An interpretable machine learning approach to estimate the influence of inflammation biomarkers on cardiovascular risk assessment

BACKGROUND AND OBJECTIVE

Cardiovascular disease has a huge impact on health care services, originating unsustainable costs at clinical, social, and economic levels. In this context, patients' risk stratification tools are central to support clinical decisions contributing to the implementation of effective preventive health care. Although useful, these tools present some limitations, in particular, some lack of performance as well as the impossibility to consider new risk factors potentially important in the prognosis of severe cardiac events. Moreover, the actual use of these tools in the daily practice requires the physicians' trust. The main goal of this work addresses these two issues: (i) evaluate the importance of inflammation biomarkers when combined with a risk assessment tool; (ii) incorporation of personalization and interpretability as key elements of that assessment.

METHODS

Firstly, machine learning based models were created to assess the potential of the inflammation biomarkers applied in secondary prevention, namely in the prediction of the six month risk of death/myocardial infarction. Then, an approach based on three main phases was created: (i) set of interpretable rules supported by clinical evidence; (ii) selection based on a machine learning classifier able to identify for a given patient the most suitable subset of rules; (iii) an ensemble scheme combining the previous subset of rules in the estimation of the patient cardiovascular risk. All the results were statistically validated (t-test, Wilcoxon-signed rank test) according to a previous verification of data normality (Shapiro-Wilk).

RESULTS

The proposed methodology was applied to a real acute coronary syndrome patients dataset (N = 1544) from the Cardiology Unit of Coimbra Hospital and Universitary centre. The first assessment was based on the GRACE tool and a Random Forest classifier, the incorporation of inflammation biomarkers achieved SE=0.83; SP=0.84 whereas the original GRACE risk factors reached SE=0.75; SP=0.85. In the second phase, the proposed approach with inflammation biomarkers achieved SE=0.763 and SP=0.778.

CONCLUSIONS

This approach confirms the potential of combining inflammation markers with the GRACE score, increasing SE and SP, when compared with the original GRACE. Additionally, it assures interpretability and personalization, which are critical issues to allow its application in the daily clinical practice.

A Fundamental Study on Compression Properties and Strain Rate Sensitivity of Spray-Dried Amorphous Solid Dispersions

Amorphous solid dispersions (ASDs) are a proven method of improving the solubility and bioavailability of poorly soluble compounds. Immediate release tablets are frequently used as final dosage form for ASDs. Increasing tableting process throughput during clinical development requires using larger, faster tablet presses which subject materials to higher strain rate. Many pharmaceutical materials show strain rate sensitivity, i.e., yield pressure sensitivity to compression speed. Currently, there is only scattered information available in scientific literature on how ASDs behave under different tablet compression speeds. The purpose of this study was to examine spray-dried ASDs' sensitivity to strain rate under compression in a comprehensive study. We also investigated the drivers for such a strain sensitive behavior. A set of sample spray-dried powders, selected for their range of properties, were compressed using a simulated Korsch XL100 profile at 3 and 30 RPM and V-profile at 0.1 and 300 mm/s on a Phoenix compaction simulator. The sample set included samples with varying API content (0-50% w/w), stabilizing polymer (HPMC, HPMC-AS, PVP-VA), particle size, and bulk densities, produced on spray driers from lab to commercial scale. We identified that all ASD samples showed plastic flow and deformation behavior and form robust compacts at slow compression speeds. At high speed, tablet defects occurred. The strain rate sensitivity observed in this study was comparable or slightly superior to that observed for microcrystalline cellulose, known to be a mildly strain rate-sensitive material. We showed that compression speed is a critical process parameter for ASD-containing tablets.

The quest for GRACE 3.0: improving our beloved risk score with machine learning

Background

Although widely recommended for risk assessment of patients with acute coronary syndrome (ACS), the Global Registry of Acute Coronary Events (GRACE) score famously lacks discriminative power. On the other hand, in-hospital serum hemoglobin levels (HG) have been shown to simultaneously forecast both thrombotic and hemorrhagic hazards.

Purpose

To ascertain the extent to which the incorporation of HG in the GRACE score is able to increase its predictive ability.

Methods

Retrospective single-center study encompassing ACS patients consecutively admitted to a Cardiac Intensive Care Unit. Inclusion criteria comprised the acquaintance of GRACE score, HG and vital status on a 6-month follow-up, which served as the outcome. 3 discriminative models were first created: (standard) GRACE score (model 1); GRACE score plus HG, by means of logistic regression (model 2); GRACE score plus HG, by means of multilayer perceptron (a class of feedforward artificial neural network) (model 3). Hereafter, if models 2 and/or 3 were to be found significantly more discriminative than model 1, a correction factor would be calculated, also allowing for the conception of the most predictive model possible (model 4). The discriminative ability was estimated by both the area under the receiver-operating characteristic curve (AUC), and the dyad sensitivity/specificity.

Results

Between April 2009 and December 2016, 1468 patients met study inclusion criteria. Mean age was 68.0±13.2 years and 29.8% were female, while 36.9% presented with ST-segment elevation myocardial infarction. Mean GRACE score was 145.5±47.0 and mean HG was 13.5±2.0. All-cause mortality reached 10.5%, at 6 months. Predictive power for models 1, 2 and 3 may be quantified as follows: AUC 0.6998, sensitivity 77.7% and specificity 62.5%; AUC 0.7818, sensitivity 36.3% and specificity 92.2%; AUC 0.7851, sensitivity 47.7% and specificity 88.5%, respectively. Both models 2 and 3 exhibited more discriminative ability than model 1 (p<0.001), due to their higher specificity. As such, a correction factor was computed (y = −7.8556x + 86.4117) and model 4 was created, displaying a sensitivity of 65.9% and a specificity of 76.5%.

Conclusion

HG single-handedly provides incremental predictive value – namely more specificity – to the GRACE score. In particular, the latter seems to overestimate ACS patients' risk if HG is normal or close to normal.

Funding Acknowledgement

Type of funding sources: None.

Emotional prosody perception in presbycusis patients after auditory rehabilitation

OBJECTIVE

Perception of emotion plays a major role in social interaction. Studies have shown that hearing loss and aging degrade emotional recognition. The main aim of the present study was to evaluate the benefit of first-time hearing aids (HA) for emotional prosody perception in presbycusis patients. Secondary objectives comprised comparison with normal-hearing subjects, and assessment of the impact of demographic and audiologic factors.

METHODS

To assess HA impact, 29 subjects with presbycusis were included. They were tested without HA and 1 month after starting to use HA. A test with emotional hearing stimuli (Montreal Affective Voice test: MAV) was performed at various intensities (50, 65 and 80dB SPL). Patients' experience was evaluated on the Profile of Emotional Competence questionnaire, before and after HA fitting. Results were compared with those of 29 normal-hearing subjects.

RESULTS

Auditory rehabilitation did not significantly improve MAV results (P>0.005), or subjective questionnaire results (P>0.005). Scores remained lower than those of normal-hearing subjects (P<0.001). MAV results, before and after HA, showed significant correlation with pure-tone average (r=-0.88, P<0.001) and age (r=0.44, P=0.018). The older the presbycusis patient and the more severe the hearing loss, the greater the difficulty in recognising emotional prosody.

CONCLUSION

Despite hearing rehabilitation, presbycusis patients' results remained poorer than in normal-hearing subjects.

Non Type-1 Brugada pattern, diagnostic yield of 5 ECG criteria in a young adult cohort

Introduction

Distinguishing between non-Type 1 Brugada pattern (non-T1BrP) and an athlete's ECG remains challenging and may have important prognostic implications. We aimed to study prevalence and the diagnostic yield of experts and non-experts for the electrocardiographic non-T1BrP criteria in the young adults from the Sudden Cardiac Death-Screening Of risk factorS (SCD-SOS) cohort.

Methods

We performed a cross-sectional study in which we reviewed 14662 ECGs of SCD-SOS survey participants and selected 2494 that presented a rSr'-pattern in V1-V2. Among these, 98 were classified by experts in hereditary arrhythmic syndromes for the presence of non-T1BrP and by non-experts who performed manual measurements of the diagnostic criteria based on triangle formed by r'-wave. We estimated intra and interobserver concordance for each criterion, and used logistic regression and receiver operating characteristics (ROC) analysis and C-statistics for diagnostic accuracy and definition of the most appropriate cut-off values.

Results

We detected a rSr'-pattern in V1-V2 in 17% of the individuals and found that it was associated with higher PQ and QTc intervals, male gender and lower BMI. The manual measurements of non-T1BrP criteria were reproducible: we had high intraobserver concordance coefficients (CC) ranging from 0.90 to 0.94 (except for d(B) that had 0.66), but interobserver CC were lower (0.45–0.68). The measurements performed were highly correlated with non-T1BrP diagnosis and the criteria with higher discriminatory capacity were the distance d(B) (AUC 0.77; 95% CI0.69–0.84) and the degree of ST-ascent (AUC 0.79; 95% CI 0.72–0.86). The cut-offs defined by other authors had very low sensitivity (8–12%), despite high specificity (98%), so we defined new cut-offs: d(A) ≥2mm, d(B) ≥1.25mm, d(B)/h ≥0.38, β-angle ≥19° and ST-ascent ≥1mm. The addition of the degree of ST-ascent to a model with these 4 parameters presented an increase in C-statistics from 0.77 (95% CI: 0.68–0.86) to 0.83 (95% CI: 0.75–0.91) for the diagnosis of non-T1BrP by an expert in Sudden Arrhythmic Death and Channelopathies.

Conclusion

A rSr'-pattern in precordial leads V1-V2 is a frequent finding and the detection of non-T1BrP by using the aforementioned 5 measurements is reproducible and accurate. In this study, we describe new cut-off values that may help untrained clinicians to identify young individuals who should be referred for provocative drug testing for Brugada Syndrome.

Accuracy of non-T1 BrP criteria

Funding Acknowledgement

Type of funding source: None

Simultaneous Synthesis and Nitrogen Doping of Free-Standing Graphene Applying Microwave Plasma

An experimental and theoretical investigation on microwave plasma-based synthesis of free-standing N-graphene, i.e., nitrogen-doped graphene, was further extended using ethanol and nitrogen gas as precursors. The in situ assembly of N-graphene is a single-step method, based on the introduction of N-containing precursor together with carbon precursor in the reactive microwave plasma environment at atmospheric pressure conditions. A previously developed theoretical model was updated to account for the new reactor geometry and the nitrogen precursor employed. The theoretical predictions of the model are in good agreement with all experimental data and assist in deeper understanding of the complicated physical and chemical process in microwave plasma. Optical Emission Spectroscopy was used to detect the emission of plasma-generated ‘‘building units’’ and to determine the gas temperature. The outlet gas was analyzed by Fourier-Transform Infrared Spectroscopy to detect the generated gaseous by-products. The synthesized N-graphene was characterized by Scanning Electron Microscopy, Raman, and X-ray photoelectron spectroscopies.

Prognostic and predictive value of the Immunoscore in stage III colon cancer patients treated with oxaliplatin in the prospective IDEA France PRODIGE-GERCOR cohort study

BACKGROUND

The Immunoscore (IS), which prognostically classifies stage I-III colon cancer (CC) patients, was evaluated in the International Duration Evaluation of Adjuvant Therapy (IDEA) France cohort study investigating 3 versus 6 months of oxaliplatin-based adjuvant chemotherapy in stage III CC patients.

PATIENTS AND METHODS

Densities of CD3+ and CD8+ T cells in the tumor and invasive margin were determined by immunohistochemistry, quantified by digital pathology, and converted to IS. Mismatch repair status was determined by immunohistochemistry or by pentaplex PCR. Prediction of disease-free survival (DFS) by IS was analyzed by a multivariable Cox regression model in each study arm. Harrell's C-statistics were used to investigate the IS performance.

RESULTS

Samples of 1322 patients were available. IS Low, Intermediate (Int), and High were observed in 43.6%, 47.0%, and 9.4% of patients, respectively. IS Low identified patients at higher risk of relapse or death compared with Int + High [hazard ratio (HR) = 1.54; 95% confidence interval (CI) 1.24-1.93, P = 0.0001]. The 3-year DFS was 66.80% (95% CI 62.23-70.94) for IS Low and 77.14% (95% CI 73.50-80.35) for IS Int + High. In multivariable analysis, IS remained significantly independently associated with DFS (P = 0.003) when adjusted for sex, histological grade, T/N stage, and microsatellite instability. For mFOLFOX6-treated patients (91.6% of the cohort), a statistical significant interaction was observed for the predictive value of IS for treatment duration (3 versus 6 months) in terms of DFS (P = 0.057). IS Int + High significantly predicted benefit of 6 months of treatment (HR = 0.53; 95% CI 0.37-0.75; P = 0.0004), including clinically low- and high-risk stage III CC (all P < 0.001). Conversely, patients with IS Low (46.4%) did not significantly benefit from the 6-month mFOLFOX6 versus the 3-month mFOLFOX6.

CONCLUSIONS

The prognostic value of IS for DFS was confirmed in patients with stage III CC treated with oxaliplatin-based chemotherapy. Its predictive value for DFS benefit of longer duration of mFOLFOX6 adjuvant treatment was found in IS Int + High. These results will be validated in an external independent cohort. CLINICALTRIALS.

GOV REGISTRATION

NCT03422601; EudraCT Number: 2009-010384-16.

Latent states extraction through Kalman Filter for the prediction of heart failure decompensation events

Cardiac function deterioration of heart failure patients is frequently manifested by the occurrence of decompensation events. One relevant step to adequately prevent cardiovascular status degradation is to predict decompensation episodes in order to allow preventive medical interventions.In this paper we introduce a methodology with the goal of finding onsets of worsening progressions from multiple physiological parameters which may have predictive value in decompensation events. The best performance was obtained for the model composed by only two features using a telemonitoring dataset (myHeart) with 41 patients. Results were achieved by applying leave-one-subject-out validation and correspond to a geometric mean of 83.67%. The obtained performance suggests that the methodology has the potential to be used in decision support solutions and assist in the prevention of this public health burden.

Microwave plasma-based direct synthesis of free-standing N-graphene

Scheme of ethanol/ammonia plasma driven decomposition pathways considering injection of the nitrogen precursor in “hot” and “mild” plasma zone.

A Clinical Interpretable Approach Applied to Cardiovascular Risk Assessment

The effectiveness of predictive models in supporting the Clinical Decision is closely related with their clinical interpretability, i.e.the model should provide clear information on how to reach a specific classification/decision. In fact, the development of interpretable and accurate predictive models assumes a key importance as these tools can be very useful in Clinical Decision Support Systems (CDSS). The development of those models may comprise two main perspectives; existent clinical knowledge (clinical expert knowledge, clinical guidelines, current models, etc.) as well as data driven approaches able to extract (new) knowledge from recent clinical datasets. This work focuses in knowledge extraction from recent datasets (data driven) based on computational intelligence techniques. The main hypothesis that supports this work is that individuals with similar characteristics present a similar risk prof ile. Thus, this work addresses the development of stratification models able to learn distinct groups (or classes) of subjects assessing the similarity between characterizing variables. In particular, in the current study a data-driven supervised cluster approach is proposed aiming the derivation of meaningful rules directly from the dataset. The validation was performed based on the largest Portuguese coronary artery disease patient's dataset, provided by the Portuguese Society of Cardiology and comprising 13902 acute coronary syndrome patients. The goal was to assess the risk of death 30 days after admission. The models' performance was assessed through the sensitivity, specificity and geometric mean values. The obtained results show the potential of this approach, as they represent an acceptable performance (GM= 72%) while the clinical interpretability of the model is assured through the derived rules. Despite the achieved results, there are several research directions to be followed in order to enhance this work.

Risk Prediction of Heart Failure Decompensation Events in Multiparametric Feature Spaces

Cardiac function deterioration of heart failure patients is frequently manifested by the occurrence of decompensation events. One relevant step to adequately prevent cardiovascular status degradation is to predict decompensation episodes in order to allow preventive medical interventions. In this paper we introduce a methodology with the goal of finding relevant feature spaces from multiple physiological parameters which may have predictive value in decompensation events. The best performance was obtained for the feature space comprising the following features: mean weight, standard deviation of the blood pressure and mean of extra-thoracic impedance in a time window of 20 days. Results were achieved by applying leave-one-out validation and correspond to a geometric mean of 88.32%. The obtained performance suggests that the methodology has the potential to be used in decision support solutions and assist in the prevention of this public health burden.

Large-scale synthesis of free-standing N-doped graphene using microwave plasma

Direct assembling of N-graphene, i.e. nitrogen doped graphene, in a controllable manner was achieved using microwave plasmas at atmospheric pressure conditions. The synthesis is accomplished via a single step using ethanol and ammonia as carbon and nitrogen precursors. Tailoring of the high-energy density plasma environment results in a selective synthesis of N-graphene (~0.4% doping level) in a narrow range of externally controlled operational conditions, i.e. precursor and background gas fluxes, plasma reactor design and microwave power. Applying infrared (IR) and ultraviolet (UV) irradiation to the flow of free-standing sheets in the post-plasma zone carries out changes in the percentage of sp², the N doping type and the oxygen functionalities. X-ray photoelectron spectroscopy (XPS) revealed the relative extension of the graphene sheets π-system and the type of nitrogen chemical functions present in the lattice structure. Scanning Electron microscopy (SEM), Transmission Electron microscopy (TEM) and Raman spectroscopy were applied to determine morphological and structural characteristics of the sheets. Optical emission and FT-IR spectroscopy were applied for characterization of the high-energy density plasma environment and outlet gas stream. Electrochemical measurements were also performed to elucidate the electrochemical behavior of NG for supercapacitor applications.

Physiologically motivated detection of Atrial Fibrillation

Atrial Fibrillation (AF) is the most common arrhythmia and it is estimated to affect 33.5 million people worldwide. AF is associated with an increased risk of mortality and morbidity, such as heart failure and stroke and affects mostly older persons and persons with other conditions (e.g. heart failure and coronary artery disease). In order to prevent such life threatening and life quality reducing conditions it is essential to provide better algorithms, capable of being integrated in low-cost personalized health systems. This paper presents a new algorithm for AF detection, which is based on the analysis of the three physiological characteristics of AF: 1) Irregularity of heart rate and; 2) Absence of P-waves and 3) Presence of fibrillatory waves. Based on these characteristics several features were extracted from 12-lead electrocardiograms (ECG) and selected according to their discrimination ability. The classification between AF and non-AF episodes was performed using a Support Vector Machine (SVM) classification model. Our results show that the identification of the fibrillatory patterns, using the proposed features, extracted from the analysis of 12-lead ECG improves the performance of the algorithm to a sensitivity of 88.5% and specificity 92.9%, when compared to our previous single-channel approach, in the same database.

Intravenous Immunoglobulin and Rituximab in HLA Highly Sensitized Kidney Transplant Recipients

INTRODUCTION

HLA-sensitized patients are penalized both in the access to kidney transplantation (KT) and, once transplanted, in the incidence of rejections and long-term allograft survival despite aggressive induction and maintenance therapy.

METHODS

This study retrospectively evaluates the impact of combining T- and B-cell-depleting agents and intravenous immunoglobulin for induction therapy in 45 highly sensitized KT patients (anti-panel reactive antibodies >60%, positive flow cytometry crossmatch or donor specific antibodies at the time of transplantation). The outcome data included the occurrence of biopsy-proven acute rejection, new-onset proteinuria, development of leukopenia, incidence of poliomavirus infection (BK or JC virus), fungal or bacterial infection after KT, de novo neoplasia, graft function, graft loss, or death with functioning KT.

RESULTS

The average panel reactive antibody was 62.5%; 41 patients (91.1%) had ≥3 HLA mismatches with the donor and 91.1% of patients had class I or II anti-HLA antibodies. Fourteen patients (31.1%) presented pre-KT donor-specific antibodies and 6 patients (13.3%) had a positive flow cytometry cross-match at the time of transplantation. The incidence of acute rejection in the first 6 months was 24.4% and the cumulative incidence was 37.8%. Two patients were diagnosed with leukopenia in the first 6 months after KT. Two patients (4.5%) had cytomegalovirus disease, 17 patients (37.8%) were diagnosed with bacterial infections. Cutaneous neoplasms were identified in 5 patients (11.1%) and solid tumors in 4 (8.9%). The death-censored graft survival was 100% in the first 6 months and 93.5% at the last evaluation. Patient survival in the same periods was 97.8% and 93.3%, respectively.

CONCLUSIONS

Induction immunosuppressive therapy with intravenous immunoglobulin and rituximab is effective; outcomes demonstrate an excellent patient and allograft survival and allograft function over the follow-up period.

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.

Microwave plasma enabled synthesis of free standing carbon nanostructures at atmospheric pressure conditions

Schematic representation of the physico-chemical processes involved in the formation of carbon nanostructures in the microwave plasma environment.

Detection of explosive cough events in audio recordings by internal sound analysis

We present a new method for the discrimination of explosive cough events, which is based on a combination of spectral content descriptors and pitch-related features. After the removal of near-silent segments, a vector of event boundaries is obtained and a proposed set of 9 features is extracted for each event. Two data sets, recorded using electronic stethoscopes and comprising a total of 46 healthy subjects and 13 patients, were employed to evaluate the method. The proposed feature set is compared to three other sets of descriptors: a baseline, a combination of both sets, and an automatic selection of the best 10 features from both sets. The combined feature set yields good results on the cross-validated database, attaining a sensitivity of 92.3±2.3% and a specificity of 84.7±3.3%. Besides, this feature set seems to generalize well when it is trained on a small data set of patients, with a variety of respiratory and cardiovascular diseases, and tested on a bigger data set of mostly healthy subjects: a sensitivity of 93.4% and a specificity of 83.4% are achieved in those conditions. These results demonstrate that complementing the proposed feature set with a baseline set is a promising approach.

Artifact detection in accelerometer signals acquired from the carotid

Manual pulse palpation is the common procedure to assess pulse in unconscious patients. This is an error prone procedure during cardiopulmonary resuscitation and therefore automatic pulse detection techniques are being investigated. Accelerometry is an interesting sensing modality for this type of applications. However, accelerometers are highly prone to movement artifacts. Hence, one challenge in designing a solution using accelerometers is to handle motion artifacts properly. In this paper we investigate computationally simple features and classifier to capture movement artifacts in accelerometer signals acquired from the carotid. In particular, based on data obtained from health subjects we show that it is possible to use simple features to achieve an artifact detection sensitivity and specificity higher than 90%.

Pattern discovery and similarity assessment for robust heart sound segmentation

Heart Sound Segmentation plays a fundamental role in pathology detection in Phonocardiogram (PCG) signals. This matter of study has been widely studied in the past decades, however the majority of algorithms' results correspond only to small databases, composed by only quality signals or signals specific to one acquisition system. In this work we proposed a robust segmentation algorithm integrated with clinical information, based on a pattern recognition approach for segmentation of the fundamental heart sounds, which is validated in several databases from different countries and with different acquisition instrumentations. The database comprises a total of 3153 recordings from 764 patients with a variety of pathological conditions. The general results were 95% and 96% of sensitivity and positive predictivity, respectively. Based on the results the algorithm is able to perform with accuracy maintaining generalization capabilities.

Pattern discovery and similarity assessment for robust Heart Sound Segmentation

Heart Sound Segmentation plays a fundamental role in pathology detection in Phonocardiogram (PCG) signals. This matter of study has been widely studied in the past decades, however the majority of algorithms' results correspond only to small databases, composed by only quality signals or signals specific to one acquisition system. In this work we proposed a robust segmentation algorithm integrated with clinical information, based on a pattern recognition approach for segmentation of the fundamental heart sounds, which is validated in several databases from different countries and with different acquisition instrumentations. The database comprises a total of 3153 recordings from 764 patients with a variety of pathological conditions. The general results were 95% and 96% of sensitivity and positive predictivity, respectively. Based on the results the algorithm is able to perform with accuracy maintaining generalization capabilities.

Towards large-scale in free-standing graphene and N-graphene sheets

One of the greatest challenges in the commercialization of graphene and derivatives is production of high quality material in bulk quantities at low price and in a reproducible manner. The very limited control, or even lack of, over the synthesis process is one of the main problems of conventional approaches. Herein, we present a microwave plasma-enabled scalable route for continuous, large-scale fabrication of free-standing graphene and nitrogen doped graphene sheets. The method's crucial advantage relies on harnessing unique plasma mechanisms to control the material and energy fluxes of the main building units at the atomic scale. By tailoring the high energy density plasma environment and complementarily applying in situ IR and soft UV radiation, a controllable selective synthesis of high quality graphene sheets at 2 mg/min yield with prescribed structural qualities was achieved. Raman spectroscopy, scanning electron microscopy, high resolution transmission electron microscopy, X-ray photoelectron spectroscopy and Near Edge X-ray-absorption fine-structure spectroscopy were used to probe the morphological, chemical and microstructural features of the produced material. The method described here is scalable and show a potential for controllable, large-scale fabrication of other graphene derivatives and promotes microwave plasmas as a competitive, green, and cost-effective alternative to presently used chemical methods.

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.

An interpretable data-driven approach for rules construction: Application to cardiovascular risk assessment

The development of models able to produce an understandable decision by the clinicians is of great importance to support their decision. Therefore, the research of methodologies able to extract useful knowledge from existing datasets, as well as to integrate this knowledge into the current clinical evidence, is a key aspect in the enhancement of the clinical decision. This work focuses on the development of interpretable models to assess the patient's condition based on supervised clustering theories, enabling the discovery of a set of features that best represents that condition. At the same time, the technique is supported on a structure that enables the formulation of simple and interpretable rules. Despite its general applicability, the proposed methodology is applied to coronary artery disease (CAD), particularly, in the risk of death assessment (30 days after the admission) of patients that have been admitted to the emergency unit. The validation is performed using a real dataset with Acute Coronary Syndromes, provided by the Portuguese Society of Cardiology. While the methodology produces simple and interpretable rules, the performance achieves an improvement of 7% in relation to geometric mean, when compared with GRACE model (commonly used in Portugal).

A non-exercise based V02max prediction using FRIEND dataset with a neural network

The main goal of this work is the development of models, based on computational intelligence techniques, in particular neural networks, to predict the maximum oxygen consumption value. While the maximum oxygen consumption is a direct mark of the cardiorespiratory fitness, several studies have also confirmed it also as a powerful predictor of risk for adverse outcomes, such as hypertension, obesity, and diabetes. Therefore, the existence of simpler and accurate models, establishing an alternative to standard cardiopulmonary exercise tests, with the potential to be employed in the stratification of the general population in daily clinical practice, would be of major importance. In the current study, different models were implemented and compared: 1) the traditional Wasserman/Hansen equation; 2) linear regression and; 3) non-linear neural networks. Their performance was evaluated based on the "FRIEND - Fitness Registry and the Importance of Exercise: The National Data Base" [1] being, in the present study, a subset of 12262 individuals employed. The accuracy of the models was performed through the computation of sensitivity and specificity values. The results show the superiority of neural networks in the prediction of maximum oxygen consumption.

Can PPG be used for HRV analysis?

Heart rate variability (HRV) represents one of the most promising markers of the autonomic nervous system (ANS) regulation. However, it requires the acquisition of the ECG signal in order to reliably detect the RR intervals, which is not always easily and comfortably available in personal health applications. Additionally, due to progress in single spot optical sensors, photoplethysmography (PPG) is an interesting alternative for heartbeat interval measurements, since it is a more convenient and a less intrusive measurement technique. Driven by the technological advances in such sensors, wrist-worn devices are becoming a commodity, and the interest in the assessment of HRV indexes from the PPG analysis (pulse rate variability - PRV) is rising. In this study, we investigate the hypothesis of using PRV features as surrogates for HRV indexes, in three different contexts: healthy subjects at rest, healthy subjects after physical exercise and subjects with cardiovascular diseases (CVD). Additionally, we also evaluate which are the characteristic points better suited for PRV analysis in these contexts, i.e. the PPG waveform characteristic points leading to the PRV features that present the best estimates of HRV (correlation and error analysis). The achieved results suggest that the PRV can be often used as an alternative for HRV analysis in healthy subjects, with significant correlations above 82%, for both time and frequency features. Contrarily, in the post-exercise and CVD subjects, time and (most importantly) frequency domain features shall be used with caution (mean correlations ranging from 68% to 88%).

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.

The use of low-dose metronomic chemotherapy in dogs-insight into a modern cancer field

The era of chemotherapy, which started in the middle of the last century, has been ruled by the routine use of dose-intense protocols, based on the "maximum-tolerated dose" concept. By promoting a balance between patient's quality of life and the goal of rapidly killing as many tumour cells as possible, these protocols still play a prominent role in veterinary oncology. However, with the opening of a new millennium, metronomic chemotherapy (MC) started to be considered a possible alternative to traditional dose-intense chemotherapy. Characterized by a long-term daily administration of lower doses of cytotoxic drugs, this new modality stands out for its unique combination of effects, namely on neovascularization, immune response and tumour dormancy. This article reviews the rationale for treatment with MC, its mechanism of action and the main studies conducted in veterinary medicine, and discusses the key challenges yet to be solved.

Protocol Design Challenges in the Detection of Awareness in Aware Subjects Using EEG Signals

Recent studies have evidenced serious difficulties in detecting covert awareness with electroencephalography-based techniques both in unresponsive patients and in healthy control subjects. This work reproduces the protocol design in two recent mental imagery studies with a larger group comprising 20 healthy volunteers. The main goal is assessing if modifications in the signal extraction techniques, training-testing/cross-validation routines, and hypotheses evoked in the statistical analysis, can provide solutions to the serious difficulties documented in the literature. The lack of robustness in the results advises for further search of alternative protocols more suitable for machine learning classification and of better performing signal treatment techniques. Specific recommendations are made using the findings in this work.

Evaluation of Pax5 expression and comparison with BLA.36 and CD79αcy in feline non-Hodgkin lymphoma

Paired box gene 5 (Pax5) is a widely used B-cell marker for human and canine non-Hodgkin's lymphoma (nHL); however, in the literature there is only one case report using Pax5 in a cat B-cell lymphoma. The purposes of this study were to investigate the expression and detection of B-cell specific activator protein (BSAP) using a monoclonal anti-Pax5 antibody in feline nHL (FnHL) tissue samples to evaluate its diagnostic relevance as a B-cell marker. A total of 45 FnHL samples in 45 cats were evaluated. B-cell lymphoma was the most common immunophenotype (51.1%) for all the samples and T-cell the most common immunophenotype (64.3%) for the gastrointestinal (GI) form. Pax5 stained 82.6% of all B-cell lymphomas and no expression was found in any of the T-cell lymphomas. Anti-Pax5 antibody staining in FnHL is similar to that reported in human and canine counterparts and may offer an excellent B-cell marker in cats.

Improving clinical models based on knowledge extracted from current datasets: a new approach

The Cardiovascular Diseases (CVD) are the leading cause of death in the world, being prevention recognized to be a key intervention able to contradict this reality. In this context, although there are several models and scores currently used in clinical practice to assess the risk of a new cardiovascular event, they present some limitations. The goal of this paper is to improve the CVD risk prediction taking into account the current models as well as information extracted from real and recent datasets. This approach is based on a decision tree scheme in order to assure the clinical interpretability of the model. An innovative optimization strategy is developed in order to adjust the decision tree thresholds (rule structure is fixed) based on recent clinical datasets. A real dataset collected in the ambit of the National Registry on Acute Coronary Syndromes, Portuguese Society of Cardiology is applied to validate this work. In order to assess the performance of the new approach, the metrics sensitivity, specificity and accuracy are used. This new approach achieves sensitivity, a specificity and an accuracy values of, 80.52%, 74.19% and 77.27% respectively, which represents an improvement of about 26% in relation to the accuracy of the original score.

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.

Robust features for detection of crackles: an exploratory study

Crackles are adventitious and explosive respiratory sounds that can be classified as fine or coarse. These sounds are usually associated with cardiopulmonary diseases such as the chronic obstructive pulmonary disease. In this work seven different features were tested with the objective to identify the best subset of features that allows a robust detection of coarse crackles. Some of the features used in this study are new, namely those based on the local entropy, on the Teager energy and on the residual fit of a Generalized Autoregressive Conditional Heteroskedasticity process. The best features as a function of the number of features used in classification were identified having into account the Matthews correlation coefficient. The best individual feature was based on the local entropy. A significant improvement in the performance was obtained by using the feature based on local entropy and the feature based on the wavelet packed stationary transform - no stationary transform. The addition of more features only allows a smaller improvement.

Minimally invasive repair of Morgagni hernia - A multicenter case series

Children may benefit from minimally invasive surgery (MIS) in the correction of Morgagni hernia (MH). The present study aims to evaluate the outcome of MIS through a multicenter study. National institutions that use MIS in the treatment of MH were included. Demographic, clinical and operative data were analyzed. Thirteen patients with MH (6 males) were operated using similar MIS technique (percutaneous stitches) at a mean age of 22.2±18.3 months. Six patients had chromosomopathies (46%), five with Down syndrome (39%). Respiratory complaints were the most common presentation (54%). Surgery lasted 95±23min. In none of the patients was the hernia sac removed; prosthesis was never used. In the immediate post-operative period, 4 patients (36%) were admitted to intensive care unit (all with Down syndrome); all patients started enteral feeds within the first 24h. With a mean follow-up of 56±16.6 months, there were two recurrences (18%) at the same institution, one of which was repaired with an absorbable suture; both with Down syndrome. The application of MIS in the MH repair is effective even in the presence of comorbidities such as Down syndrome; the latter influences the immediate postoperative recovery and possibly the recurrence rate. Removal of hernia sac does not seem necessary. Non-absorbable sutures may be more appropriate.

Can behavioural differences in Platypus cylindrus (Coleoptera: Platypodinae) from Portugal and Tunisia be explained by genetic and morphological traits?

Platypus cylindrus is an important wood borer of cork oak trees (Quercus suber) in the Mediterranean region, namely Portugal, Morocco and Algeria where its presence has drastically increased in the past few decades. On the contrary, the insect is not a relevant pest in Tunisia. The aim of this work is to analyze morphological and genetic differences among Tunisian and Portuguese populations in order to understand their role in the diverse population dynamics (e.g., aggressiveness) of the insect. The information could be used as a novel tool to implement protective measures. Insects were collected from cork oak stands in Tunisia (Ain Beya, Babouch and Mzara) and Portugal (Chamusca and Crato). Morphological traits of female and male mycangial pits were determined, using scanning electron microscopy but no significant differences were found. Genetic differences were analyzed using nuclear (internal simple sequence repeat polymerase chain reaction) and mitochondrial (cytochrome oxidase I (COI)) molecular markers. The results showed a very low level of intraspecific polymorphism and genetic diversity. The alignment of COI sequences showed high percentage of identical sites (99%) indicating a very low variation in nucleotide composition. Other variables related with the ecology of the insect and its associated fungi must be studied for a better understanding of the differences in the insect population's dynamic in Mediterranean countries.

A retrospective analysis of chemotherapy switch suggests improved outcome in surgically removed, biologically aggressive canine haemangiosarcoma

Haemangiosarcoma (HSA) has an aggressive biological behaviour and carries a poor prognosis, with less than 10% of treated dogs surviving longer than 1 year. In this retrospective study a varied metronomic chemotherapy (MC) regimen preceded by adjuvant doxorubicin-based maximum-tolerated dose chemotherapy (MTDC) was compared with MTDC, in terms of efficacy [time to metastasis, (TTM) and survival time (ST)] and safety in dogs with biologically aggressive HSA. Dogs were eligible if they had no metastasis after MTDC and received either no further chemotherapy or MC maintenance. Twelve dogs received MTDC, and 10 received MC thereafter. Median TTM and ST were significantly longer for dogs receiving MTDC-MC (not reached versus 150 days, P = 0.028; and not reached versus 168 days, P = 0.030, respectively). Treatment was well tolerated. MTDC followed by MC is safe and suggests improved TTM and ST in dogs with surgically removed, biologically aggressive HSA that are treated in the microscopic setting.