Back to the future: the novel art of digital auscultation applied in a prospective observational study of critically ill Covid-19 patients.

BACKGROUND: The Covid-19 pandemic has caused immense pressure on Intensive Care Units (ICU). In patients with severe ARDS due to Covid-19, respiratory mechanics are important for determining the severity of lung damage. Lung auscultation could not be used during the pandemic despite its merit. The main objective of this study was to investigate associations between lung auscultatory sound features and lung mechanical properties, length of stay (LOS) and survival, in adults with severe Covid-19 ARDS. METHODS: Consecutive patients admitted to a large ICU between 2020 and 2021 (n = 173) were included. Digital stethoscopes obtained auscultatory sounds and stored them in an on-line database for replay and further processing using advanced AI techniques. Correlation and regression analysis explored relationships between digital auscultation findings and lung mechanics or the ICU outcome. The resulting annotated lung sounds database is also publicly available as supplementary material. RESULTS: The presence of squawks was associated with the ICU LOS, outcome and 90-day mortality. Other features (age, SOFA score & oxygenation index upon admission, minimum crackle entropy) had significant impact on outcome. Additional features affecting the 90-d survival were age and mean crackle entropy. Multivariate logistic regression showed that survival was affected by age, baseline SOFA, baseline oxygenation index and minimum crackle entropy. CONCLUSIONS: Respiratory mechanics were associated with various adventitious sounds, whereas the lung sound analytics and the presence of certain adventitious sounds correlated with the ICU outcome and the 90-d survival. Spectral features of crackles sounds can serve as prognostic factors for survival, highlighting the importance of digital auscultation.

A Comparison Study of Deep Learning Methodologies for Music Emotion Recognition.

Classical machine learning techniques have dominated Music Emotion Recognition. However, improvements have slowed down due to the complex and time-consuming task of handcrafting new emotionally relevant audio features. Deep learning methods have recently gained popularity in the field because of their ability to automatically learn relevant features from spectral representations of songs, eliminating such necessity. Nonetheless, there are limitations, such as the need for large amounts of quality labeled data, a common problem in MER research. To understand the effectiveness of these techniques, a comparison study using various classical machine learning and deep learning methods was conducted. The results showed that using an ensemble of a Dense Neural Network and a Convolutional Neural Network architecture resulted in a state-of-the-art 80.20% F1 score, an improvement of around 5% considering the best baseline results, concluding that future research should take advantage of both paradigms, that is, combining handcrafted features with feature learning.

BRACETS: Bimodal repository of auscultation coupled with electrical impedance thoracic signals.

BACKGROUND AND OBJECTIVE: Respiratory diseases are among the most significant causes of morbidity and mortality worldwide, causing substantial strain on society and health systems. Over the last few decades, there has been increasing interest in the automatic analysis of respiratory sounds and electrical impedance tomography (EIT). Nevertheless, no publicly available databases with both respiratory sound and EIT data are available. METHODS: In this work, we have assembled the first open-access bimodal database focusing on the differential diagnosis of respiratory diseases (BRACETS: Bimodal Repository of Auscultation Coupled with Electrical Impedance Thoracic Signals). It includes simultaneous recordings of single and multi-channel respiratory sounds and EIT. Furthermore, we have proposed several machine learning-based baseline systems for automatically classifying respiratory diseases in six distinct evaluation tasks using respiratory sound and EIT (A1, A2, A3, B1, B2, B3). These tasks included classifying respiratory diseases at sample and subject levels. The performance of the classification models was evaluated using a 5-fold cross-validation scheme (with subject isolation between folds). RESULTS: The resulting database consists of 1097 respiratory sounds and 795 EIT recordings acquired from 78 adult subjects in two countries (Portugal and Greece). In the task of automatically classifying respiratory diseases, the baseline classification models have achieved the following average balanced accuracy: Task A1 - 77.9±13.1%; Task A2 - 51.6±9.7%; Task A3 - 38.6±13.1%; Task B1 - 90.0±22.4%; Task B2 - 61.4±11.8%; Task B3 - 50.8±10.6%. CONCLUSION: The creation of this database and its public release will aid the research community in developing automated methodologies to assess and monitor respiratory function, and it might serve as a benchmark in the field of digital medicine for managing respiratory diseases. Moreover, it could pave the way for creating multi-modal robust approaches for that same purpose.

Automatic wheeze segmentation using harmonic-percussive source separation and empirical mode decomposition.

Wheezes are adventitious respiratory sounds commonly present in patients with respiratory conditions. The presence of wheezes and their time location are relevant for clinical reasons, such as understanding the degree of bronchial obstruction. Conventional auscultation is usually employed to analyze wheezes, but remote monitoring has become a pressing need during recent years. Automatic respiratory sound analysis is required to reliably perform remote auscultation. In this work we propose a method for wheeze segmentation. Our method starts by decomposing a given audio excerpt into intrinsic mode frequencies using empirical mode decomposition. Then, we apply harmonic-percussive source separation to the resulting audio tracks and get harmonic-enhanced spectrograms, which are processed to obtain harmonic masks. Subsequently, a series of empirically derived rules are applied to find wheeze candidates. Finally, the candidates stemming from the different audio tracks are merged and median filtered. In the evaluation stage, we compare our method to three baselines on the ICBHI 2017 Respiratory Sound Database, a challenging dataset containing various noise sources and background sounds. Using the full dataset, our method outperforms the baselines, achieving an F1 of 41.9%. Our method's performance is also better than the baselines across several stratified results focusing on five variables: recording equipment, age, sex, body-mass index, and diagnosis. We conclude that, contrary to what has been reported in the literature, wheeze segmentation has not been solved for real life scenario applications. Adaptation of existing systems to demographic characteristics might be a promising step in the direction of algorithm personalization, which would make automatic wheeze segmentation methods clinically viable.

Automated Lung Sound Classification Using a Hybrid CNN-LSTM Network and Focal Loss Function.

Respiratory diseases constitute one of the leading causes of death worldwide and directly affect the patient's quality of life. Early diagnosis and patient monitoring, which conventionally include lung auscultation, are essential for the efficient management of respiratory diseases. Manual lung sound interpretation is a subjective and time-consuming process that requires high medical expertise. The capabilities that deep learning offers could be exploited in order that robust lung sound classification models can be designed. In this paper, we propose a novel hybrid neural model that implements the focal loss (FL) function to deal with training data imbalance. Features initially extracted from short-time Fourier transform (STFT) spectrograms via a convolutional neural network (CNN) are given as input to a long short-term memory (LSTM) network that memorizes the temporal dependencies between data and classifies four types of lung sounds, including normal, crackles, wheezes, and both crackles and wheezes. The model was trained and tested on the ICBHI 2017 Respiratory Sound Database and achieved state-of-the-art results using three different data splitting strategies-namely, sensitivity 47.37%, specificity 82.46%, score 64.92% and accuracy 73.69% for the official 60/40 split, sensitivity 52.78%, specificity 84.26%, score 68.52% and accuracy 76.39% using interpatient 10-fold cross validation, and sensitivity 60.29% and accuracy 74.57% using leave-one-out cross validation.

Melanoma classification using light-Fields with morlet scattering transform and CNN: Surface depth as a valuable tool to increase detection rate.

Medical image classification through learning-based approaches has been increasingly used, namely in the discrimination of melanoma. However, for skin lesion classification in general, such methods commonly rely on dermoscopic or other 2D-macro RGB images. This work proposes to exploit beyond conventional 2D image characteristics, by considering a third dimension (depth) that characterises the skin surface rugosity, which can be obtained from light-field images, such as those available in the SKINL2 dataset. To achieve this goal, a processing pipeline was deployed using a morlet scattering transform and a CNN model, allowing to perform a comparison between using 2D information, only 3D information, or both. Results show that discrimination between Melanoma and Nevus reaches an accuracy of 84.00, 74.00 or 94.00% when using only 2D, only 3D, or both, respectively. An increase of 14.29pp in sensitivity and 8.33pp in specificity is achieved when expanding beyond conventional 2D information by also using depth. When discriminating between Melanoma and all other types of lesions (a further imbalanced setting), an increase of 28.57pp in sensitivity and decrease of 1.19pp in specificity is achieved for the same test conditions. Overall the results of this work demonstrate significant improvements over conventional approaches.

Classification of Electrical Impedance Tomography Data Using Machine Learning.

Patients suffering from pulmonary diseases typically exhibit pathological lung ventilation in terms of homogeneity. Electrical Impedance Tomography (EIT) is a non- invasive imaging method that allows to analyze and quantify the distribution of ventilation in the lungs. In this article, we present a new approach to promote the use of EIT data and the implementation of new clinical applications for differential diagnosis, with the development of several machine learning models to discriminate between EIT data from healthy and nonhealthy subjects. EIT data from 16 subjects were acquired: 5 healthy and 11 non-healthy subjects (with multiple pulmonary conditions). Preliminary results have shown accuracy percentages of 66% in challenging evaluation scenarios. The results suggest that the pairing of EIT feature engineering methods with machine learning methods could be further explored and applied in the diagnostic and monitoring of patients suffering from lung diseases. Also, we introduce the use of a new feature in the context of EIT data analysis (Impedance Curve Correlation).

Detection of squawks in respiratory sounds of mechanically ventilated COVID-19 patients.

Mechanically ventilated patients typically exhibit abnormal respiratory sounds. Squawks are short inspiratory adventitious sounds that may occur in patients with pneumonia, such as COVID-19 patients. In this work we devised a method for squawk detection in mechanically ventilated patients by developing algorithms for respiratory cycle estimation, squawk candidate identification, feature extraction, and clustering. The best classifier reached an F1 of 0.48 at the sound file level and an F1 of 0.66 at the recording session level. These preliminary results are promising, as they were obtained in noisy environments. This method will give health professionals a new feature to assess the potential deterioration of critically ill patients.

Skin lesion classification using features of 3D border lines.

Machine learning algorithms are progressively assuming important roles as computational tools to support clinical diagnosis, namely in the classification of pigmented skin lesions using RGB images. Most current classification methods rely on common 2D image features derived from shape, colour or texture, which does not always guarantee the best results. This work presents a contribution to this field, by exploiting the lesions' border line characteristics using a new dimension - depth, which has not been thoroughly investigated so far. A selected group of features is extracted from the depth information of 3D images, which are then used for classification using a quadratic Support Vector Machine. Despite class imbalance often present in medical image datasets, the proposed algorithm achieves a top geometric mean of 94.87%, comprising 100.00% sensitivity and 90.00% specificity, using only depth information for the detection of Melanomas. Such results show that potential gains can be achieved by extracting information from this often overlooked dimension, which provides more balanced results in terms of sensitivity and specificity than other settings.

Automatic Classification of Adventitious Respiratory Sounds: A (Un)Solved Problem?

(1) Background: Patients with respiratory conditions typically exhibit adventitious respiratory sounds (ARS), such as wheezes and crackles. ARS events have variable duration. In this work we studied the influence of event duration on automatic ARS classification, namely, how the creation of the Other class (negative class) affected the classifiers' performance. (2) Methods: We conducted a set of experiments where we varied the durations of the other events on three tasks: crackle vs. wheeze vs. other (3 Class); crackle vs. other (2 Class Crackles); and wheeze vs. other (2 Class Wheezes). Four classifiers (linear discriminant analysis, support vector machines, boosted trees, and convolutional neural networks) were evaluated on those tasks using an open access respiratory sound database. (3) Results: While on the 3 Class task with fixed durations, the best classifier achieved an accuracy of 96.9%, the same classifier reached an accuracy of 81.8% on the more realistic 3 Class task with variable durations. (4) Conclusion: These results demonstrate the importance of experimental design on the assessment of the performance of automatic ARS classification algorithms. Furthermore, they also indicate, unlike what is stated in the literature, that the automatic classification of ARS is not a solved problem, as the algorithms' performance decreases substantially under complex evaluation scenarios.

A Wearable Stethoscope for Long-Term Ambulatory Respiratory Health Monitoring.

Lung sounds acquired by stethoscopes are extensively used in diagnosing and differentiating respiratory diseases. Although an extensive know-how has been built to interpret these sounds and identify diseases associated with certain patterns, its effective use is limited to individual experience of practitioners. This user-dependency manifests itself as a factor impeding the digital transformation of this valuable diagnostic tool, which can improve patient outcomes by continuous long-term respiratory monitoring under real-life conditions. Particularly patients suffering from respiratory diseases with progressive nature, such as chronic obstructive pulmonary diseases, are expected to benefit from long-term monitoring. Recently, the COVID-19 pandemic has also shown the lack of respiratory monitoring systems which are ready to deploy in operational conditions while requiring minimal patient education. To address particularly the latter subject, in this article, we present a sound acquisition module which can be integrated into a dedicated garment; thus, minimizing the role of the patient for positioning the stethoscope and applying the appropriate pressure. We have implemented a diaphragm-less acousto-electric transducer by stacking a silicone rubber and a piezoelectric film to capture thoracic sounds with minimum attenuation. Furthermore, we benchmarked our device with an electronic stethoscope widely used in clinical practice to quantify its performance.

WELCOME ­ innovative integrated care platform using wearable sensing and smart cloud computing for COPD patients with comorbidities.

We propose WELCOME, an innovative integrated care platform using wearable sensors and smart cloud computing for Chronic Obstructive Pulmonary Disease (COPD) patients with co-morbidities. WELCOME aims to bring about a change in the reactive nature of the management of chronic diseases and its comorbidities, in particular through the development of a patient centred and proactive approach to COPD management. The aim of WELCOME is to support healthcare services to give early detection of complications (potentially reducing hospitalisations) and the prevention and mitigation of comorbidities (Heart Failure, Diabetes, Anxiety and Depression). The system incorporates patient hub, where it interacts with the patient via a light vest including a large number of non-invasive chest sensors for monitoring various relevant parameters. In addition, interactive applications to monitor and manage diabetes, anxiety and lifestyle issues will be provided to the patient. Informal carers will also be supported in dealing with their patients. On the other hand, welcome smart cloud platform is the heart of the proposed system where all the medical records and the monitoring data are managed and processed via the decision support system. Healthcare professionals will be able to securely access the WELCOME applications to monitor and manage the patient's conditions and respond to alerts on personalized level.