Stratifying asthma severity in children using cough sound analytic technology

Feature fusion method for pulmonary tuberculosis patient detection based on cough sound

Since the COVID-19, cough sounds have been widely used for screening purposes. Intelligent analysis techniques have proven to be effective in detecting respiratory diseases. In 2021, there were up to 10 million TB-infected patients worldwide, with an annual growth rate of 4.5%. Most of the patients were from economically underdeveloped regions and countries. The PPD test, a common screening method in the community, has a sensitivity of as low as 77%. Although IGRA and Xpert MTB/RIF offer high specificity and sensitivity, their cost makes them less accessible. In this study, we proposed a feature fusion model-based cough sound classification method for primary TB screening in communities. Data were collected from hospitals using smart phones, including 230 cough sounds from 70 patients with TB and 226 cough sounds from 74 healthy subjects. We employed Bi-LSTM and Bi-GRU recurrent neural networks to analyze five traditional feature sets including the Mel frequency cepstrum coefficient (MFCC), zero-crossing rate (ZCR), short-time energy, root mean square, and chroma_cens. The incorporation of features extracted from the speech spectrogram by 2D convolution training into the Bi-LSTM model enhanced the classification results. With traditional futures, the best TB patient detection result was achieved with the Bi-LSTM model, with 93.99% accuracy, 93.93% specificity, and 92.39% sensitivity. When combined with a speech spectrogram, the classification results showed 96.33% accuracy, 94.99% specificity, and 98.13% sensitivity. Our findings underscore that traditional features and deep features have good complementarity when fused using Bi LSTM modelling, which outperforms existing PPD detection methods in terms of both efficiency and accuracy.

An AI-Enabled Bias-Free Respiratory Disease Diagnosis Model Using Cough Audio

Cough-based diagnosis for respiratory diseases (RDs) using artificial intelligence (AI) has attracted considerable attention, yet many existing studies overlook confounding variables in their predictive models. These variables can distort the relationship between cough recordings (input data) and RD status (output variable), leading to biased associations and unrealistic model performance. To address this gap, we propose the Bias-Free Network (RBF-Net), an end-to-end solution that effectively mitigates the impact of confounders in the training data distribution. RBF-Net ensures accurate and unbiased RD diagnosis features, emphasizing its relevance by incorporating a COVID-19 dataset in this study. This approach aims to enhance the reliability of AI-based RD diagnosis models by navigating the challenges posed by confounding variables. A hybrid of a Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks is proposed for the feature encoder module of RBF-Net. An additional bias predictor is incorporated in the classification scheme to formulate a conditional Generative Adversarial Network (c-GAN) that helps in decorrelating the impact of confounding variables from RD prediction. The merit of RBF-Net is demonstrated by comparing classification performance with a State-of-The-Art (SoTA) Deep Learning (DL) model (CNN-LSTM) after training on different unbalanced COVID-19 data sets, created by using a large-scale proprietary cough data set. RBF-Net proved its robustness against extremely biased training scenarios by achieving test set accuracies of 84.1%, 84.6%, and 80.5% for the following confounding variables-gender, age, and smoking status, respectively. RBF-Net outperforms the CNN-LSTM model test set accuracies by 5.5%, 7.7%, and 8.2%, respectively.

Real-time counting of wheezing events from lung sounds using deep learning algorithms: Implications for disease prediction and early intervention

This pioneering study aims to revolutionize self-symptom management and telemedicine-based remote monitoring through the development of a real-time wheeze counting algorithm. Leveraging a novel approach that includes the detailed labeling of one breathing cycle into three types: break, normal, and wheeze, this study not only identifies abnormal sounds within each breath but also captures comprehensive data on their location, duration, and relationships within entire respiratory cycles, including atypical patterns. This innovative strategy is based on a combination of a one-dimensional convolutional neural network (1D-CNN) and a long short-term memory (LSTM) network model, enabling real-time analysis of respiratory sounds. Notably, it stands out for its capacity to handle continuous data, distinguishing it from conventional lung sound classification algorithms. The study utilizes a substantial dataset consisting of 535 respiration cycles from diverse sources, including the Child Sim Lung Sound Simulator, the EMTprep Open-Source Database, Clinical Patient Records, and the ICBHI 2017 Challenge Database. Achieving a classification accuracy of 90%, the exceptional result metrics encompass the identification of each breath cycle and simultaneous detection of the abnormal sound, enabling the real-time wheeze counting of all respirations. This innovative wheeze counter holds the promise of revolutionizing research on predicting lung diseases based on long-term breathing patterns and offers applicability in clinical and non-clinical settings for on-the-go detection and remote intervention of exacerbated respiratory symptoms.

A scoping review of asthma and machine learning

OBJECTIVE

The objective of this study was to determine the extent of machine learning (ML) application in asthma research and to identify research gaps while mapping the existing literature.

DATA SOURCES

We conducted a scoping review. PubMed, ProQuest, and Embase Scopus databases were searched with an end date of September 18, 2020.

STUDY SELECTION

DistillerSR was used for data management. Inclusion criteria were an asthma focus, human participants, ML techniques, and written in English. Exclusion criteria were abstract only, simulation-based, not human based, or were reviews or commentaries. Descriptive statistics were presented.

RESULTS

A total of 6,317 potential articles were found. After removing duplicates, and reviewing the titles and abstracts, 102 articles were included for the full text analysis. Asthma episode prediction (24.5%), asthma phenotype classification (16.7%), and genetic profiling of asthma (12.7%) were the top three study topics. Cohort (52.9%), cross-sectional (20.6%), and case-control studies (11.8%) were the study designs most frequently used. Regarding the ML techniques, 34.3% of the studies used more than one technique. Neural networks, clustering, and random forests were the most common ML techniques used where they were used in 20.6%, 18.6%, and 17.6% of studies, respectively. Very few studies considered location of residence (i.e. urban or rural status).

CONCLUSIONS

The use of ML in asthma studies has been increasing with most of this focused on the three major topics (>50%). Future research using ML could focus on gaps such as a broader range of study topics and focus on its use in additional populations (e.g. location of residence).

Supplemental data for this article is available online at http://dx.doi.org/ .

A smartphone-based algorithm comprising cough analysis and patient-reported symptoms identifies acute exacerbations of asthma: a prospective, double blind, diagnostic accuracy study

Objective: Early and accurate recognition of asthma exacerbations reduces the duration and risk of hospitalization. Current diagnostic methods depend upon patient recognition of symptoms, expert clinical examination, or measures of lung function. Here, we aimed to develop and test the accuracy of a smartphone-based diagnostic algorithm that analyses five cough events and five patient-reported features (age, fever, acute or productive cough and wheeze) to detect asthma exacerbations.Methods: We conducted a double-blind, prospective, diagnostic accuracy study comparing the algorithm with expert clinical opinion and formal lung function testing. Results: One hundred nineteen participants >12 years with a physician-diagnosed history of asthma were recruited from a hospital in Perth, Western Australia: 46 with clinically confirmed asthma exacerbations, 73 with controlled asthma. The groups were similar in median age (54yr versus 60yr, p=0.72) and sex (female 76% versus 70%, p=0.5). The algorithm's positive percent agreement (PPA) with the expert clinical diagnosis of asthma exacerbations was 89% [95% CI: 76%, 96%]. The negative percent agreement (NPA) was 84% [95% CI: 73%, 91%]. The algorithm's performance for asthma exacerbations diagnosis exceeded its performance as a detector of patient-reported wheeze (sensitivity, 63.7%). Patient-reported wheeze in isolation was an insensitive marker of asthma exacerbations (PPA=53.8%, NPA=49%). Conclusions: Our diagnostic algorithm accurately detected the presence of an asthma exacerbation as a point-of-care test without requiring clinical examination or lung function testing. This method could improve the accuracy of telehealth consultations and might be helpful in Asthma Action Plans and patient-initiated therapy.

Feasibility and clinical utility of ambulatory cough monitoring in an outpatient clinical setting: a real-world retrospective evaluation

RESEARCH QUESTION

Objective quantification of cough is rarely utilised outside of research settings and the role of cough frequency monitoring in clinical practice has not been established. This study examined the clinical utility of cough frequency monitoring in an outpatient clinical setting.

METHODS

The study involved a retrospective review of cough monitor data. Participants included 174 patients referred for treatment of cough and upper airway symptoms (103 chronic cough; 50 inducible laryngeal obstruction; 21 severe asthma) and 15 controls. Measures, taken prior to treatment, included 24-h ambulatory cough frequency using the Leicester Cough Monitor, the Leicester Cough Questionnaire and Laryngeal Hypersensitivity Questionnaire. Post-treatment data were available for 50 participants. Feasibility and clinical utility were also reported.

RESULTS

Analysis time per recording was up to 10 min. 75% of participants could use the monitors correctly, and most (93%) recordings were interpretable. The geometric mean cough frequency in patients was 10.1±2.9 (mean±sd) compared to 2.4±2.0 for healthy controls (p=0.003). There was no significant difference in cough frequency between clinical groups (p=0.080). Cough frequency decreased significantly following treatment (p<0.001). There was a moderate correlation between cough frequency and both cough quality of life and laryngeal hypersensitivity. Cough frequency monitoring was responsive to therapy and able to discriminate differences in cough frequency between diseases.

CONCLUSION

While ambulatory cough frequency monitoring remains a research tool, it provides useful clinical data that can assist in patient management. Logistical issues may preclude use in some clinical settings, and additional time needs to be allocated to the process.

Deep Neural Network-Based Respiratory Pathology Classification Using Cough Sounds

Intelligent systems are transforming the world, as well as our healthcare system. We propose a deep learning-based cough sound classification model that can distinguish between children with healthy versus pathological coughs such as asthma, upper respiratory tract infection (URTI), and lower respiratory tract infection (LRTI). To train a deep neural network model, we collected a new dataset of cough sounds, labelled with a clinician's diagnosis. The chosen model is a bidirectional long-short-term memory network (BiLSTM) based on Mel-Frequency Cepstral Coefficients (MFCCs) features. The resulting trained model when trained for classifying two classes of coughs-healthy or pathology (in general or belonging to a specific respiratory pathology)-reaches accuracy exceeding 84% when classifying the cough to the label provided by the physicians' diagnosis. To classify the subject's respiratory pathology condition, results of multiple cough epochs per subject were combined. The resulting prediction accuracy exceeds 91% for all three respiratory pathologies. However, when the model is trained to classify and discriminate among four classes of coughs, overall accuracy dropped: one class of pathological coughs is often misclassified as the other. However, if one considers the healthy cough classified as healthy and pathological cough classified to have some kind of pathology, then the overall accuracy of the four-class model is above 84%. A longitudinal study of MFCC feature space when comparing pathological and recovered coughs collected from the same subjects revealed the fact that pathological coughs, irrespective of the underlying conditions, occupy the same feature space making it harder to differentiate only using MFCC features.

Identifying acute exacerbations of chronic obstructive pulmonary disease using patient-reported symptoms and cough feature analysis

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are commonly encountered in the primary care setting, though the accurate and timely diagnosis is problematic. Using technology like that employed in speech recognition technology, we developed a smartphone-based algorithm for rapid and accurate diagnosis of AECOPD. The algorithm incorporates patient-reported features (age, fever, and new cough), audio data from five coughs and can be deployed by novice users. We compared the accuracy of the algorithm to expert clinical assessment. In patients with known COPD, the algorithm correctly identified the presence of AECOPD in 82.6% (95% CI: 72.9–89.9%) of subjects (n = 86). The absence of AECOPD was correctly identified in 91.0% (95% CI: 82.4–96.3%) of individuals (n = 78). The diagnostic agreement was maintained in milder cases of AECOPD (PPA: 79.2%, 95% CI: 68.0–87.8%), who typically comprise the cohort presenting to primary care. The algorithm may aid early identification of AECOPD and be incorporated in patient self-management plans.

Diagnosing community-acquired pneumonia via a smartphone-based algorithm: a prospective cohort study in primary and acute-care consultations

BACKGROUND

Community-acquired pneumonia (CAP) is an essential consideration in patients presenting to primary care with respiratory symptoms; however, accurate diagnosis is difficult when clinical and radiological examinations are not possible, such as during telehealth consultations.

AIM

To develop and test a smartphone-based algorithm for diagnosing CAP without need for clinical examination or radiological inputs.

DESIGN AND SETTING

A prospective cohort study using data from participants aged >12 years presenting with acute respiratory symptoms to a hospital in Western Australia.

METHOD

Five cough audio-segments were recorded and four patient-reported symptoms (fever, acute cough, productive cough, and age) were analysed by the smartphone-based algorithm to generate an immediate diagnostic output for CAP. Independent cohorts were recruited to train and test the accuracy of the algorithm. Diagnostic agreement was calculated against the confirmed discharge diagnosis of CAP by specialist physicians. Specialist radiologists reported medical imaging.

RESULTS

The smartphone-based algorithm had high percentage agreement (PA) with the clinical diagnosis of CAP in the total cohort (n = 322, positive PA [PPA] = 86.2%, negative PA [NPA] = 86.5%, area under the receiver operating characteristic curve [AUC] = 0.95); in participants 22-<65 years (n = 192, PPA = 85.7%, NPA = 87.0%, AUC = 0.94), and in participants aged ≥65 years (n = 86, PPA = 85.7%, NPA = 87.5%, AUC = 0.94). Agreement was preserved across CAP severity: 85.1% (n = 80/94) of participants with CRB-65 scores 1 or 2, and 87.7% (n = 57/65) with a score of 0, were correctly diagnosed by the algorithm.

CONCLUSION

The algorithm provides rapid and accurate diagnosis of CAP. It offers improved accuracy over current protocols when clinical evaluation is difficult. It provides increased capabilities for primary and acute care, including telehealth services, required during the COVID-19 pandemic.

The diagnosis of respiratory disease in children using a phone-based cough and symptom analysis algorithm: The smartphone recordings of cough sounds 2 (SMARTCOUGH-C 2) trial design

The diagnosis of acute respiratory diseases in children can be challenging, and no single objective diagnostic test exists for common pediatric respiratory diseases. Previous research has demonstrated that ResAppDx, a cough sound and symptom-based analysis algorithm, can identify common respiratory diseases at the point of care. We present the study protocol for SMARTCOUGH-C 2, a prospective diagnostic accuracy trial of a cough and symptom-based algorithm in a cohort of children presenting with acute respiratory diseases. The objective of the study is to assess the performance characteristics of the ResAppDx algorithm in the diagnosis of common pediatric acute respiratory diseases.

Diagnosing Chronic Obstructive Airway Disease on a Smartphone Using Patient-Reported Symptoms and Cough Analysis: Diagnostic Accuracy Study

Background

Rapid and accurate diagnosis of chronic obstructive pulmonary disease (COPD) is problematic in acute care settings, particularly in the presence of infective comorbidities.

Objective

The aim of this study was to develop a rapid smartphone-based algorithm for the detection of COPD in the presence or absence of acute respiratory infection and evaluate diagnostic accuracy on an independent validation set.

Methods

Participants aged 40 to 75 years with or without symptoms of respiratory disease who had no chronic respiratory condition apart from COPD, chronic bronchitis, or emphysema were recruited into the study. The algorithm analyzed 5 cough sounds and 4 patient-reported clinical symptoms, providing a diagnosis in less than 1 minute. Clinical diagnoses were determined by a specialist physician using all available case notes, including spirometry where available.

Results

The algorithm demonstrated high positive percent agreement (PPA) and negative percent agreement (NPA) with clinical diagnosis for COPD in the total cohort (N=252; PPA=93.8%, NPA=77.0%, area under the curve [AUC]=0.95), in participants with pneumonia or infective exacerbations of COPD (n=117; PPA=86.7%, NPA=80.5%, AUC=0.93), and in participants without an infective comorbidity (n=135; PPA=100.0%, NPA=74.0%, AUC=0.97). In those who had their COPD confirmed by spirometry (n=229), PPA was 100.0% and NPA was 77.0%, with an AUC of 0.97.

Conclusions

The algorithm demonstrated high agreement with clinical diagnosis and rapidly detected COPD in participants presenting with or without other infective lung illnesses. The algorithm can be installed on a smartphone to provide bedside diagnosis of COPD in acute care settings, inform treatment regimens, and identify those at increased risk of mortality due to seasonal or other respiratory ailments.

Trial Registration

Australian New Zealand Clinical Trials Registry ACTRN12618001521213; http://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=375939