Tiny Machine Learning (TinyML) Based Self Diagnostic Kit for Respiratory Diseases

Heart failure (HF) is a severe cardiovascular disease often worsened by respiratory infections like influenza, COVID-19, and community-acquired pneumonia (CAP). This study aims to uncover the molecular commonalities among these respiratory diseases and their impact on HF, identifying key mediating genes. By performing differential expression analysis on GEO database data, we found 51 common molecules of three respiratory diseases. The gene module of HF was identified by weighted gene co-expression network analysis, and 10 characteristic genes of respiratory diseases that aggravate HF were obtained. GO and KEGG enrichment analysis showed that these genes were mainly involved in innate immune response, inflammation and coagulation pathways. By using three machine learning algorithms, LASSO, RF and SVM-RFE, we identified RSAD2 and IFI44L as key genes, and the Receiver Operating Characteristic (ROC) curve verification results showed high accuracy (Area Under the Curve, AUC > 0.7). ssGSEA showed that RSAD2 was involved in complement and coagulation cascade reactions, while IFI44L was related to myocardial contraction in the progression of heart failure. DSigDB prediction results showed that 6 drugs such as acetohexamide may have potential therapeutic effects on HF aggravated by respiratory diseases. Immune infiltration analysis revealed significant differences in eight immune cell types between HF patients and healthy controls. Our findings enhance the understanding of molecular interactions between respiratory diseases and heart failure, paving the way for future research and therapeutic strategies.

Electret-based flexible pressure sensor for respiratory diseases auxiliary diagnosis system using machine learning technique

Respiratory disease in human respiratory diseases is a leading cause of mortality worldwide, claiming nearly 900,000 lives annually. Early identification is crucial for reducing mortality rates. This review explores the innovative use of machine learning and deep learning technologies in detecting and classifying respiratory diseases, highlighting recent advancements. The review overviews machine learning approaches, and discusses various deep learning algorithms and specialized architectures. Performance evaluation includes support vector machine, logistic regression, artificial neural network, convolutional neural network, recurrent neural network, and long short-term memory, using metrics such as accuracy, precision, recall, F1-Score, and AUC. Among these, the recurrent neural network stands out with an accuracy of (83%), precision of (87%), F1-Score of (91%), and AUC of (91%). However, the artificial neural network shows a higher recall of (96%) compared to other algorithms.

Even though extreme containment and mitigation strategies were implemented by numerous governments around the world to slow down the spread of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), the number of critically ill patients and fatalities keeps rising. This crisis has highlighted the socioeconomic disparities of health care systems within and among countries. As new CoVID policies and responses are implemented to lessen the impact of the virus, it is imperative (1) to consider additional mitigation strategies critical for the development of effective countermeasures, (2) to promote long-term policies and strict regulations of the trade of wildlife and live animal markets, and (3) to advocate for necessary funding and investments in global health, specifically for the prevention of and response to natural and manmade pandemics. This document considers some of these challenges.

Chapter 15 - Clinical decision support systems to improve the diagnosis and management of respiratory diseases

Lung cancer management challenges amidst COVID-19 pandemic: hope lives here.

Non-adherence to prescribed medication is a major public health concern that escalates the risk of morbidity and death as well as incurring extra expenses associated with hospitalisation. According to the World Health Organization (WHO), only 50% of people suffering from chronic diseases follow the treatment recommendations despite the counsel provided to patients on the importance of medication adherence (MA). Early detection of non-communicable disease (NCD) patients poorly adhering to recommended medications using analytics based on machine learning (ML) may improve the outcomes of NCD patients positively. This paper presents a systematic review of literature involving the application of ML in evaluating MA amongst NCD patients. The articles considered in this study were extracted from Web of Science, Google Scholar, PubMed, and IEEE Explore. Twenty-five articles in total met the criteria for inclusion. These were articles that utilised ML techniques to analyse MA in NCDs, with patients suffering from diabetes (n = 8), hypertension (n = 3), cardiovascular disease (CVD) and statin adherence (n = 6), cancer (n = 3), respiratory diseases (n = 2), and other NCD conditions (n = 3). The proportion of days covered (PDC) was typically used to evaluate MA. It emerged from the study that for MA to be considered high, the adherence threshold should be at least 75% of the PDC, a universally accepted threshold. In MA analytics research and practice, a PDC ≥80% threshold is typically regarded as a high level of adherence to prescription medication. Logistic regression (LR) (n = 12), random forest (RF) (n = 11), support vector machine (SVM) (n = 7), neural net (n = 6), ensemble learning (n = 6), MLPs (n = 4), XGBoost (n = 3), Bayesian network (BN) (n = 3), and gradient boosting (n = 3) were the most frequently applied ML techniques in the analytics of MA amongst NCD patients. It should be underscored that leveraging standard ML, deep learning (DL), and ensemble learning has enormous potential for measuring MA amongst NCD patients based on various analytics such as prediction, regression, classification, and clustering. Moreover, a further study could be conducted to comprehend how the application of alternative ML-based techniques can be used to measure MA among patients with chronic infectious diseases.

Targeted omics and systems medicine: personalising care

The computing everywhere paradigm is paving the way for the pervasive diffusion of tiny devices (such as Internet-of-Things or edge computing devices) endowed with intelligent abilities. Achieving this goal requires machine and deep learning solutions to be completely redesigned to fit the severe technological constraints on computation, memory, and power consumption typically characterizing these tiny devices. The aim of this paper is to explore tiny machine learning (TinyML) and introduce tiny deep learning (TinyDL) for the design, development, and deployment of machine and deep learning solutions for (an ecosystem of) tiny devices, hence supporting intelligent and pervasive applications following the computing everywhere paradigm.KeywordsTiny machine learning (TinyML)Tiny deep learning (TinyDL)Internet of thingsEdge computing

A respiratory disorder that attacks COVID-19 patients requires intensive supervision of medical practitioners during the isolation period. A non-contact monitoring device will be a suitable solution for reducing the spread risk of the virus while monitoring the COVID-19 patient. This study uses Frequency-Modulated Continuous Wave (FMCW) radar and Machine Learning (ML) to obtain respiratory information and analyze respiratory signals, respectively. Multiple subjects in a room can be detected simultaneously by calculating the Angle of Arrival (AoA) of the received signal and utilizing the Multiple Input Multiple Output (MIMO) of FMCW radar. Fast Fourier Transform (FFT) and some signal processing are implemented to obtain a breathing waveform. ML helps the system to analyze the respiratory signals automatically. This paper also compares the performance of several ML algorithms such as Multinomial Logistic Regression (MLR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), CatBoosting (CB) Classifier, Multilayer Perceptron (MLP), and three proposed stacked ensemble models, namely Stacked Ensemble Classifier (SEC), Boosting Tree-based Stacked Classifier (BTSC), and Neural Stacked Ensemble Model (NSEM) to obtain the best ML model. The results show that the NSEM algorithm achieves the best performance with 97.1% accuracy. In the real-time implementation, the system could simultaneously detect several objects with different breathing characteristics and classify the respiratory signals into five different classes.

Respiration monitoring plays a vital role in human health monitoring, as it is an essential indicator of vital signs. Respiration monitoring can help determine the physiological state of the human body and provide insight into certain illnesses. Recently, non-contact respiratory illness sensing methods have drawn much attention due to user acceptance and great potential for real-world deployment. Such methods can reduce stress on healthcare facilities by providing modern digital health technologies. This digital revolution in the healthcare sector will provide inexpensive and unobstructed solutions. Non-contact respiratory illness sensing is effective as it does not require users to carry devices and avoids privacy concerns. The primary objective of this research work is to develop a system for continuous real-time sensing of respiratory illnesses. In this research work, the non-contact software-defined radio (SDR) based RF technique is exploited for respiratory illness sensing. The developed system measures respiratory activity imprints on channel state information (CSI). For this purpose, an orthogonal frequency division multiplexing (OFDM) transceiver is designed, and the developed system is tested for single-person and multi-person cases. Nine respiratory illnesses are detected and classified using machine learning algorithms (ML) with maximum accuracy of 99.7% for a single-person case. Three respiratory illnesses are detected and classified with a maximum accuracy of 93.5% and 88.4% for two- and three-person cases, respectively. The research provides an intelligent, accurate, continuous, and real-time solution for respiratory illness sensing. Furthermore, the developed system can also be deployed in office and home environments.

11132 Background: Multimorbidity is common among patients with cancer; however, research on cancer prognosis has predominantly focused on cancers in isolation. This study identified comorbidity clusters among patients with cancer using machine learning and examine their associations with survival outcomes in a nationally representative sample of the US. Methods: This study used 10 survey cycles of the National Health and Nutrition Examination Survey from 1999 to 2018. Participants aged ≥20 years with a self-reported history of cancer were included. Comorbidities were ascertained through self-reports and quantitative measurements. Machine learning techniques, including Bernoulli mixture model and partition-based methods, were used to identify comorbidity clusters. Cox proportional hazards models were used to analyze the associations between comorbidity clusters and mortality outcomes, including all-cause mortality and cause-specific mortality (cancer, cardiovascular disease [CVD], and respiratory diseases), adjusting for relevant covariates. Results: The study included 4,390 participants. Four comorbidity clusters were identified: Low Comorbidity, Metabolic, CVD, and Respiratory. After adjusting for confounders, participants in the Respiratory Cluster had the highest risk of all-cause mortality (adjusted hazard ratio[aHR]=1.62, 95% confidence interval [CI]=1.26-2.08, p<0.001), followed by the CVD Cluster (aHR=1.50, 95%CI = 1.26-1.80, p<0.001) and the Metabolic Cluster (aHR=1.15, 95%CI=1.02-1.29, p=0.03) compared to the Low Comorbidity Cluster. The Metabolic, CVD, and Respiratory clusters were associated with higher risks of cardiovascular disease-related mortality (aHR=1.48-3.05, p<0.003), but no significant differences in cancer mortality were observed among the clusters. The effects of comorbidity clusters on all-cause mortality were modified by income-to-poverty ratio ( p for interaction=0.04), diet quality ( p for interaction=0.02), time since cancer diagnosis ( p for interaction=0.009), and cancer prognosis ( p for interaction=0.005). Conclusions: High comorbidity burden clusters showed increased all-cause and CVD-related mortality. Moreover, diet quality and socioeconomic status modified these associations. Machine learning approaches can provide valuable insights into complex multimorbidity profiles in patients with cancer. Further research is needed to deepen understanding of the relationships between multimorbidity, mortality, and cancer-specific outcomes.

Rationale Multimodal clinical audio data, including speech and breath sounds, play a critical role in assessing respiratory function in disorders affecting the thoracic cage, including neurodegenerative, infectious, and respiratory disorders. Machine learning (ML) has shown promise in identifying patterns to enhance predictive accuracy in these disorders due to its data processing power with chronic respiratory conditions that disrupt airflow, such as Chronic Obstructive Pulmonary Disease (COPD) and Interstitial Lung Disease (ILD), being notable examples. This study investigates the feasibility of using an ML model to predict clinical parameters including pulmonary function, smoking status, and disease presence by analyzing breath and vocal characteristics. Methods Two-hundred patients with various respiratory and thoracic pathologies were recruited via convenience sampling from a university medical center. Data were collected using digital stethoscopes and microphones to record breath sounds at eight anatomical locations along with vocal samples during a variety of vocal tasks (Figure 1). Audio data were converted into Mel-spectrograms and used to train a convolutional neural network (CNN) model to predict patient outcomes, including gender, smoking status, COPD, and ILD. The dataset was divided into a 90:10 training-to-validation ratio. Fig. 1. Representation of study workflow showing data collection, data transformation, and subsequent model training. Results In predicting smoking status, breath recordings achieved 80% accuracy, 95% specificity, 63% sensitivity, and an AUC of 0.82 compared to 65%, 69%, 57%, and 0.58 for the voice-based model. In COPD classification, breath data yielded 75% accuracy, 62% specificity, and 80% sensitivity in contrast to 62.5%, 80%, and 57% for the vocal model. In ILD classification, breath data yielded 65% accuracy, 58% specificity, and 75% sensitivity in contrast to 65%, 77%, and 55% for the vocal model. Additionally, in predicting pulmonary function via Forced Vital Capacity (FVC) and the FEV1/FVC ratio, breath sounds had lower mean squared error (MSE) values relative to voice models (0.7 vs. 10.2, and 111.6 vs. 2308.6), demonstrating superior performance. Conclusion Breath sounds offer significant predictive value in identifying pulmonary conditions and estimating functional metrics, outperforming vocal and cough data in most tasks. Given the low cost of obtaining clinical audio data, uncovering reliable biomarkers in audio patterns via ML methods may assist in diagnosis, risk stratification, and identification of patients in whom preventative care may be beneficial. Future work will require larger datasets to improve model robustness and explore more complex biomarker integration, such as combining breath and voice data, for enhanced clinical prediction.

Detecting acute respiratory diseases in the pediatric population using cough sound features and machine learning: A systematic review

Purpose of reviewAssessment of underlying conditions that contribute to breathlessness is fundamental for symptom management. This review aims to summarize the knowledge from the past two years on the most common underlying conditions among individuals with breathlessness in the general population and to identify research gaps.Recent findingsNine studies from the last two years were included in the review; two studies systematically assessed underlying conditions among breathless individuals in the general population. The modified Medical Research Council (mMRC) scale was used in eight of nine studies. Respiratory diseases were the main underlying condition (40–57%), of which asthma was the most common (approx. 25%), and chronic obstructive pulmonary disease was particularly strongly associated with breathlessness. Other conditions prevalent among breathless individuals included heart diseases, anxiety, depression, and obesity, and several conditions often co-existed.SummaryBreathlessness in the general population is common and associated with several underlying conditions. Respiratory disease is the most commonly reported underlying condition. Refined methods such as machine learning could be useful to study the complex interplay between multiple underlying causes of breathlessness and impact on outcomes such as quality of life and survival.