Artificial intelligence for quality control of oscillometry measures

BACKGROUND

The forced oscillation technique (FOT) allows non-invasive lung function testing during quiet breathing even without expert guidance. However, it still relies on an operator for excluding breaths with artefacts such as swallowing, glottis closure and coughing. This manual selection is operator-dependent and time-consuming. We evaluated supervised machine learning methods to exclude breaths with artefacts from data analysis automatically.

METHODS

We collected 932 FOT measurements (Resmon Pro Full, Restech) from 155 patients (6-87 years) following the European Respiratory Society (ERS) technical standards. Patients were randomly assigned to either a training (70%) or test set. For each breath, we computed 71 features (including anthropometric, pressure stimulus, breathing pattern, and oscillometry data). Univariate filter, multivariate filter and wrapper methods for feature selection combined with several classification models were considered.

RESULTS

Trained operators identified 4333 breaths with- and 10244 without artefacts. Features selection performed by a wrapper method combined with an AdaBoost tree model provided the best performance metrics on the test set: Balanced Accuracy = 85%; Sensitivity = 79%; Specificity = 91%; AUC-ROC = 0.93. Differences in FOT parameters computed after manual or automatic breath selection was less than ∼0.25 cmH₂O*s/L for 95% of cases.

CONCLUSION

Supervised machine-learning techniques allow reliable artefact detection in FOT diagnostic tests. Automating this process is fundamental for enabling FOT for home monitoring, telemedicine, and point-of-care diagnostic applications and opens new scenarios for respiratory and community medicine.

Machine learning approach in mortality rate prediction for hemodialysis patients

Kernel support vector machine algorithm and K-means clustering algorithm are used to determine the expected mortality rate for hemodialysis patients. The national nephrology database of Montenegro has been used to conduct this research. Mortality rate prediction is realized with accuracy up to 94.12% and up to 96.77%, when a complete database is observed and when a reduced database (that contains data for the three most common basic diseases) is observed, respectively. Additionally, it is shown that just a few parameters, most of which are collected during the sole patient examination, are enough for satisfying results.

Technical standards for respiratory oscillometry

Oscillometry (also known as the forced oscillation technique) measures the mechanical properties of the respiratory system (upper and intrathoracic airways, lung tissue and chest wall) during quiet tidal breathing, by the application of an oscillating pressure signal (input or forcing signal), most commonly at the mouth. With increased clinical and research use, it is critical that all technical details of the hardware design, signal processing and analyses, and testing protocols are transparent and clearly reported to allow standardisation, comparison and replication of clinical and research studies. Because of this need, an update of the 2003 European Respiratory Society (ERS) technical standards document was produced by an ERS task force of experts who are active in clinical oscillometry research.The aim of the task force was to provide technical recommendations regarding oscillometry measurement including hardware, software, testing protocols and quality control.The main changes in this update, compared with the 2003 ERS task force document are 1) new quality control procedures which reflect use of "within-breath" analysis, and methods of handling artefacts; 2) recommendation to disclose signal processing, quality control, artefact handling and breathing protocols (e.g. number and duration of acquisitions) in reports and publications to allow comparability and replication between devices and laboratories; 3) a summary review of new data to support threshold values for bronchodilator and bronchial challenge tests; and 4) updated list of predicted impedance values in adults and children.

Groundtruth: A Matlab GUI for Artifact and Feature Identification in Physiological Signals

Groundtruth is a Matlab Graphical User Interface (GUI) developed for the identification of key features and artifacts within physiological signals. The ultimate aim of this GUI is to provide a simple means of assessing the performance of new sensors. Secondary, to this is providing a means of providing marked data, enabling assessment of automated artifact rejection and feature identification algorithms. With the emergence of new wearable sensor technologies, there is an unmet need for convenient assessment of device performance, and a faster means of assessing new algorithms. The proposed GUI allows interactive marking of artifact regions as well as simultaneous interactive identification of key features, e.g., respiration peaks in respiration signals, R-peaks in Electrocardiography signals, etc. In this paper, we present the base structure of the system, together with an example of its use for two simultaneously worn respiration sensors. The respiration rates are computed for both original as well as artifact removed data and validated using Bland–Altman plots. The respiration rates computed based on the proposed GUI (after artifact removal process) demonstrated consistent results for two respiration sensors after artifact removal process. Groundtruth is customizable, and alternative processing modules are easy to add/remove. Groundtruth is intended for open-source use.

Wearable sensors for Parkinson's disease: which data are worth collecting for training symptom detection models

Machine learning algorithms that use data streams captured from soft wearable sensors have the potential to automatically detect PD symptoms and inform clinicians about the progression of disease. However, these algorithms must be trained with annotated data from clinical experts who can recognize symptoms, and collecting such data are costly. Understanding how many sensors and how much labeled data are required is key to successfully deploying these models outside of the clinic. Here we recorded movement data using 6 flexible wearable sensors in 20 individuals with PD over the course of multiple clinical assessments conducted on 1 day and repeated 2 weeks later. Participants performed 13 common tasks, such as walking or typing, and a clinician rated the severity of symptoms (bradykinesia and tremor). We then trained convolutional neural networks and statistical ensembles to detect whether a segment of movement showed signs of bradykinesia or tremor based on data from tasks performed by other individuals. Our results show that a single wearable sensor on the back of the hand is sufficient for detecting bradykinesia and tremor in the upper extremities, whereas using sensors on both sides does not improve performance. Increasing the amount of training data by adding other individuals can lead to improved performance, but repeating assessments with the same individuals—even at different medication states—does not substantially improve detection across days. Our results suggest that PD symptoms can be detected during a variety of activities and are best modeled by a dataset incorporating many individuals.