (CLL). Given the emerging multifarious interactions between CLL and bystander T cells, we sought to determine whether antigen(s) are also selecting T cells in CLL. We performed a large-scale, next-generation sequencing (NGS) study of the T-cell repertoire, focusing on major stereotyped subsets representing CLL subgroups with undisputed antigenic drive, but also included patients carrying non-subset IG rearrangements to seek for T-cell immunogenetic signatures ubiquitous in CLL. Considering the inherent limitations of NGS, we deployed bioinformatics algorithms for qualitative curation of T-cell receptor rearrangements, and included multiple types of controls. Overall, we document the clonal architecture of the T-cell repertoire in CLL. These T-cell clones persist and further expand overtime, and can be shared by different patients, most especially patients belonging to the same stereotyped subset. Notably, these shared clonotypes appear to be disease-specific, as they are found in neither public databases nor healthy controls. Altogether, these findings indicate that antigen drive likely underlies T-cell expansions in CLL and may be acting in a CLL subset-specific context. Whether these are the same antigens interacting with the malignant clone or tumor-derived antigens remains to be elucidated.

Influence of torso and arm positions on chest examinations by electrical impedance tomography.

Electrical impedance tomography (EIT) is increasingly used in patients suffering from respiratory disorders during pulmonary function testing (PFT). The EIT chest examinations often take place simultaneously to conventional PFT during which the patients involuntarily move in order to facilitate their breathing. Since the influence of torso and arm movements on EIT chest examinations is unknown, we studied this effect in 13 healthy subjects (37 ± 4 years, mean age ± SD) and 15 patients with obstructive lung diseases (72 ± 8 years) during stable tidal breathing. We carried out the examinations in an upright sitting position with both arms adducted, in a leaning forward position and in an upright sitting position with consecutive right and left arm elevations. We analysed the differences in EIT-derived regional end-expiratory impedance values, tidal impedance variations and their spatial distributions during all successive study phases. Both the torso and the arm movements had a highly significant influence on the end-expiratory impedance values in the healthy subjects (p = 0.0054 and p < 0.0001, respectively) and the patients (p < 0.0001 in both cases). The global tidal impedance variation was affected by the torso, but not the arm movements in both study groups (p = 0.0447 and p = 0.0418, respectively). The spatial heterogeneity of the tidal ventilation distribution was slightly influenced by the alteration of the torso position only in the patients (p = 0.0391). The arm movements did not impact the ventilation distribution in either study group. In summary, the forward torso movement and the arms' abduction exert significant effects on the EIT waveforms during tidal breathing. We recommend strict adherence to the upright sitting position during PFT when EIT is used.

Multiparametric modeling of the ineffective efforts in assisted ventilation within an ICU.

In the context of assisted ventilation in ICU, it is of vital importance to keep a high synchronization between the patient's attempt to breath and the assisted ventilation event, so that the patient receives the ventilation support requested. In this work, experimental equipment is employed, which allows for unobtrusive and continuous monitoring of a multiple relevant bioparameters. These are meant to guide the medical professionals in appropriately adapting the treatment and fine-tune the ventilation. However, synchronization phenomena of different origin (neurological, mechanical, ventilation parameters) may occur, which vary among patients, and during the course of monitoring of a single patient, the timely recognition of which is challenging even for experts. The dynamics and complex causal relations among bioparameters and the ventilation synchronization are not well studied. The purpose of this work is to elaborate on a methodology toward modeling the ventilation synchronization failures based on the evolution of monitored bioparameters. Principal component analysis is employed for the transformation into a small number of features and the investigation of repeating patterns and clusters within measurements. Using these features, nonlinear prediction models based on support vector machines regression are explored, in terms of what past knowledge is required and what is the future horizon that can be predicted. The proposed model shows good correlation (over 0.74) with the actual outputs, constituting an encouraging step toward understanding of ICU ventilation dynamic phenomena.

Detection of crackle events using a multi-feature approach.

The automatic detection of adventitious lung sounds is a valuable tool to monitor respiratory diseases like chronic obstructive pulmonary disease. Crackles are adventitious and explosive respiratory sounds that are usually associated with the inflammation or infection of the small bronchi, bronchioles and alveoli. In this study a multi-feature approach is proposed for the detection of events, in the frame space, that contain one or more crackles. The performance of thirty-five features was tested. These features include thirty-one features usually used in the context of Music Information Retrieval, a wavelet based feature as well as the Teager energy and the entropy. The classification was done using a logistic regression classifier. Data from seventeen patients with manifestations of adventitious sounds and three healthy volunteers were used to evaluate the performance of the proposed method. The dataset includes crackles, wheezes and normal lung sounds. The optimal detection parameters, such as the number of features, were chosen based on a grid search. The performance of the detection was studied taking into account the sensitivity and the positive predictive value. For the conditions tested, the best results were obtained for the frame size equal to 128 ms and twenty-seven features.

Detection of different types of noise in lung sounds.

Lung sound signal processing has proven to be a great improvement to the traditional acoustic interpretation of lung sounds. However, that analysis can be seriously hindered by the presence of different types of noise originated in the acquisition environment or caused by physiological processes. Consequently, the diagnostic accuracy of pulmonary diseases can be severely affected, especially if the implementation of telemonitoring systems is considered. The present study is focused on the implementation of an algorithm able to identify noisy periods, either voluntarily (vocalizations, chest movement and background voices) or involuntarily produced during acquisitions of lung sounds. The developed approach also had to deal with the presence of simulated cough events, that carry important diagnostic information regarding several pulmonary diseases. Features such as Katz fractal dimension, Teager-Kaiser energy operator and normalized mutual information, were extracted from the time domain of healthy and a pathological lung signals. Noise detection was the result of a good discrimination between uncontaminated lung sounds and both cough and noise episodes and a slightly worse classification of cough events. In fact, detection of cough periods carrying diagnostic information was influenced by the presence of two other types of noise having similar signal characteristics.

A data model to support the evaluation of coordinating care EU programmes in the context of the ACT programme.

Advancing Care Coordination and Telehealth Deployment (ACT) is a European Union (EU) project, completed last October, which has developed a framework for evaluating and improving pioneering health care programs regarding coordinating care and telehealth (CC & TH) across specific EU regions. In this paper we present the key design decisions of the project's data model and the challenges faced. We focus on the definition of the multi-dimensional indicators in order to overcome data incompleteness and heterogeneity issues. Finally, we also suggest a graph based approach that could facilitate development of such data models in similar projects.

Evaluation of lung ventilation distribution in chronic obstructive pulmonary disease patients using the global inhomogeneity index.

The global inhomogeneity (GI) index is a electrical impedance tomography (EIT) parameter that quantifies the tidal volume distribution within the lung. In this work the global inhomogeneity index was computed for twenty subjects in order to evaluate his potential use in the detection and follow up of chronic obstructive pulmonary disease (COPD) patients. EIT data of 17 subjects were acquired: 14 patients with the main diagnoses of COPD and 3 healthy subjects which served as a control group. Two or three datasets of around 30 seconds were acquired at 33 scans/s and analysed for each subject. After reconstruction, a tidal EIT image was computed for each breathing cycle and a GI index calculated from it. Results have shown significant differences in GI values between the two groups (0.745 ± 0.007 for COPD and 0.668 ± 0.006 for lung-healthy subject, p <; 0.005). The GI values obtained for each subject have shown small variance between them, which is a good indication of stability. The results suggested that the GI may be useful for the identification and follow up of ventilation problems in patients with COPD.

Medical Informatics Education & Research in Greece.

OBJECTIVES: This paper aims to present an overview of the medical informatics landscape in Greece, to describe the Greek ehealth background and to highlight the main education and research axes in medical informatics, along with activities, achievements and pitfalls. METHODS: With respect to research and education, formal and informal sources were investigated and information was collected and presented in a qualitative manner, including also quantitative indicators when possible. RESULTS: Greece has adopted and applied medical informatics education in various ways, including undergraduate courses in health sciences schools as well as multidisciplinary postgraduate courses. There is a continuous research effort, and large participation in EU-wide initiatives, in all the spectrum of medical informatics research, with notable scientific contributions, although technology maturation is not without barriers. Wide-scale deployment of eHealth is anticipated in the healthcare system in the near future. While ePrescription deployment has been an important step, ICT for integrated care and telehealth have a lot of room for further deployment. CONCLUSIONS: Greece is a valuable contributor in the European medical informatics arena, and has the potential to offer more as long as the barriers of research and innovation fragmentation are addressed and alleviated.

Ineffective efforts in ICU assisted ventilation: Exploring causalities via multiscale analysis.

Intensive Care Unit (ICU) is a data intensive environment, requiring continuous monitoring of patient's physiology and response to treatment. In assisted ventilation, where patient effort that triggers the ventilator and there is need for patient-ventilator coupling, attention is required in cases where patient's effort that doesn't trigger the ventilator at all. When synchronization between the patient's attempt to breath and the assisted ventilation event is lost, an ineffective effort (IE) event takes place. A series of relevant bioparameters continuously monitored, are meant to guide the medical professionals in appropriately adapting the operation and treatment, in order to minimize IEs. The purpose of this work is to investigate the causal relations between physiological or ventilation parameters and IE events. A multiscale approach is proposed, based on wavelet similarity and localized phase relationship. The proposed method indicates the existence of distinct frequency zones correlated with the IE experienced by the patient.

Detection of wheezes using their signature in the spectrogram space and musical features.

In this work thirty features were tested in order to identify the best feature set for the robust detection of wheezes. The features include the detection of the wheezes signature in the spectrogram space (WS-SS) and twenty-nine musical features usually used in the context of Music Information Retrieval. The method proposed to detect the signature of wheezes imposes a temporal Gaussian regularization and a reduction of the false positives based on the (geodesic) morphological opening by reconstruction operator. Our dataset contains wheezes, crackles and normal breath sounds. Four selection algorithms were used to rank the features. The performance of the features was asserted having into account the Matthews correlation coefficient (MCC). All the selection algorithms ranked the WS-SS feature as the most important. A significant boost in performance was obtained by using around ten features. This improvement was independent of the selection algorithm. The use of more than ten features only allows for a small increase of the MCC value.