A High performance electronic nose system for the recognition of myocardial infarction and coronary artery diseases

doi:10.1016/j.bspc.2020.102247

Biomedical Signal Processing and Control

Volume 64, February 2021, 102247

https://doi.org/10.1016/j.bspc.2020.102247 Get rights and content

Highlights

•
An electronic nose has been made and the breaths of cardiovascular diseases patients were analyzed.
•
The MI patients were separated from the healthy subjects and SCAD patients with a classification accuracy rate of 97.19%.
•
The SCAD patients were separated from the healthy control subjects with a classification accuracy performance of 81.48%.
•
The proposed study could be a guide study in the field of diagnosis of cardiology diseases by using electronic nose.

Abstract

Electronic noses are devices that detect the number and level of chemicals in an odor. This is accomplished by means of chemical gas sensors in the device’s structure. The device recognizes the odors that were introduced to it via the software. When we breathe air, several gas exchanges take place in our lungs (e.g., O2-CO2 at the alveola-capillary level). Some gases in our blood are presented with our breath. The detection of these gases can provide information about our health. Nowadays, many diseases can be diagnosed with very high accuracy by using an electronic nose. Despite advances in diagnoses and treatments, cardiovascular disease is still the leading cause of mortality worldwide. Coronary artery disease is the leading cause of cardiovascular mortality and morbidity. The ability to diagnose coronary artery disease from the breath will accelerate the diagnosis, and thus, the initiation of treatment, which may save many lives. This study, involved an investigation of whether or not diseases (e.g., myocardial infarction, stable coronary artery disease) can be diagnosed from exhaled respiratory air using an electronic nose. This involved collecting data on exhaled breath from 33 patients diagnosed with myocardial infarction that underwent a primary percutaneous coronary intervention, 22 patients with stable coronary artery disease and 26 patients without heart disease. An electronic nose containing 19 gas sensors was manufactured for this study. The respondents’ breath was collected in a sterile manner. The statistical features including mean, skewness, kurtosis and derivative variance were extracted from the breath samples. These features were classified for the entire database using the support vector machine classifier by selecting 66 % as a training set and 34 % as a test set. The breath from the myocardial infarction patients were separated from that of the healthy individuals and the stable coronary artery disease patients with a classification accuracy rate of 97.19 %. The breath from the stable coronary artery disease patients were separated from the breath of the healthy control subjects with a classification accuracy rate of 81.48 %. The results reveal that the proposed method has great potential for myocardial infarction, stable coronary artery disease and healthy subjects when the electronic nose is used to record the exhaled respiratory air of the participants.

Introduction

An electronic nose is a device that uses electrochemical sensors to determine the proportions of volatile chemicals dissolved in the air. Electronic noses contain a variety of gas sensors. They are also commonly used commercially. An extensive number of studies have been conducted recently to determine the freshness [1,2], spoilage [3,4], quality [5,6] and type [7,8] of various foods and beverages using an electronic nose. Electronic noses can also determine indoor air quality [9], and detect air pollution and toxic gases [10,11]. They are also used for environmental monitoring and to detect odors in the cosmetic and fragrance industry [12,13].

In recent years, the detection of disease by the electronic nose has attracted considerable attention. The attraction of an electronic nose diagnosis is the time and cost savings, as well as the ability to take samples from the patient in a non-invasive manner. Many studies have been conducted on diagnosing diseases using an electronic nose. More specifically, lung cancer disease was diagnosed with high accuracy [14,15]. The breath of 10 lung cancer patients was successfully separated from the breath of 10 chronic obstructive pulmonary disease (COPD) patients and 10 healthy individuals [16]. Asthma has been diagnosed [17]. The breath of 30 COPD patients was successfully separated from the breath of 20 asthma patients with 96 % success and 20 healthy smokers with 66 % success [18]. Breath samples of 18 children with obstructive sleep apnea syndrome and 10 healthy children were classified with 64 % accuracy [19]. The urine samples of type 2 diabetic mellitus patients and healthy control groups were separated from each other with 94 % success [20]. The breath of 27 decompensated heart failure (DHF) patients, 25 compensated heart failure (CHF) patients and 28 healthy individuals (control group (CON)) were classified with respective accuracies of 76 % DHF−CON, 87 % DHF−CHF and 85 % CON–CHF by considering 10 times the body odors from their skin with the e-nose [21]. In another study, the exhaled breaths of 10 DHF patients, 16 CHF patients, and 13 healthy subjects were classified with accuracies of 91 % for DHF−CHF, 97 % for CHF−CON and 100 % for DHF−CON, respectively [22].

As stated previously, researchers have generally focused on the diagnosis of lung and respiratory diseases with the electronic nose. This is not where the electronic nose studies should end. Gas chromatography results provided by Machado et al. [23] illustrate that the human breath contains more than two hundred volatile organic compounds, most of which appear in the exhaled breath given by the diffusion of some organic compounds from the alveolar capillary to inhaled air in the lung [24]. This has been the starting point of our hypothesis. In our opinion, diseases that are currently detectable by blood tests are likely to be detected by the breath, despite similar or different chemical mechanisms in the blood.

In light of this hypothesis, we conducted our studies on myocardial infarction (MI). MI is the most important cause of mortality and disability worldwide [25]. Diagnosis is based on clinical features, electrocardiography (ECG) and elevated values of biochemical markers. The most common type of myocardial infarction is an atherosclerotic plaque rupture. Early diagnosis is particularly valuable for the MI group, so they can be revascularized early. The ST segment elevation observed on the ECG is essential for percutaneous coronary intervention (PCI) guidance in MI patients (ST elevation myocardial infarction (STMI)). In patients presenting chest pain, but no ST elevation in the ECG, this diagnosis can only be made by observing the elevation of the cardiac biomarkers. Diagnosis of non−ST elevation myocardial infarction (NSTMI) can only be determined by measuring the level of troponin in the blood; this blood test takes approximately 45–60 min. Therefore, patients suspected of NSTMI experience a minimum delay of 1 h before starting their treatment.

It is very important to quickly diagnose a heart attack to begin treatment as soon as possible. Troponin is a protein released in the cardiac myocytes with impaired ischemic damage; it is also the main biochemical marker in the diagnosis of these patients. Troponin is not the only substance that is released. A significant portion of these other substances are eliminated by renal and hepatic clearance. However, some may be excreted by exhaled breath. This has yet to be scientifically determined.

The diagnosis of stable coronary artery disease (SCAD) is different from that of myocardial infarction. Non-invasive imaging methods for a SCAD diagnosis include: Exercise electrocardiography (ExECG), myocardial nuclear perfusion imaging, single-photon emission computed tomography and computed tomography. If necessary, the diagnosis can be confirmed and revascularization can be performed by coronary angiography (CAG). In this study, SCAD patients consisted of patients who were pre-diagnosed with non-invasive tests and confirmed by CAG; their serious lesions were revascularized with PCI.

In this study, we aimed to demonstrate that the breath of myocardial infarction patients, SCAD patients and healthy individuals, can successfully be separated with an electronic nose to a high accuracy.

Section snippets

Subjects and study design

Thirty-five patients who were directed to the coronary intensive care unit of the Cardiology Clinic from the Emergency Department of the Hitit University Erol Olçok Education and Research Hospital with a diagnosis of MI who underwent CAG for primary PCI, were included in the study. The STMI and NSTMI diagnoses were established by clinical features, ECG findings and elevated myocardial necrosis markers.

The inclusion and exclusion criteria were as follows. Our patient group was formed by patients

Results

In this study, we classified the MI, SCAD and healthy subjects using an electronic nose which recorded the exhaled respiratory air of the participants. The divide and conquer multi-class classification technique was used to determine the final label of the trials. In the first classification node, the trials of the MI patients and the rest of the trials were classified. As previously mentioned, the most effective features were determined after performing the forward feature selection method a

Discussion and conclusion

In this paper, we proposed a statistical features-based method for classifying MI, SCAD and healthy subjects using an electronic nose, which acquired the exhaled breath of the participants. Our approach was successfully applied to our dataset, which consists of 362 breath trials collected from 81 subjects. The experiments proved that the SVM has more capability than the k-NN and NN classifiers. It achieved average CA rates of 97.19 % and 81.48 % on the test data for MI versus the rest and the

CRediT authorship contribution statement

Bilge Han Tozlu: Conceptualization, Methodology, Software, Investigation, Resources, Writing - original draft, Writing - review & editing, Supervision, Project administration. Cemaleddin Şimşek: Conceptualization, Methodology, Writing - original draft, Writing - review & editing, Visualization. Onder Aydemir: Conceptualization, Methodology, Software, Validation, Formal analysis, Writing - original draft, Writing - review & editing, Visualization. Yusuf Karavelioglu: Conceptualization,

Declaration of Competing Interest

None.

References (29)

A. Galdikas et al.
Response time based output of metal oxide gas sensors applied to evaluation of meat freshness with neural signal analysis
Sens. Actuators B Chem.
(2000)
A. Poghossian et al.
Rapid methods and sensors for milk quality monitoring and spoilage detection
Biosens. Bioelectron.
(2019)
T. Yakhno
Chapter 13 - drying drop technology in wine and hard drinks quality control
Handbook of Food Bioengineering, Food Control and Biosecurity
(2018)
M. Pardo
Data preprocessing enhances the classification of different brands of espresso coffee with an electronic nose
Sens. Actuators B Chem.
(2000)
J. He et al.
A high precise e-nose for daily indoor air quality monitoring in living environment
Integration
(2017)
G. Abedi et al.
The survey of analytical methods for sample preparation and analysis of fragrances in cosmetics and personal care products
TrAC Trends Anal. Chem.
(2018)
A. Branca et al.
Electronic nose based discrimination of a perfumery compound in a fragrance
Sens. Actuators B Chem.
(2003)
A. D’Amico
An investigation on electronic nose diagnosis of lung cancer
Lung Cancer Amst. Neth.
(2010)
S. Dragonieri
An electronic nose in the discrimination of patients with non-small cell lung cancer and COPD
Lung Cancer Amst. Neth.
(2009)
P. Montuschi
Diagnostic performance of an electronic nose, fractional exhaled nitric oxide, and lung function testing in asthma
Chest
(2010)

P. Benedek et al.

Exhaled biomarker pattern is altered in children with obstructive sleep apnoea syndrome

Int. J. Pediatr. Otorhinolaryngol.

(2013)

S. Dragonieri et al.

An electronic nose in the discrimination of patients with asthma and controls

J. Allergy Clin. Immunol.

(2007)

R. van de Goor et al.

Training and validating a portable electronic nose for lung cancer screening

J. Thorac. Oncol.

(2018)

S. Guney et al.

An electronic nose system for assessing horse mackerel freshness

Cited by (22)

Open-set adversarial domain match for electronic nose drift compensation and unknown gas recognition
2024, Expert Systems with Applications
In practical applications, electronic noses (EN) must tackle not only sensor drift but also resist interference from unknown gases. In prolonged open environments, electronic noses often encounter unknown gases that cannot be predicted in advance. Current gas detection methods typically rely on prior knowledge of target analytes but struggle to comprehensively address sensor drift and interference from unknown gases. Such as indoor air quality monitoring, long-term stability is crucial, and not all relevant analytes can be predetermined. However, the recent research cannot solve both the sensor drift and unknown gas intrusion problem simultaneously well. In this work, we unify above problems in to the open-set risk boundary. We propose an open-set adversarial domain match (OSADM) model and introduce the considers of open-set domain adaptation (OSDA). OSDA trains a target classifier through matching the domain distribution to recognize the known and unknown gases. First, a binary adversarial loss divides the class boundary. Secondly, adversarial domain adaptation unifies the distribution of different domains. Compared with the metric methods, it avoids complex distribution computation and parameter adjustment to reduce negative transfer. Extensive experimental results on two benchmark datasets, Gas Sensor Array Drift and Twin gas sensor arrays Dataset show that OSADM outperformance of the existing open-set models.
Safety and effectiveness of nano composite hydrogel stent implantation in the treatment of coronary cardiovascular disease: A preclinical study
2023, Preventive Medicine
With the improvement of people's quality of life, various cardiovascular diseases are the most common diseases. Therefore, the main site of disease atherosclerosis is blood vessels, so we can see that its flow rate has obvious changes. Through the analysis of coronary heart disease, this paper studies the relationship between coronary artery disease and cardiovascular disease, which is helpful to evaluate the risk of disease, and also provides the best prevention and treatment plan to overcome cardiovascular disease. As the material of artificial cartilage repair, nanocomposite hydrogel has excellent application value and attraction, because nanocomposite hydrogel has a structure similar to the extracellular matrix of natural chondrocytes. The patients in the experimental group were treated with nano composite hydrogel stent implantation. The other group of patients used the traditional way to carry out the comparative experiment. In the perfusion data of each ventricular wall in the coronary angiography and anterior wall perfusion group, the percentage of lateral wall in the normal proportion was the highest, 69.2%, 59.3% in the anterior wall, 39.5% in the inferior wall, and 19.7% in the apical value and interval. The percentage of LAD stenosis in anterior wall perfusion was O. The highest percentage in the lateral wall was 69.2%, and the lowest in the septum and apex was 19.7%. Nanocomposite hydrogel stent implantation can effectively treat coronary heart disease. The research shows that it is safe and effective in application.
Breath as the mirror of our body is the answer really blowing in the wind? Recent technologies in exhaled breath analysis systems as non-invasive sensing platforms
2021, TrAC - Trends in Analytical Chemistry
Citation Excerpt :
These devices consist of an across-reactive sensor array to replicate the receptors of a human nose and analyze the markers qualitatively based on different sensing techniques [88]. Many efforts have been done to develop highly specific e-nose devices to diagnose diseases such as lung cancer diagnosis [89]. Liu et al. developed an electrochemical e-nose sensor for patients with pulmonary diseases by detecting VOCs in exhaled breath [90].
Health care monitoring is an enormous field of research that has great potential to solve many problems in human life. In recent years, non-invasive health monitoring has become a prerequisite for early diagnosis of various diseases such as lung cancer, oxidative stress, diabetes, to enable the prompt treatment and screening of crucial chemicals. Although analyzing of exhaled breath has been correlated with advanced analytical techniques such as gas chromatography and infrared spectroscopy, breath analyzing biosensing systems offer a cost-effective, sensitive platform for a straightforward analysis. However, current non-invasive sensing strategies have been lacking in practicality in terms of the design and usage, on-site ability and accessibility. This review will critically discuss current commercialized breath analyzers, the recent achievements for the use of the detection towards chemical and biological substances from exhaled breath as non-invasive sensing systems including challenges/drawbacks by addressing the practical applications and concerns in the field. The different fabrication strategies, methodology of detection techniques involved in the development of the breath analyzing systems will be overviewed and discussed along with the future opportunities for possible point of care applications with smartphone integration in this review. The scientific and technological challenges in the field are discussed in the conclusion.
A breathalyzer for the assessment of chronic kidney disease patients’ breathprint: Breath flow dynamic simulation on the measurement chamber and experimental investigation
2021, Biomedical Signal Processing and Control
Citation Excerpt :
Future medical diagnostic procedures are dependent on non-invasive breath analysis by electronic nose systems [1–3].
Chemical sensor technologies with appropriate chamber design influence the kinetic adsorption of exhaled breath. Moreover, emphasize on feature extraction from breathprint may improve the accuracy of e-nose system. In this work, optimization of hardware and software have been elucidated for the improvement in the accuracy of classification between healthy and Chronic Kidney Disease (CKD) breathprint using a commercial sensor (MQ 135). Dimensioning the chamber to accommodate complete alveolar breath along with sensor positioning in the effective flow path of breath sample have been modeled using Computational Fluid Dynamics (CFD) which solved Navier-stokes equation for 3D chambers. The simulation results showed the cylindrical chamber with tangential sensor position to the flow path have effective adsorption of breath sample. The average of saturation voltage of breathprint acquired from 51 hemodialysis patients using the optimized hardware was 13.1% higher than the average saturation voltage of 47 healthy subjects. Fast Fourier Transform (FFT) coefficients extracted from breathprints exhibited high statistical significance between groups (p < 0.05). Susequently, the subset of features ranked by gain ratio algorithm showed highest accuracy of 85.7% for Support Vector Machine (SVM) classifier. These results indicated the importance of hardware and software optimization of e-nose systems for the potential applications in real time disease diagnosis.
Estimation of Concentration Values of Different Gases Based on Long Short-Term Memory by Using Electronic Nose
2021, Biomedical Signal Processing and Control
Citation Excerpt :
VOCs are indicator of different disease types [3]. For these reasons, e-nose can recognize different diseases according to VOCs [4–7]. E-nose [8–10] is a device based on electronic sensing technology, which determines, defines and classifies various features by making precise measurements on volatile components emitted from biological materials with its chemical sensor array.
An electronic nose (e-nose) is commonly used in different areas. In the e-nose studies, one of the most important subjects is the estimation of the different concentration values of different gases. An accurate estimation of gas concentrations plays a very important role in sensitive issues such as disease detection. This study has been carried out to increase the classification and regression successes of concentration values of four different gases detected by 4 metal oxide gas sensors. The different methods are used to compare the success of the classification of the concentration levels and the success of the estimation of concentration values of these all gases. In order to realize these classification and regression processes, first a preprocessing and a feature extraction steps were applied to the raw data. The focus of this study is to increase the success achieved in classification and regression by performing the feature extraction using the proposed method. In the proposed method, “Fully Connected Layer” of Long Short-Term Memory networks was used as a feature extraction. Then, these extracted features were used. The results of the proposed method are compared the other traditional methods. It was observed that there was an improvement in both the classification and regression results with the proposed method. The highest accuracy rate in the classification were obtained in the Support Vector Machine method with 90.8% and in the regression problem, the best mean square errors were obtained with Gaussian Process Regression by using the proposed method.
Machine Learning and XAI approaches for Allergy Diagnosis
2021, Biomedical Signal Processing and Control
Citation Excerpt :
Computer-aided systems aim at assisting decision-makers in decision-making tasks. Different forms of computer-aided systems for enhancing clinical diagnosis and decision support are reported in literature; these include Clinical Decision Support Systems (CDSS) [12], Diagnosis Support Systems [13], Automated Medical Diagnostics [14], Healthcare Information System [15], Electronic Health Record System [16], Intelligent Healthcare Systems [17], Knowledge-based Systems [18], Medical Expert Systems [19], Evidence-based Systems [20], and many others. These systems aid physicians in providing safe and reliable medical services in various domains such as Cardiology [21], Pulmonology [22], Respirology [23], Hepatology [24], Nephrology [25], Dentistry [26], Ophthalmology [26], Dermatology [27] and many others.
This work presents a computer-aided framework for allergy diagnosis which is capable of handling comorbidities. The system was developed using datasets collected from allergy testing centers in South India. Intradermal skin test results of 878 patients were recorded and it was observed that the data contained very few samples for comorbid conditions. Modified data sampling techniques were applied to handle this data imbalance for improving the efficiency of the learning algorithms. The algorithms were cross-validated to choose the optimal trained model for multi-label classification. The transparency of the machine learning models was ensured using post-hoc explainable artificial intelligence approaches. The system was tested by verifying the performance of a trained random forest model on the test data. The training and validation accuracy rate of the decision tree, support vector machine and random forest are 81.62, 81.04 and 83.07 respectively. During evaluation, random forest achieved a rate of 86.39 accuracy overall, and 75% sensitivity for the comorbid Rhinitis-Urticaria class. The framework along with all the functionalities were deployed on mobile devices. The average performance of the clinicians before and after using the decision support system were 77.21% and 81.80% respectively. The diagnosis system integrated with mobile applications serves as a source of information whereby junior clinicians can use it to confirm their diagnostic predictions.

View all citing articles on Scopus

View full text

A High performance electronic nose system for the recognition of myocardial infarction and coronary artery diseases

Highlights

Abstract

Introduction

Section snippets

Subjects and study design

Results

Discussion and conclusion

CRediT authorship contribution statement

Declaration of Competing Interest

Sens. Actuators B Chem.

Biosens. Bioelectron.

Sens. Actuators B Chem.

Integration

TrAC Trends Anal. Chem.

Sens. Actuators B Chem.

Lung Cancer Amst. Neth.

Lung Cancer Amst. Neth.

Chest

Int. J. Pediatr. Otorhinolaryngol.

J. Allergy Clin. Immunol.

J. Thorac. Oncol.

An electronic nose system for assessing horse mackerel freshness