Breath sounds, asthma, and the mobile phone

Artificial intelligence and machine learning for anaphylaxis algorithms

PURPOSE OF REVIEW

Anaphylaxis is a severe, potentially life-threatening allergic reaction that requires rapid identification and intervention. Current management includes early recognition, prompt administration of epinephrine, and immediate medical attention. However, challenges remain in accurate diagnosis, timely treatment, and personalized care. This article reviews the integration of artificial intelligence and machine learning in enhancing anaphylaxis management.

RECENT FINDINGS

Artificial intelligence and machine learning can analyze vast datasets to identify patterns and predict anaphylactic episodes, improve diagnostic accuracy through image and biomarker analysis, and personalize treatment plans. Artificial intelligence-powered wearable devices and decision support systems can facilitate real-time monitoring and early intervention. The ethical considerations of artificial intelligence use, including data privacy, transparency, and bias mitigation, are also discussed.

SUMMARY

Future directions include the development of predictive models, enhanced diagnostic tools, and artificial intelligence-driven educational resources. By leveraging artificial intelligence and machine learning, healthcare providers can improve the management of anaphylaxis, ensuring better patient outcomes and advancing personalized medicine.

Machine Learning-based Voice Assessment for the Detection of Positive and Recovered COVID-19 Patients

Many virological tests have been implemented during the Coronavirus Disease 2019 (COVID-19) pandemic for diagnostic purposes, but they appear unsuitable for screening purposes. Furthermore, current screening strategies are not accurate enough to effectively curb the spread of the disease. Therefore, the present study was conducted within a controlled clinical environment to determine eventual detectable variations in the voice of COVID-19 patients, recovered and healthy subjects, and also to determine whether machine learning-based voice assessment (MLVA) can accurately discriminate between them, thus potentially serving as a more effective mass-screening tool. Three different subpopulations were consecutively recruited: positive COVID-19 patients, recovered COVID-19 patients and healthy individuals as controls. Positive patients were recruited within 10 days from nasal swab positivity. Recovery from COVID-19 was established clinically, virologically and radiologically. Healthy individuals reported no COVID-19 symptoms and yielded negative results at serological testing. All study participants provided three trials for multiple vocal tasks (sustained vowel phonation, speech, cough). All recordings were initially divided into three different binary classifications with a feature selection, ranking and cross-validated RBF-SVM pipeline. This brough a mean accuracy of 90.24%, a mean sensitivity of 91.15%, a mean specificity of 89.13% and a mean AUC of 0.94 across all tasks and all comparisons, and outlined the sustained vowel as the most effective vocal task for COVID discrimination. Moreover, a three-way classification was carried out on an external test set comprised of 30 subjects, 10 per class, with a mean accuracy of 80% and an accuracy of 100% for the detection of positive subjects. Within this assessment, recovered individuals proved to be the most difficult class to identify, and all the misclassified subjects were declared positive; this might be related to mid and short-term vocal traces of COVID-19, even after the clinical resolution of the infection. In conclusion, MLVA may accurately discriminate between positive COVID-19 patients, recovered COVID-19 patients and healthy individuals. Further studies should test MLVA among larger populations and asymptomatic positive COVID-19 patients to validate this novel screening technology and test its potential application as a potentially more effective surveillance strategy for COVID-19.

Evaluating Listening Performance for COVID-19 Detection by Clinicians and Machine Learning: A Comparative Study

BACKGROUND

To date, performance comparisons between men and machines have been performed in many health domains. Yet, machine learning models and human performance comparisons in audio-based respiratory diagnosis remain largely unexplored.

OBJECTIVE

The primary objective of this study is to compare human clinicians and a machine learning model in predicting COVID-19 from respiratory sound recordings.

METHODS

In this study, we compare human clinicians and a machine learning model in predicting COVID-19 from respiratory sound recordings. Prediction performance on 24 audio samples (12 tested positive) made by 36 clinicians with experience in treating COVID-19 or other respiratory illnesses is compared with predictions made by a machine learning model trained on 1,162 samples. Each sample consists of voice, cough, and breathing sound recordings from one subject, and the length of each sample is around 20 seconds. We also investigated whether combining the predictions of the model and human experts could further enhance the performance, in terms of both accuracy and confidence.

RESULTS

The machine learning model outperformed the clinicians, yielding a sensitivity of 0.75 and a specificity of 0.83, while the best performance achieved by the clinician was 0.67 in terms of sensitivity and 0.75 in terms of specificity. Integrating clinicians' and model's predictions, however, could enhance performance further, achieving a sensitivity of 0.83 and a specificity of 0.92.

CONCLUSIONS

Our findings suggest that the clinicians and the machine learning model could make better clinical decisions via a cooperative approach and achieve higher confidence in audio-based respiratory diagnosis.

Interrater reliability of spectrogram for detecting wheezing in children

BACKGROUND

Auscultation is an easy way to evaluate and diagnose patients with lung conditions but has the shortcoming of being subjective. Using the spectrogram, it is possible to visualize wheezing. We therefore conducted a study to compare the efficacy of diagnosing wheezing by auscultation versus diagnosing wheezing by spectrogram.

METHODS

This was an investigation of interrater reliability and agreement in which the subject population consisted of children, and the rater population consisted of pediatric pulmonologists. We recorded 55 respiratory sound files from June to November 2019. Three pediatric pulmonologists listened to the respiratory sound files and assessed whether wheezing was present. All respiratory sound files were also converted into spectrograms; the same pulmonologists viewed these and assessed whether wheezing was present. We tested for interrater reliability and agreement between the auscultation results and spectrographic results and investigated the diagnostic reliability of auscultation versus spectrogram.

RESULTS

Agreement among the three raters of our auscultation respiratory recordings was 88% and reliability was good (κ = 0.76, P < 0.001). Agreement among the three raters of our spectrograms was 83% and reliability was good (κ =0.66, P < 0.001). The level of agreement between each rater's spectrographic findings and diagnosed wheezing was 91%, 75%, and 93%, respectively. Reliability was accordingly very good, moderate, and very good (κ = 0.82, 0.49, 0.85, P < 0.001), respectively.

CONCLUSIONS

A spectrogram may be a valuable tool for evaluating wheezing in children. It may also be used to improve a young clinician's ability to accurately diagnose wheezing in the future.

Wheezing recognition algorithm using recordings of respiratory sounds at the mouth in a pediatric population

BACKGROUND

Respiratory diseases in children are a common reason for physician visits. A diagnostic difficulty arises when parents hear wheezing that is no longer present during the medical consultation. Thus, an outpatient objective tool for recognition of wheezing is of clinical value.

METHOD

We developed a wheezing recognition algorithm from recorded respiratory sounds with a Smartphone placed near the mouth. A total of 186 recordings were obtained in a pediatric emergency department, mostly in toddlers (mean age 20 months). After exclusion of recordings with artefacts and those with a single clinical operator auscultation, 95 recordings with the agreement of two operators on auscultation diagnosis (27 with wheezing and 68 without) were subjected to a two phase algorithm (signal analysis and pattern classifier using machine learning algorithms) to classify records.

RESULTS

The best performance (71.4% sensitivity and 88.9% specificity) was observed with a Support Vector Machine-based algorithm. We further tested the algorithm over a set of 39 recordings having a single operator and found a fair agreement (kappa=0.28, CI95% [0.12, 0.45]) between the algorithm and the operator.

CONCLUSIONS

The main advantage of such an algorithm is its use in contact-free sound recording, thus valuable in the pediatric population.

Image and diagnosis quality of X-ray image transmission via cell phone camera: a project study evaluating quality and reliability

Introduction

Developments in telemedicine have not produced any relevant benefits for orthopedics and trauma surgery to date. For the present project study, several parameters were examined during assessment of x-ray images, which had been photographed and transmitted via cell phone.

Materials and Methods

A total of 100 x-ray images of various body regions were photographed with a Nokia cell phone and transmitted via email or MMS. Next, the transmitted photographs were reviewed on a laptop computer by five medical specialists and assessed regarding quality and diagnosis.

Results

Due to their poor quality, the transmitted MMS images could not be evaluated and this path of transmission was therefore excluded. Mean size of transmitted x-ray email images was 394 kB (range: 265–590 kB, SD ±59), average transmission time was 3.29 min ±8 (CI 95%: 1.7–4.9). Applying a score from 1–10 (very poor - excellent), mean image quality was 5.8. In 83.2±4% (mean value ± SD) of cases (median 82; 80–89%), there was agreement between final diagnosis and assessment by the five medical experts who had received the images. However, there was a markedly low concurrence ratio in the thoracic area and in pediatric injuries.

Discussion

While the rate of accurate diagnosis and indication for surgery was high with a concurrence ratio of 83%, considerable differences existed between the assessed regions, with lowest values for thoracic images. Teleradiology is a cost-effective, rapid method which can be applied wherever wireless cell phone reception is available. In our opinion, this method is in principle suitable for clinical use, enabling the physician on duty to agree on appropriate measures with colleagues located elsewhere via x-ray image transmission on a cell phone.

The effects of mobile phones on pacemaker function

OBJECTIVES

The electromagnetic field generated by different systems have well-recognized adverse effects on pacemaker functions. The aim of this study is to evaluate the adverse effects of mobile phones on pacemaker functions.

METHODS AND RESULTS

A total of 679 patients with permanent pacemakers were enrolled in this study. The study was performed in two steps. Pacemaker lead polarity was unipolar in the first step and bipolar in the second step. Pacemaker sensitivity was first at nominal values, it was then reduced to the minimal value for that pacemaker and tested again. Two mobile phones were symmetrically located on both sides of the pacemaker pocket with the antennas being equidistant at 50, 30, 20 and 10 cm and in close contact with the pocket. The tests were performed when both mobiles were opened, on stand-by, were receiving a call, during the call and were closed. Thirty-seven patients with pacemakers were adversely affected (5.5%) (33 VVI-R pacemakers were converted to asynchronous mode, and 3 were inhibited, 1 DDD-R pacemaker developed ventricular triggering). When the lead polarity was unipolar, the rate of adverse effect was higher when compared to the bipolar state (4.12% and 1.40%, p<0.01). The increase in sensitivity was not an independent factor on the rate of being affected (p>0.05). The rate of observing an adverse effect increased as the pacemaker got older (p<0.05).

CONCLUSIONS

Mobile phones might have adverse effects on pacemaker functions under certain conditions. This does not result in any symptoms other than the inhibition of pacemakers, and pacemaker functions return to normal when the mobile phones are removed away from the patient.

Automatic wheeze detection based on auditory modelling

Automatic wheeze detection has several potential benefits compared with reliance on human auscultation: it is experience independent, an automated historical record can easily be kept, and it allows quantification of wheeze severity. Previous attempts to detect wheezes automatically have had partial success but have not been reliable enough to become widely accepted as a useful tool. In this paper an improved algorithm for automatic wheeze detection based on auditory modelling is developed, called the frequency- and duration-dependent threshold algorithm. The mean frequency and duration of each wheeze component are obtained automatically. The detected wheezes are marked on a spectrogram. In the new algorithm, the concept of a frequency- and duration-dependent threshold for wheeze detection is introduced. Another departure from previous work is that the threshold is based not on global power but on power corresponding to a particular frequency range. The algorithm has been tested on 36 subjects, 11 of whom exhibited characteristics of wheeze. The results show a marked improvement in the accuracy of wheeze detection when compared with previous algorithms.

Improving asthma control through telemedicine: a study of short-message service

Home peak expiratory flow (PEF) measurement is recommended by asthma guidelines. In a 16-week randomized controlled study on 16 subjects with asthma (24.6 6.5 years old, asthma duration small ze, Cyrillic 6 months), we examined Global System for Mobile Communications (GSM) mobile telephone short-message service (SMS) as a novel means of telemedicine in PEF monitoring. All subjects received asthma education, self-management plan, and standard treatment. All measured PEF three times daily and kept a symptom diary. In the study group, therapy was adjusted weekly by an asthma specialist according to PEF values received daily from the patients. There was no significant difference between the groups in absolute PEF, but PEF variability was significantly smaller in the study group (16.12 +/- 6.93% vs. 27.24 +/- 10.01%, p = 0.049). forced expiratory flow in 1 second (FEV1; % predicted) in the study group was slightly but significantly increased (81.25 +/- 17.31 vs. 77.63 +/- 14.80, p = 0.014) and in the control group, unchanged (78.25 +/- 21.09 vs. 78.88 +/- 22.02, p = 0.497). Mean FEV1 was similar in the two groups both before and after the study. Controls had significantly higher scores for cough (1.85 +/- 0.43 vs. 1.42 +/- 0.28, p < 0.05) and night symptoms (1.22 +/- 0.23 vs. 0.85 +/- 0.32, p < 0.05). There was no significant difference between the groups in daily consumption of inhaled medicine, forced vital capacity, or compliance. Per patient, per week, the additional cost of follow-up by SMS was Euros 1.67 (equivalent to approximately $1.30 per 1 Euro), and SMS transmission required 11.5 minutes. Although a study group of 40 patients is needed for the follow-up study to achieve the power of 80% within the 95% confidence interval, we conclude that SMS is a convenient, reliable, affordable, and secure means of telemedicine that may improve asthma control when added to a written action plan and standard follow-up.

Cellular radio telecommunication for health care: benefits and risks

Cellular radio telecommunication has increased exponentially with many applications to health care reported. The authors attempt to summarize published applications with demonstrated effect on health care, review briefly the rapid evolution of hardware and software standards, explain current limitations and future potential of data quality and security, and discuss issues of safety.

Outpatient monitoring of asthma

Monitoring asthma outcomes is an essential step to the successful implementation of national asthma guidelines. Symptoms, airflow obstruction and exacerbations can be monitored by patients with asthma and by physicians. Patients who practise self-monitoring in conjunction with use of a written action plan and regular medical review have significantly fewer hospitalizations, emergency room visits and lost time from work. Additional monitoring tools are under evaluation, and these include measures of airway responsiveness, airway inflammation, and Internet-based monitoring systems.