Review of the algorithms used in exhaled breath analysis for the detection of diabetes

Noninvasive Total Cholesterol Level Measurement Using an E-Nose System and Machine Learning on Exhaled Breath Samples

In this paper, the first e-nose system coupled with machine learning algorithm for noninvasive measurement of total cholesterol level based on exhaled air sample was proposed. The study was conducted with the participation of 151 people, from whom a breath sample was collected, and the level of total cholesterol was measured. The breath sample was examined using e-nose and gas sensors, such as TGS1820, TGS2620, TGS2600, MQ3, Semeatech 7e4 NO2 and 7e4 H2S, SGX_NO2, SGX_H2S, K33, AL-03P, and AL-03S. The LGBMRegressor algorithm was used to predict cholesterol level based on the breath sample. Machine learning algorithms were developed for the entire measurement range and for the norm range ≤200 mg/dL achieving MAPE 13.7% and 8%, respectively. The results show that it is possible to develop a noninvasive device to measure total cholesterol level from breath.

Enhanced Diabetes Detection and Blood Glucose Prediction Using TinyML-Integrated E-Nose and Breath Analysis: A Novel Approach Combining Synthetic and Real-World Data

Diabetes mellitus, a chronic condition affecting millions worldwide, necessitates continuous monitoring of blood glucose level (BGL). The increasing prevalence of diabetes has driven the development of non-invasive methods, such as electronic noses (e-noses), for analyzing exhaled breath and detecting biomarkers in volatile organic compounds (VOCs). Effective machine learning models require extensive patient data to ensure accurate BGL predictions, but previous studies have been limited by small sample sizes. This study addresses this limitation by employing conditional generative adversarial networks (CTGAN) to generate synthetic data from real-world tests involving 29 healthy and 29 diabetic participants, resulting in over 14,000 new synthetic samples. These data were used to validate machine learning models for diabetes detection and BGL prediction, integrated into a Tiny Machine Learning (TinyML) e-nose system for real-time analysis. The proposed models achieved an 86% accuracy in BGL identification using LightGBM (Light Gradient Boosting Machine) and a 94.14% accuracy in diabetes detection using Random Forest. These results demonstrate the efficacy of enhancing machine learning models with both real and synthetic data, particularly in non-invasive systems integrating e-noses with TinyML. This study signifies a major advancement in non-invasive diabetes monitoring, underscoring the transformative potential of TinyML-powered e-nose systems in healthcare applications.

Nanowire Array Breath Acetone Sensor for Diabetes Monitoring

Diabetic ketoacidosis (DKA) is a life-threatening acute complication of diabetes characterized by the accumulation of ketone bodies in the blood. Breath acetone, a ketone, directly correlates with blood ketones. Therefore, monitoring breath acetone can significantly enhance the safety and efficacy of diabetes care. In this work, the design and fabrication of an InP/Pt/chitosan nanowire array-based chemiresistive acetone sensor is reported. By incorporation of chitosan as a surface-functional layer and a Pt Schottky contact for efficient charge transfer processes and photovoltaic effect, self-powered, highly selective acetone sensing is achieved. The sensor has exhibited an ultra-wide acetone detection range from sub-ppb to >100 000 ppm level at room temperature, covering those in the exhaled breath from healthy individuals (300-800 ppb) to people at high risk of DKA (>75 ppm). The nanowire sensor has also been successfully integrated into a handheld breath testing prototype, the Ketowhistle, which can successfully detect different ranges of acetone concentrations in simulated breath samples. The Ketowhistle demonstrates the immediate potential for non-invasive ketone monitoring for people living with diabetes, in particular for DKA prevention.

Noninvasive Diabetes Detection through Human Breath Using TinyML-Powered E-Nose

Volatile organic compounds (VOCs) in exhaled human breath serve as pivotal biomarkers for disease identification and medical diagnostics. In the context of diabetes mellitus, the noninvasive detection of acetone, a primary biomarker using electronic noses (e-noses), has gained significant attention. However, employing e-noses requires pre-trained algorithms for precise diabetes detection, often requiring a computer with a programming environment to classify newly acquired data. This study focuses on the development of an embedded system integrating Tiny Machine Learning (TinyML) and an e-nose equipped with Metal Oxide Semiconductor (MOS) sensors for real-time diabetes detection. The study encompassed 44 individuals, comprising 22 healthy individuals and 22 diagnosed with various types of diabetes mellitus. Test results highlight the XGBoost Machine Learning algorithm's achievement of 95% detection accuracy. Additionally, the integration of deep learning algorithms, particularly deep neural networks (DNNs) and one-dimensional convolutional neural network (1D-CNN), yielded a detection efficacy of 94.44%. These outcomes underscore the potency of combining e-noses with TinyML in embedded systems, offering a noninvasive approach for diabetes mellitus detection.

A review of physical, chemical and biological synthesis methods of bimetallic nanoparticles and applications in sensing, water treatment, biomedicine, catalysis and hydrogen storage

This article provides an in-depth analysis of various fabrication methods of bimetallic nanoparticles (BNP), including chemical, biological, and physical techniques. The review explores BNP's diverse uses, from well-known applications such as sensing water treatment and biomedical uses to less-studied areas like breath sensing for diabetes monitoring and hydrogen storage. It cites results from over 1000 researchers worldwide and >300 peer-reviewed articles. Additionally, the article discusses current trends, actionable recommendations, and the importance of synthetic analysis for industry players looking to optimize manufacturing techniques for specific applications. The article also evaluates the pros and cons of various fabrication methods, highlighting the potential of plant extract synthesis for mass production of capped BNPs. However, it warns that this method may not be suitable for certain applications requiring ligand-free surfaces. In contrast, physical methods like laser ablation offer better control and reactivity, especially for applications where ligand-free surfaces are critical. The report underscores the environmental benefits of plant extract synthesis compared to chemical methods that use hazardous chemicals and pose risks to extraction, production, and disposal. The article emphasizes the need for life cycle assessment (LCA) articles in the literature, given the growing volume of research on nanotechnology materials. This article caters to researchers at all stages and applies to various fields applying nanomaterials.

Research progress of electronic nose technology in exhaled breath disease analysis

Exhaled breath analysis has attracted considerable attention as a noninvasive and portable health diagnosis method due to numerous advantages, such as convenience, safety, simplicity, and avoidance of discomfort. Based on many studies, exhaled breath analysis is a promising medical detection technology capable of diagnosing different diseases by analyzing the concentration, type and other characteristics of specific gases. In the existing gas analysis technology, the electronic nose (eNose) analysis method has great advantages of high sensitivity, rapid response, real-time monitoring, ease of use and portability. Herein, this review is intended to provide an overview of the application of human exhaled breath components in disease diagnosis, existing breath testing technologies and the development and research status of electronic nose technology. In the electronic nose technology section, the three aspects of sensors, algorithms and existing systems are summarized in detail. Moreover, the related challenges and limitations involved in the abovementioned technologies are also discussed. Finally, the conclusion and perspective of eNose technology are presented.

Artificial Neural Processing-Driven Bioelectronic Nose for the Diagnosis of Diabetes and Its Complications

Diabetes and its complications affect the younger population and are associated with a high mortality rate; however, early diagnosis can contribute to the selection of appropriate treatment regimens that can reduce mortality. Although diabetes diagnosis via exhaled breath has great potential for early diagnosis, research on such diagnosis is restricted to disease detection, requiring in-depth examination to diagnose and classify diseases and their complications. This study demonstrates the use of an artificial neural processing-based bioelectronic nose to accurately diagnose diabetes and classify diabetic types (type I and II) and their complications, such as heart disease. Specifically, an M13 phage-based electronic nose (e-nose) is used to explore the features of subjects with diabetes at various levels of cellular and organismal organization (cells, liver organoids, and mice). Exhaled breath samples are collected during culturing and exposed to the phage-based e-nose. Compared with cells, liver organoids cultured under conditions mimicking a diabetic environment display properties that closely resemble the characteristics of diabetic mice. Using neural pattern separation, the M13 phage-based e-nose achieves a classification success rate of over 86% for four conditions in mice, namely, type 1 diabetes, type 2 diabetes, diabetic cardiomyopathy, and cardiomyopathy.

Nanosensor Based on Thermal Gradient and Machine Learning for the Detection of Methanol Adulteration in Alcoholic Beverages and Methanol Poisoning

Methanol, naturally present in small quantities in the distillation of alcoholic beverages, can lead to serious health problems. When it exceeds a certain concentration, it causes blindness, organ failure, and even death if not recognized in time. Analytical techniques such as chromatography are used to detect dangerous concentrations of methanol, which are very accurate but also expensive, cumbersome, and time-consuming. Therefore, a gas sensor that is inexpensive and portable and capable of distinguishing methanol from ethanol would be very useful. Here, we present a resistive gas sensor, based on tin oxide nanowires, that works in a thermal gradient. By combining responses at various temperatures and using machine learning algorithms (PCA, SVM, LDA), the device can distinguish methanol from ethanol in a wide range of concentrations (1–100 ppm) in both dry air and under different humidity conditions (25–75% RH). The proposed sensor, which is small and inexpensive, demonstrates the ability to distinguish methanol from ethanol at different concentrations and could be developed both to detect the adulteration of alcoholic beverages and to quickly recognize methanol poisoning.

Polyurethane-Based Porous Carbons Suitable for Medical Application

The main aim of the study was to synthesize and analyze spectral data to determine the structure and stereometry of the carbon-based porous material internal structure. Samples of a porous biomaterial were synthesized through anionic polymerization following our own patent and then carbonized. The samples were investigated using MALDI ToF MS, FTIR ATR spectroscopy, optic microscopy, SEM, confocal laser scanning microscopy and CMT imaging. The analysis revealed the chemical and stereological structure of the obtained porous biomaterial. Then, the parameters characterizing the pore geometry and the porosity of the samples were calculated. The developed material can be used to collect adsorption of breathing phase samples to determine the parity composition of exhaled air.

A Review of Gas Measurement Set-Ups

Measurements of the properties of gas-sensitive materials are a subject of constant research, including continuous developments and improvements of measurement methods and, consequently, measurement set-ups. Preparation of the test set-up is a key aspect of research, and it has a significant impact on the tested sensor. This paper aims to review the current state of the art in the field of gas-sensing measurement and provide overall conclusions of how the different set-ups impact the obtained results.