Nanosensors for Exhaled Breath Condensate: Explored Models, Analytes, and Prospects

Abstract

Exhaled breath condensate (EBC) has gained attention as a diagnostic gateway for lung diseases, brain–gut microbiota dysbiosis, and biobanking. Due to its non-invasive and fast collection method, EBC collection is not under temporal or volume limitations. Nonetheless, conventional EBC screening methods are complex and require high operational costs and expertise. Thus, the advent of nanotechnology has introduced efforts for using nanosensors as EBC analyzers. Over the past decade, multiple EBC-based studies reported the successful use of functionalized nanosensors to trace oxidative stress, tissue damage, and respiratory diseases. The EBC signature includes biomarkers such as gases (H₂O₂ and VOCs), cations (polyamines), fatty acids, cytokines, and aldehydes, in addition to traces of drugs and antibiotics. A common feature of nanosensors is their ability to amplify signals and rapidly detect EBC analytes with high sensitivity and specificity. Based on the collected data, standardizing the collection protocol and read-out methods across laboratories is essential for optimal data comparability. Larger cohorts should be considered to ensure a reliable reproducibility of the reported outputs. Future research directions should employ EBC-based nanosensors to unravel the unexplored omics of lung diseases, especially those linked to the brain–gut microbiota that might influence airway immunity.

1. Introduction

EBC, also known as the condensate of exhaled lung gas, is classified as a biological fluid of the respiratory tract and has recently gained attention for preservation in biobanks [1,2,3,4]. EBC is composed of three major constituents with volatile organic compounds (VOCs) and airway lining fluid (ALF), occupying less than 1% of EBC volume diluted in water vapor [5]. ALF droplets are thought to be released into the airway primarily through mechanical forces acting on the lung and airway surfaces during normal breathing or more forceful respiratory events [6]. The EBC mixture offers a plethora of inflammatory biomarkers that delineate chronic airway diseases such as asthma and cystic fibrosis [7]. With the multi-omics approach, EBC is also analyzed for metabolites, proteins, and genes [8] to diagnose not only airway diseases, but also lung cancer and systemic disorders [9]. Many methods are employed to screen EBC samples, mainly liquid chromatography/tandem mass spectrometry (LC-MS/MS), low-dose computed tomography (LDCT), mass spectrometry, enzyme-linked immunosorbent assay (ELISA), UV spectrophotometry, high-performance thin-layer chromatography (HPTLC), gas chromatography (GC), and voltammetry. However, most of these methods are time-consuming, utilize expensive and hazardous solvents, rely on specimen pre-treatment steps, and require skilled operation. Moreover, large samples are required and are usually collected by invasive techniques such as bronchoalveolar lavage and induced sputum sample collection [10]. Other impediments are linked to weak sensitivity, selectivity, and variability of the EBC signature [11]. This is most likely influenced by intrinsic and extrinsic factors such as the EBC collection device, storage condition, gender, drug intake, and health status [12,13]. Hence, scientists are opting for a simple and non-invasive method for EBC collection and screening [14], achieved by cooling fresh tidal breath [15] using a condenser at −10 to −20 °C [12]. With the advent of nanotechnology, efforts have been improved in the last decade to analyze EBC content through the use of bioactive nanoparticles (NPs).

NPs usually range from 1 to 100 nm [16,17] and possess remarkable physicochemical characteristics with an excellent surface-to-volume ratio and unique optical, chemical, and electrical properties [18,19]. NPs can form diverse shapes, including rods, spheres, and wires, and some can dynamically change shape and chemical properties upon induction with external stimuli [20,21]. Interestingly, NPs are easily functionalized with dyes, polymers, antibodies, and hydrophobic molecular probes, making them ideal for specimen analysis across various fields, including engineering, agriculture, food safety, and mainly in biomedical studies. Remarkably, nanosensors can detect liquid, solid, and gaseous specimens at the nanoscale level. This has prompted the use of NPs to analyze EBC analytes, such as VOCs, interleukins, drugs, and cancer antigens [22,23]. The choice of the nanomaterial and coating layer influences the overall mechanism of action and sensitivity of the nanosensor. NPs are usually classified as lipids, polymers, dendrimers, carbon nanotubes, nanofibers, and other types based on their semiconductor property, such as metal oxide, titanium, or aluminum oxide. In this article, nanoparticles are classified based on their core elemental composition used during synthesis, including gold, silver, copper, carbon nanotubes, quantum dots, and polymers. In addition, NPs can be categorized based on their associated transducer and detection signals. Transducers transfer electrochemical, optical, and thermistor signals to detectors. Detectors are connected to microprocessors equipped with various amplification and read-out mechanisms. For example, optical signals—such as aggregation-induced quenching or energy transfer between the probe and analytes—can be readily detected using a spectrophotometer [24,25]. Notably, the integration of nanoparticles with electrochemically resistant sensors has demonstrated strong potential for the sensitive detection of VOCs in EBC [26,27]. Figure 1 summarizes the key steps in EBC analysis—from sample collection using an EBC condenser, through nano-based biosensing and detection platforms, to their integration into clinical diagnostics.

Currently, a lack of consensus exists regarding the optimal nanomaterial, functionalization strategies, and analytical techniques to standardize clinical applications. Moreover, the integration of nanosensors with multi-omics platforms—such as transcriptomics, proteomics, and metabolomics—has yet to be fully explored, particularly in relation to the brain–gut axis and its impact on airway immunity. This review compiles the existing nanosensor-based approaches for EBC analysis, focusing on the design of the sensor, its mode of detection, sensitivity, and specificity. This article addresses the limitations of various nanoformulations with key controversies surrounding EBC biomarker reproducibility and standardization. Greater emphasis should be placed on gold nanoparticles (AuNPs) due to their superior properties [28,29] in detecting unexplored genomics and transcriptomics analytes in EBC.

2. Data Collection

An extensive search of relevant studies, without implementing date restrictions, was performed using Google Scholar as a search engine, given its wide indexation algorithm. The sentence ‘Exhaled breath condensate detection using nanoparticles’ was used for the first round of screening, followed by more specific terms, namely ‘Exhaled Breath Condensate’ AND ‘Biomarker Detection’ AND ‘Nanoparticles’ AND ‘Nanosensors’. All studies reported on the first 20 pages of each search were screened for relevancy. Accordingly, 400 studies were retrieved, with only 30 studies meeting our criteria. Before 2012, only two studies had been published, highlighting the early stages of nanotechnology integration into EBC analysis. Over the past decade, advancements in nanosensor design have progressed gradually, with only six publications emerging between 2023 and 2025. To date, research has focused on eight core nanoparticle (NP) types—iron oxide, gold, silver, copper, quantum dots, cobalt oxide, polymers, and carbon nanotubes—with findings systematically categorized into five tables as depicted in the next section. Each table is divided into two sections: (a) parameters used in the EBC methodology and nanosensor-based analysis, and (b) categorized nanosensor performance in reference to age range, group size, and health status. Overall, the EBC signature was tested for 15 biomarkers (7 inflammatory, 4 cancer-related, 4 stress-related) and 16 other molecules linked to drugs/antibiotics.

3. Nanoparticle-Based Sensors for Biomolecule Detection in EBC

3.1. Iron Oxide/Magnetic Nanoparticles

Magnetic iron oxide NPs (MIONPs) have been explored in therapeutic applications, magnetic resonance imaging (MRI), in vivo diagnostics [30], and lately have showed promising applications in EBC screening with high reproducibility, sensitivity, and selectivity (Table 1a,b). In the following section, various forms of MIONPs are described with a general overview of the model design and performance.

MIONPs can be functionalized with hydrophobic octadecyl [31], carbon (C) [32], or sodium dodecyl sulfate (SDS) [33] for improved selectivity to biomolecules. For quantification, biomolecules are separated from MIONPs using magnetic solid-phase extraction (MSPE).

Octadecyl hydrocarbons, Tween-20 magnetic nanoparticles (TW-C8-Fe₃O₄), can detect trace amounts of aldehydes in the EBC of healthy and lung cancer patients. Tween-20 surfactants reduced the adsorption of macromolecules from the EBC sample matrix. Aldehydes were quantified using high-performance liquid chromatography–photo diode array detector (HPLC-PDA). After the routine optimization of the nanosystem using aldehyde-spiked EBC samples, a 21.5–29 nmol/L limit of detection (LOD) range was obtained, with the successful quantification of aldehydes in the EBC of healthy people and lung cancer patients. Higher pentanal and nonanal aldehyde concentrations were detected in the EBC samples of lung cancer patients compared with healthy individuals, whereas heptanal was exclusively found in healthy EBC samples. This method did not account for crucial habits such as brushing, smoking, and environmental toxin exposure [31].

In another study, a simpler spectrofluorometric analysis employed carbon-coated magnetic nanoparticles (C/Fe₃O₄ MNs) for the detection of the daclatasvir drug in EBC. EBC was collected from subjects during tidal breathing and then spiked with DAC for further system optimization. Interestingly, a wide linear range [34] of 0.5–15 ng/mL and a low LOD of 0.12 ng/mL with good DAC selectivity were presented by C/Fe₃O₄ MNs. Moreover, no interference of co-administered drugs with C/Fe₃O₄ MNs performance was detected [32]. On the other hand, SDS-MIONPs were used to test the metoprolol drug in plasma, urine, and EBC samples. Even though SDS-MIONPs successfully detected metoprolol in metoprolol-spiked samples, the method was not sensitive to the EBC samples of metoprolol-treated subjects compared to urine and plasma [33]. In a different model, a more sensitive hybrid multi-plate fluorescent intensity reader (BioTek Synergy H1) was designed using iron/iron oxide (Fe/Fe₃O₄)-based nanosensors. This approach presented femtomolar sensitivity outperforming other MIONPs, most likely due to the use of R-Tube equipment and a resin platform topped with NPs and peptides [35,36,37,38,39]. Despite the small pool size of asthmatic versus non-asthmatic participants (3 subjects each), the design approach is promising to detect cytokines, such as IL-6 and CCL20, from EBC. Therefore, non-invasive EBC sampling is more suitable for future clinical applications with a standardized approach for EBC chemokine and enzymatic cleavage activity [40].

A metal–organic nanocomposite hydrogel was developed for quantifying metformin [13] and deferiprone (DEF) in EBC. The method exhibited a linear range [34] of 0.005–2.0 µg/mL and an LOD of 0.001 µg/mL for MET, enabling precise diabetes management and drug metabolism assessment. Additionally, the hydrogel facilitated DEF analysis in EBC, highlighting its potential for non-invasive therapeutic drug monitoring in personalized medicine [41,42].

Table 1. Iron oxide nanoparticles/magnetic nanoparticles.

(a) EBC Extraction Methodology and Nanosensor-Based Analysis
EBC Methodology					Nanosensor
Refs	Condenser	Extraction Temp/Time	Vol.	Storage Temp/Time	Biomarker/Drug	Outcomes
Xu et al., 2014 [31]	Commercial RTubeTM technique	Magnetic solid-phase extraction n/a 10 min	0.7–2.5 mL	−20 °C n/a *	Aldehydes (butanal, pentanal, hexanal, heptanal, octanal, nonanal)	Dual function of TW-C8-Fe3O4 NPs: selective capturing of low-molecular-weight compounds by hydrophobic interaction with simultaneous size exclusion and hydrophilic barrier to macromolecules. Proven optimal MNP sensitivity to aldehydes in 2 mL sample solution (containing 20 µL of formic acid), 0.245 mmol/L of DNPH, 3 mg of TW-C8-Fe3O4 NPs, 100 µL of methanol desorption solvent, and a 5 min reaction time Good TW-C8-Fe3O4 NPs dispersibility in aqueous solution with high selectivity, fast extraction, and convenient operation Heptanal detection only in EBCs of healthy people
Tarfiei et al., 2020 [33]	Lab-made setup	n/a 20 min	n/a	−20 °C n/a	Metoprolol drug	Optimal performance: a pH of 3–4, 10 mL of water, SDS 5–6 mM, and 2.0 mL of acetonitrile for desorption The adsorbent can be reused at least up to 4 times Retained MIONP stability after 9 months of storage Precise and accurate findings according to the FDA guidelines No reliable concentration of metoprolol detected in the EBC of real metoprolol-treated subjects Metoprolol detection in higher EBC concentrations
Heidari et al., 2022 [32]	Lab-made setup	n/a	n/a	n/a	Daclatasvir drug	Optimal performance: 0.5 mL of acetonitrile, 6.45 (%w/v; NaCl), 11.75 pH, and a 5 min desorption time C/Fe3O4 MNs could be used a minimum of eight times without reducing the extraction efficiency No interference by co-administered drugs Good selectivity for measuring DAC in EBC samples
Zambrano et al., 2022 [40]	RTubeTM technique	Room temp/ 5–15 min	1 mL	−80 °C n/a	ADAM 33; Granzyme B Chemokine (CCL20) IL-6; MMP-8; NSE; arginase	Both mouse bronchoalveolar lavage and EBC showed highly sensitive nanosensors as tracked by energy quenching or emission using a quenching sequence. Mild asthma EBC with significant increase in all biomarkers compared to control, except for ADAM33 and NE
(b) Performance Parameters of Iron Oxide Nanoparticles/Magnetic Nanoparticles
Refs.	Age	Groups (N)/Health Status		Read-Out Method	Reproducibility/Storage	Linear Range [µmol/L]	LOD [nmol/L]
Xu et al., 2014 [31]	n/a	>Healthy (8) (used for calibration after spiking with aldehydes) >Lung cancer (6)		NP-aldehyde metabolites detected by high-performance liquid chromatography–photo diode array detector HPLC/PDA	RSD of 2.9–13.1% for all intra- and inter-day standard determinations	Butanal 0.02–7 Pentanal 0.01–9 Hexanal 0.05–9 Heptanal 0.005–7 Octanal 0.005–9 Nonanal 0.05–10	Butanal 13.2 Pentanal 2.9 Hexanal 21.5 Heptanal 4 Octanal 3 Nonanal 21.4
Tarfiei et al., 2020 [33]	n/a	>Healthy (n/a) (spiked with metoprolol) >Metoprolol-treated (n/a)		Spectrofluorophotometer	Intra- and inter-day precision values were lower than 5%	n/a	2.29 ng/mL
Heidari et al., 2022 [32]	n/a	Healthy (n/a) (spiked with daclatasvir)		FP-750 spectrofluorometer	Intra- and inter-day RSD of 3.2 and 3.9%	0.5–15 ng/mL	0.12 ng/mL
Zambrano et al., 2022 [40]	n/a	>Mice (8) >Non-asthmatic patients (3) >Mild asthma patients (3)		Fluorescence intensity	Stored > 1 year at 253 K under argon	n/a	Sub-femtomolar (10 × more sensitive than ELISA or other immunoassays) at least 10−14 M

* n/a indicates that the item has not been assigned.

Table 1. Iron oxide nanoparticles/magnetic nanoparticles.

(a) EBC Extraction Methodology and Nanosensor-Based Analysis
EBC Methodology					Nanosensor
Refs	Condenser	Extraction Temp/Time	Vol.	Storage Temp/Time	Biomarker/Drug	Outcomes
Xu et al., 2014 [31]	Commercial RTubeTM technique	Magnetic solid-phase extraction n/a 10 min	0.7–2.5 mL	−20 °C n/a *	Aldehydes (butanal, pentanal, hexanal, heptanal, octanal, nonanal)	Dual function of TW-C8-Fe3O4 NPs: selective capturing of low-molecular-weight compounds by hydrophobic interaction with simultaneous size exclusion and hydrophilic barrier to macromolecules. Proven optimal MNP sensitivity to aldehydes in 2 mL sample solution (containing 20 µL of formic acid), 0.245 mmol/L of DNPH, 3 mg of TW-C8-Fe3O4 NPs, 100 µL of methanol desorption solvent, and a 5 min reaction time Good TW-C8-Fe3O4 NPs dispersibility in aqueous solution with high selectivity, fast extraction, and convenient operation Heptanal detection only in EBCs of healthy people
Tarfiei et al., 2020 [33]	Lab-made setup	n/a 20 min	n/a	−20 °C n/a	Metoprolol drug	Optimal performance: a pH of 3–4, 10 mL of water, SDS 5–6 mM, and 2.0 mL of acetonitrile for desorption The adsorbent can be reused at least up to 4 times Retained MIONP stability after 9 months of storage Precise and accurate findings according to the FDA guidelines No reliable concentration of metoprolol detected in the EBC of real metoprolol-treated subjects Metoprolol detection in higher EBC concentrations
Heidari et al., 2022 [32]	Lab-made setup	n/a	n/a	n/a	Daclatasvir drug	Optimal performance: 0.5 mL of acetonitrile, 6.45 (%w/v; NaCl), 11.75 pH, and a 5 min desorption time C/Fe3O4 MNs could be used a minimum of eight times without reducing the extraction efficiency No interference by co-administered drugs Good selectivity for measuring DAC in EBC samples
Zambrano et al., 2022 [40]	RTubeTM technique	Room temp/ 5–15 min	1 mL	−80 °C n/a	ADAM 33; Granzyme B Chemokine (CCL20) IL-6; MMP-8; NSE; arginase	Both mouse bronchoalveolar lavage and EBC showed highly sensitive nanosensors as tracked by energy quenching or emission using a quenching sequence. Mild asthma EBC with significant increase in all biomarkers compared to control, except for ADAM33 and NE
(b) Performance Parameters of Iron Oxide Nanoparticles/Magnetic Nanoparticles
Refs.	Age	Groups (N)/Health Status		Read-Out Method	Reproducibility/Storage	Linear Range [µmol/L]	LOD [nmol/L]
Xu et al., 2014 [31]	n/a	>Healthy (8) (used for calibration after spiking with aldehydes) >Lung cancer (6)		NP-aldehyde metabolites detected by high-performance liquid chromatography–photo diode array detector HPLC/PDA	RSD of 2.9–13.1% for all intra- and inter-day standard determinations	Butanal 0.02–7 Pentanal 0.01–9 Hexanal 0.05–9 Heptanal 0.005–7 Octanal 0.005–9 Nonanal 0.05–10	Butanal 13.2 Pentanal 2.9 Hexanal 21.5 Heptanal 4 Octanal 3 Nonanal 21.4
Tarfiei et al., 2020 [33]	n/a	>Healthy (n/a) (spiked with metoprolol) >Metoprolol-treated (n/a)		Spectrofluorophotometer	Intra- and inter-day precision values were lower than 5%	n/a	2.29 ng/mL
Heidari et al., 2022 [32]	n/a	Healthy (n/a) (spiked with daclatasvir)		FP-750 spectrofluorometer	Intra- and inter-day RSD of 3.2 and 3.9%	0.5–15 ng/mL	0.12 ng/mL
Zambrano et al., 2022 [40]	n/a	>Mice (8) >Non-asthmatic patients (3) >Mild asthma patients (3)		Fluorescence intensity	Stored > 1 year at 253 K under argon	n/a	Sub-femtomolar (10 × more sensitive than ELISA or other immunoassays) at least 10−14 M

* n/a indicates that the item has not been assigned.

3.2. Gold Nanoparticles

Gold nanoparticles (AuNPs) possess beneficial optical and electronic properties, are easily synthesized, highly stable, and present controllable size- and shape-dependent properties [43]. AuNPs possess anti-cancer and antimicrobial properties and are widely applied for drug delivery and biomarker nanosensing [44]. Interestingly, AuNPs present a great affinity to biomolecules [45,46], with significantly low cytotoxicity on healthy cells [47].

AuNP-based nanosensors have been applied to screen for several volatile biomarkers with only one study on drug detection (Table 2a,b). Peng et al. used gold nanosensors to compare the profile of endogenous VOCs from the EBCs of 177 healthy versus cancer patients (breast, lung, prostate, colon) to distinguish the cancer subtype and stage. Volunteers were asked to refrain from any alcohol or coffee intake for 12 h prior to EBC collection. Gold nanosensors were functionalized using 14 different organic compounds and deposited on silicon wafers capped with thermal oxide. EBC was filtered directly from the mouthpiece to extract > 99.99% of exogenous VOCs and collected in a bag for analysis during a 4-day period. EBC was either exposed to a nanosensor connected to a computer to track the electrical resistance or wired to gas chromatography–mass spectrometry (GC-MS) for direct analysis. In comparison, fresh air from volunteers was directly analyzed by the GC-MS without the use of nanosensors. Based on their results, nanosensors, unlike GC-MS, are insensitive to confounding factors such as ethnicity, age, work pollution, drug exposure, and family cancer history. The general analysis of VOC clusters or patterns was decoded between healthy and cancer cases. Both lung and breast cancer patients showed distinguishable VOC clusters with minimal overlap. For prostate and colon cancer, clusters are similar, with some VOCs overlapping between control and patient groups. Yet, specific VOCs were only detected by GC-MS from freshly withdrawn EBC [48]. The sensor module was adopted from a previous study examined by the same team to distinguish VOCs from lung cancer patients [34].

In another attempt, a surface acoustic wave (SAW) immunosensor, composed of ∼30 nm spherical AuNPs, was coated with anti-carcinoembryonic antigen (CEA) antibodies for detection in EBC. A LR of 1–16 ng/mL was observed between the immunosensor response and CEA concentration, with a 1 ng/mL LOD. Additional testing was performed on 1 to 3 mL human EBC samples collected by the EcoScreen commercial breath condensate collection device at −10 °C for 10 min (immediate use). Comparing the performance of the immunosensor with a commercial instrument, a high data correlation coefficient (r = 0.999) confirmed the efficiency of the immunosensor for a CEA range between 1 and 10 ng/mL [49]. This model was validated to specifically detect CEA in EBC after it was originally introduced as a preliminary comparison between CEA, neuron-specific enolase [21], and squamous cell carcinoma antigen (SCC) detection in the EBCs of healthy and lung cancer patients [50]. On the other hand, a stronger sensitivity for CEA, neuron-specific enolase [21], and SCC, in picogram levels, was achieved by a supreme technique based on surface-enhanced chemiluminescence (SEECL). In this model, AuNPs reacted with Na2SiO3 in the presence of HCl to form Au@SiO_2, then topped with Ru(bpy)₃²⁺. The SEECL was designed using a 16-hole electrode (4 × 4) layered with graphene–COOH prior to the addition of primary antibodies (anti-CEA, NSE, SCC). Then, the antigens were sandwiched under a layer of secondary antibodies specific to the tested biomarkers. EBC antigens were collected from volunteers who had fasted for 8 hrs. The ECL signal was exposed for 10 s and captured with an sCOMS camera followed by signal analysis. The immunoassay demonstrated comparable results, with high sensitivity and selectivity in picogram ranges for multiple antigens tested simultaneously in the same electrode array [51]. The team also investigated CEA markers using SEECL from the EBC collected using the same experimental setup [52]. Thus, the model relied on Au@SiO₂ topped with Ru(bpy)₃²⁺, but instead of using antibodies, aptamers were selected, and another layer of Au–graphite-like carbon nitride nanohybrids acted as energy donors, with Rub(bpy)₃²⁺ as an energy acceptor. A resonance energy transfer signal in the presence of CEA was generated with detection limits in picogram range superior to values obtained from ELISA in the nanogram ranges. Moreover, sensors lacking AuNPs showed a much lower signal. This could be attributed to the localized plasmon surface resonance (LPSR) of gold NPs and the stability mediated by NPs in the model. Overall, the sensors showed high stability with a relative standard deviation (RSD) < 2% of six different sensors tested at different time points. Similarly, selectivity and sensitivity displayed an accurate output with low noise to background signals as collected by cyclic voltammetry (CV) readings using an ultra-weak luminescence apparatus. Cumulatively, biosensors show high reliability and potential to be applied in clinical trials.

Following the extensive applications of imprinting matrices with polymers to detect ions from organic or inorganic compounds [53,54], a screen-printed gold electrode (Au-SPE) was used to immobilize sodium nitrite (NaNO₂), then it was layered with polyvinyl alcohol (PVA). The complex was then crosslinked with glutaraldehyde (GA) to generate Ion-Imprinted Polymer (IIP) sensors [55]. The sensitivity of nitrites from freshly withdrawn EBC was not affected by the simple mouthpiece lab-based collection method. The detection limit was at micromolar levels comparable to other electrodes reported in the literature. Moreover, the selectivity was specific to nitrites without any background signals to other compounds such as nitrate, acetate, or ammonium nitrate. The team just condensed air using a freezing pipe, and the liquid was collected in a beaker. Volunteers were asked to wash their mouths prior to sample collection with frequent saliva swallowing to keep a dry mouth condition. The citrate assisted in the stabilization of the AuNPs and prevented aggregation by forming an electrostatic repulsion layer. However, the use of ferri/ferrocyanide to probe surface changes carries a lot of misinterpretations. Several studies have discussed the concerns and drawbacks of using such a redox couple with gold-plated surfaces, mainly due to plate damage caused by unbound cyanide ions [56], the instability of thiol groups at the gold surface [57], and the presence of chloride ions in the PBS, which may affect the kinetics of the redox couple [58].

Another NP-based tool used citrate-capped gold nanoparticles (AuNP–citrate) to attract positively charged polyamines from 1 mL of EBC samples. EBC was collected from 12 healthy volunteers using a commercially available EBC kit at room temperature and stored at −20 °C. The extraction capacity peaked after 20 min of EBC incubation with AuNP concentration at around 90 nM. In their method, to release polyamines, 2-mercaptoethanol (2-ME) at 0.1 M was used, as it possesses stronger covalent bonding with gold surfaces, forming Au-S bonds. The read-out assay was based on capillary electrophoresis (CE) linked to a conductivity detector (C4D) and field-amplified sample stacking (FASS). AuNPs showed promising sensitivity compared to conventional extraction methods, with a high enrichment of polyamines in satisfactory volumes of 1 mL. The limitation in this method was the involvement of only 12 healthy volunteers without enrolling patients [59].

The AuNP-based sensor for EBC drug detection was designed as sucrose–AuNPs (Suc-AuNPs) while relying on sucrose hydroxyl and oxygen groups interacting with the analyte amine groups. The targeted analyte in EBC was daclatasvir [60,61], which was detected at low levels limited to 8 ng/mL. The nanoprobes proved to be highly stable, selective, and reproducible. Like other studies, high recoveries of up to 99 and 100% were tested at different analyte spike concentrations with the RSD not exceeding 1%. The team also compared DAC from plasma samples, yet the ratio of DAC to EBC/plasma was not evaluated to test its usage in clinical applications. The assay was quite fast, with only a 5 min incubation of a small EBC volume (200 µL) in the presence of the same volume of Suc-AuNPs and 100 µL of PBS. This highlights the simplicity of using this method, especially since EBC does not undergo any pre-treatment and can be used fresh at room temperature with a colorimetric change from pink to blue, accompanied by UV–Vis spectroscopy at a λmax of 532 nm. The model holds promising clinical applications for monitoring drugs or disease breath signatures [62,63]. Another drug, phenytoin, was detected in EBC patients treated for epilepsy. AuNPs were modified using beta-cyclodextrin, and the method relied on the aggregation of -OH groups from beta-cyclodextrin with the amines of phenytoin. A linear relationship was shown with a phenytoin concentration in the range of 0.005–0.6 µg/mL, with a limit of detection of 0.003 µg/mL [64]. An enzyme-mimic optical probe utilizing UiO-66/Au NP-PVA nanocomposite hydrogel was developed for chlordiazepoxide detection in EBC. The probe catalyzed the H₂O₂–tetramethylbenzidine reaction, achieving a detection limit of 0.0032 µg/mL and an LR of 0.005–2.0 µg/mL under optimal conditions [65].

Table 2. Gold nanoparticles.

(a) EBC Extraction Methodology and Nanosensor-Based Analysis
EBC Methodology						Nanosensor
Refs.	Condenser		Extraction Temp/Time	Vol.	Storage Temp/ Time	Biomarker/Drug	Outcomes
Peng et al., 2009 [34]	Mouthpiece with filter (EcoMedicDuerten)		Room temp/ 5 min	n/a *	Air stored in Mylar bags; analyzed within 2 days of collection	Lung cancer VOCs	Proper functionalization of AuNPs with organic moieties Pattern analysis of VOCs showed well-confined and non-overlapping clusters of VOCs between healthy and lung cancer patients Air analyte detected without pre-treatment or dehumidification Electrical resistance to analytes with low detection limits
Peng et al., 2010 [48]	Mouthpiece with filter (EcoMedic Duerten)		Room temp/ 3–5 min	n/a	Air stored in Mylar bags; analyzed within 4 days of collection	Endogenous cancer VOCs	Air trapped in bag directly sensed by GC-MS or by electrical resistance of GNP sensors connected to a computer detection system Cluster and pattern analysis of VOC cancer subtypes is plotted in comparison to healthy samples Nanosensors are insensitive to cofounding factors (gender, age, stress, food additives, drug exposure, work pollution)
Zou et al., 2014 [50]	Commercial EBC collection device (EcoScreen (Jaeger)		−10 °C/ 10 min	n/a	Immediate use	CEA, NSE, and SCC	Small 50 μL EBC sample required Rapid detection, low cost, and simple procedure Higher levels of CEA, NSE, and SCC detection in EBCs of patients than in EBCs of controls Only the concentration of CEA showed significant differences between control and case groups
Zhang et al., 2015 [49]	Commercial EBC collection device (EcoScreen (Jaeger)		−10 °C/ 10 min	1–3 mL	Immediate use	CEA	Significant detection data correlation between the immunosensor and a commercial instrument (coefficient = 0.999) Rapid detection through real-time measurement (in minutes) High specificity and insignificant potential interference from other biomarkers (neuron-specific enolase (NSE) and squamous cell carcinoma (SCC))
Chen et al., 2019 [59]	Commercial disposable EBC collection system (Ecoscreen Jaeger)		n/a	Diluted to 900 μL with ultrapure water	−20 °C n/a	Polyamines	Proven optimal system performance in 1 mL of EBC, 87 nM AuNPs concentration, a 20 min incubation period, 0.10 M 2-ME, and 10 μL of anhydrous methanol dispersant Low extraction time and sample consumption by AuNP-ME/FASS-CE-C4D method Successful polyamine separation from potential coexisting substances and direct determination without derivatization
Ding et al., 2020 [52]	Frozen pipe mouthpiece		n/a	30 µL	Freshly used	Nitrite ions	Sensitive sensor to micromolar nitrite levels Selective sensor to nitrites, without signals to nitrate, acetate, or ammonium nitrate
Ding et al., 2022 [51]	Condensate device		Room temp	0.5 mL	EBC centrifuged at 5000 rpm and supernatant freshly used	CEA	High sensitivity: superior detection level in pg/mL compared to ELISA in ng/mL Energy transfer to Ru(bpy)32+ was amplified with Au NPs High selectivity and stability Au@SiO2-Ru showed an almost 20 nm NP size with a 5 nm silicon corona shell Increasing CEA concentrations generate closer aptamer-to-aptamer binding; better energy transfer from Au-g-C3N4 to Au@SiO2-Ru
Diouf et al., 2020 [55]	Condensate device		Room temp/ n/a	0.5 mL	EBC centrifuged at 5000 rpm and supernatant freshly used	CEA, NSE, and SCC	Novel design of an electrode array (4 × 4) to check 3 cancer biomarkers High sensitivity and selectivity to the tested biomarkers as compared with other non-target proteins Comparable values using the sensor and ELISA tests
Karimzadeh et al., 2022 [61]	Lab-made setup		Room temp/ n/a	200 µL	n/a	Daclatasvir drug	DAC interacts with the oxygen and hydroxyl groups of sucrose, causing aggregation and color change from pink to blue Increase in DAC lowers LSPR peaks The sensor detects DAC both in EBC and plasma samples with comparable recoveries Readings should be made in <5 min
(b) Design and Performance Parameters of AuNPs Nanosensors
Refs.	Age	Groups (N)/Health Status			Biomarkers/Drugs	Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Peng et al., 2009 [34]	28–60	>Healthy (56) >Lung cancer (40)			Endogenous cancer VOCs	Electrical resistance signals detected by computer system	>90% reproducibility for readings on different days	n/a	1 to 5 ppb
Peng et al., 2010 [48]	20–75	>Lung cancer (30) Colon cancer (26) >Breast cancer (22) >Healthy (22–59)			Endogenous cancer VOCs	Electrical resistance signals detected by computer system	>87% reproducibility for a specific volunteer who was examined multiple times over a period of 6 months	300–400 different VOCs/ breath	n/a
Zou et al., 2014 [50]	n/a	>Healthy (13) (spiked with CEA or NSE) >Lung cancer (17)			CEA, neuron-specific enolase, and SCC	Love Wave sensor; surface acoustic wave (SAW) immunosensor	10 replicates for each biomarker RSD (n/a) used to calculate precision and accuracy	n/a	CEA, 0.967 ng/mL NSE, 1.598 ng/mL SCC, 0.663 ng/mL
Zhang et al., 2015 [49]	n/a	Healthy (28)			CEA	Love Wave sensor	3 repetitions performed to confirm linearity	1–16 ng/mL	1 ng/mL
Chen et al., 2019 [59]	n/a	>Healthy (12) (spiked with polyamines)			Polyamines	Field-amplified sample stacking (FASS) coupled with capillary electrophoresis and FASS-CEC4D	RSDs for intraday reproducibility: 1.0–1.8% and 1.1–2.7% (n = 7), respectively; RSDs for inter-day reproducibility: within 4.7% (n = 5)	0.20–20 ng/mL	0.070–0.17 ng/mL
Ding et al., 2020 [52]	n/a	Healthy (8)			Carcinoembryonic antigen (CEA)	SEECL with resonance energy transfer [66] recorded by CV using ultra-weak luminescence	10 consecutive repetitions without CEA with RSD of 1.67%; 6 different sensors tested at different time points with CEA; RSD 2.14%	1.0 pg/mL–5.0 ng/mL	0.3 pg/mL
Ding et al., 2022 [51]	n/a	Healthy (n/a)			CEA, neuron-specific enolase, and SCC	SEECL-I compared with ELISA results	6 batches tested with 1 ng/mL of antigens with RSD for CEA 3.02%, NSE 2.91%, and SCC 2.68%	CEA, 0.5 pg/mL NSE, 1.0 pg/mL SCC, 1.0 pg/mL	CEA, 0.17 pg/mL NSE, 0.33 pg/mL SCC, 0.33 pg/mL
Diouf et al., 2020 [55]	n/a	Healthy (n/a)			Nitrite ion	Electrochemical Impedance Spectroscopy and Differential Pulse Voltammetry	Reproducibility only checked for 4 replicates with RSD of 4%	0.5–50 μg/mL	4 μmol/L
Karimzadeh et al., 2022 [61]	n/a	Healthy (n/a) (spiked with daclatasvir)			Daclatasvir	Colorimetric change from pink to blue LSPR using UV–vis spectrophotometer	3 replicates at different DAC concentrations. At DAC of 1 μg/mL, recovery of 99.27%, and RSD of 0.992 ± 0.018	0.01–1.0 µg/mL	0.008 µg/mL

* n/a indicates that the item has not been assigned.

Table 2. Gold nanoparticles.

(a) EBC Extraction Methodology and Nanosensor-Based Analysis
EBC Methodology						Nanosensor
Refs.	Condenser		Extraction Temp/Time	Vol.	Storage Temp/ Time	Biomarker/Drug	Outcomes
Peng et al., 2009 [34]	Mouthpiece with filter (EcoMedicDuerten)		Room temp/ 5 min	n/a *	Air stored in Mylar bags; analyzed within 2 days of collection	Lung cancer VOCs	Proper functionalization of AuNPs with organic moieties Pattern analysis of VOCs showed well-confined and non-overlapping clusters of VOCs between healthy and lung cancer patients Air analyte detected without pre-treatment or dehumidification Electrical resistance to analytes with low detection limits
Peng et al., 2010 [48]	Mouthpiece with filter (EcoMedic Duerten)		Room temp/ 3–5 min	n/a	Air stored in Mylar bags; analyzed within 4 days of collection	Endogenous cancer VOCs	Air trapped in bag directly sensed by GC-MS or by electrical resistance of GNP sensors connected to a computer detection system Cluster and pattern analysis of VOC cancer subtypes is plotted in comparison to healthy samples Nanosensors are insensitive to cofounding factors (gender, age, stress, food additives, drug exposure, work pollution)
Zou et al., 2014 [50]	Commercial EBC collection device (EcoScreen (Jaeger)		−10 °C/ 10 min	n/a	Immediate use	CEA, NSE, and SCC	Small 50 μL EBC sample required Rapid detection, low cost, and simple procedure Higher levels of CEA, NSE, and SCC detection in EBCs of patients than in EBCs of controls Only the concentration of CEA showed significant differences between control and case groups
Zhang et al., 2015 [49]	Commercial EBC collection device (EcoScreen (Jaeger)		−10 °C/ 10 min	1–3 mL	Immediate use	CEA	Significant detection data correlation between the immunosensor and a commercial instrument (coefficient = 0.999) Rapid detection through real-time measurement (in minutes) High specificity and insignificant potential interference from other biomarkers (neuron-specific enolase (NSE) and squamous cell carcinoma (SCC))
Chen et al., 2019 [59]	Commercial disposable EBC collection system (Ecoscreen Jaeger)		n/a	Diluted to 900 μL with ultrapure water	−20 °C n/a	Polyamines	Proven optimal system performance in 1 mL of EBC, 87 nM AuNPs concentration, a 20 min incubation period, 0.10 M 2-ME, and 10 μL of anhydrous methanol dispersant Low extraction time and sample consumption by AuNP-ME/FASS-CE-C4D method Successful polyamine separation from potential coexisting substances and direct determination without derivatization
Ding et al., 2020 [52]	Frozen pipe mouthpiece		n/a	30 µL	Freshly used	Nitrite ions	Sensitive sensor to micromolar nitrite levels Selective sensor to nitrites, without signals to nitrate, acetate, or ammonium nitrate
Ding et al., 2022 [51]	Condensate device		Room temp	0.5 mL	EBC centrifuged at 5000 rpm and supernatant freshly used	CEA	High sensitivity: superior detection level in pg/mL compared to ELISA in ng/mL Energy transfer to Ru(bpy)32+ was amplified with Au NPs High selectivity and stability Au@SiO2-Ru showed an almost 20 nm NP size with a 5 nm silicon corona shell Increasing CEA concentrations generate closer aptamer-to-aptamer binding; better energy transfer from Au-g-C3N4 to Au@SiO2-Ru
Diouf et al., 2020 [55]	Condensate device		Room temp/ n/a	0.5 mL	EBC centrifuged at 5000 rpm and supernatant freshly used	CEA, NSE, and SCC	Novel design of an electrode array (4 × 4) to check 3 cancer biomarkers High sensitivity and selectivity to the tested biomarkers as compared with other non-target proteins Comparable values using the sensor and ELISA tests
Karimzadeh et al., 2022 [61]	Lab-made setup		Room temp/ n/a	200 µL	n/a	Daclatasvir drug	DAC interacts with the oxygen and hydroxyl groups of sucrose, causing aggregation and color change from pink to blue Increase in DAC lowers LSPR peaks The sensor detects DAC both in EBC and plasma samples with comparable recoveries Readings should be made in <5 min
(b) Design and Performance Parameters of AuNPs Nanosensors
Refs.	Age	Groups (N)/Health Status			Biomarkers/Drugs	Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Peng et al., 2009 [34]	28–60	>Healthy (56) >Lung cancer (40)			Endogenous cancer VOCs	Electrical resistance signals detected by computer system	>90% reproducibility for readings on different days	n/a	1 to 5 ppb
Peng et al., 2010 [48]	20–75	>Lung cancer (30) Colon cancer (26) >Breast cancer (22) >Healthy (22–59)			Endogenous cancer VOCs	Electrical resistance signals detected by computer system	>87% reproducibility for a specific volunteer who was examined multiple times over a period of 6 months	300–400 different VOCs/ breath	n/a
Zou et al., 2014 [50]	n/a	>Healthy (13) (spiked with CEA or NSE) >Lung cancer (17)			CEA, neuron-specific enolase, and SCC	Love Wave sensor; surface acoustic wave (SAW) immunosensor	10 replicates for each biomarker RSD (n/a) used to calculate precision and accuracy	n/a	CEA, 0.967 ng/mL NSE, 1.598 ng/mL SCC, 0.663 ng/mL
Zhang et al., 2015 [49]	n/a	Healthy (28)			CEA	Love Wave sensor	3 repetitions performed to confirm linearity	1–16 ng/mL	1 ng/mL
Chen et al., 2019 [59]	n/a	>Healthy (12) (spiked with polyamines)			Polyamines	Field-amplified sample stacking (FASS) coupled with capillary electrophoresis and FASS-CEC4D	RSDs for intraday reproducibility: 1.0–1.8% and 1.1–2.7% (n = 7), respectively; RSDs for inter-day reproducibility: within 4.7% (n = 5)	0.20–20 ng/mL	0.070–0.17 ng/mL
Ding et al., 2020 [52]	n/a	Healthy (8)			Carcinoembryonic antigen (CEA)	SEECL with resonance energy transfer [66] recorded by CV using ultra-weak luminescence	10 consecutive repetitions without CEA with RSD of 1.67%; 6 different sensors tested at different time points with CEA; RSD 2.14%	1.0 pg/mL–5.0 ng/mL	0.3 pg/mL
Ding et al., 2022 [51]	n/a	Healthy (n/a)			CEA, neuron-specific enolase, and SCC	SEECL-I compared with ELISA results	6 batches tested with 1 ng/mL of antigens with RSD for CEA 3.02%, NSE 2.91%, and SCC 2.68%	CEA, 0.5 pg/mL NSE, 1.0 pg/mL SCC, 1.0 pg/mL	CEA, 0.17 pg/mL NSE, 0.33 pg/mL SCC, 0.33 pg/mL
Diouf et al., 2020 [55]	n/a	Healthy (n/a)			Nitrite ion	Electrochemical Impedance Spectroscopy and Differential Pulse Voltammetry	Reproducibility only checked for 4 replicates with RSD of 4%	0.5–50 μg/mL	4 μmol/L
Karimzadeh et al., 2022 [61]	n/a	Healthy (n/a) (spiked with daclatasvir)			Daclatasvir	Colorimetric change from pink to blue LSPR using UV–vis spectrophotometer	3 replicates at different DAC concentrations. At DAC of 1 μg/mL, recovery of 99.27%, and RSD of 0.992 ± 0.018	0.01–1.0 µg/mL	0.008 µg/mL

* n/a indicates that the item has not been assigned.

3.3. Silver Nanoparticles

The unique silver properties, especially silver nanoparticles (AgNPs), allow their utilization in numerous analytical techniques, such as for catalysis and photocatalysis, bioimaging and biosensing, and antibacterial and anti-cancer therapies. However, AgNPs are very fragile. Any slight alteration in their microenvironment from the reducing agent to the stabilizer might directly impact their performance, stability, and toxicity [67].

AgNPs were successfully employed as nanosensors for drug detection in EBC (Table 3a,b). AgNPs’ effect was tested on the fluorescence emission of the terbium ion–deferiprone drug (Tb³⁺–DEF) in EBC. Samples were collected from sixteen individuals using a lab-made setup at −20 °C, with a maximal sample storage time of 3 weeks at −20 °C. Also, precaution steps, including mouth washing and nasal obstruction by clips, were followed to obtain clean EBC samples. AgNP surface plasmon electrons increased the fluorescent intensity of the Tb³⁺–DEF complex by 2-fold. Parameters such as AgNP and Tb³⁺ concentrations, pH, buffer concentrations, and EBC storage were calibrated using DEF-spiked EBC samples to ensure the highest DEF detection potential. Moreover, an LR was obtained between 0.06 and 1.50 μg/mL DEF concentrations with successful DEF quantification in sixteen EBC samples at very low concentrations (0.06–0.17 μg/mL) [68].

The remaining studies employing AgNPs for EBC drug detection relied solely on spectrophotometric analysis while granting the AgNPs specificity toward EBC biomolecules through capping with amidosulfonic acid (ASA) [69], doxorubicin (DOX) [70], indoxyl sulfate (InS) [71], or SDS [72,73]. ASA-AgNPs were used for lamotrigine (LTG) drug detection in human EBC in a dose-dependent manner. This reaction was read using simple UV-Vis measurements as an absorption peak appearing at longer wavelengths. After the optimization and calibration of the reaction in terms of ASA/ASA-AgNPs concentrations, pH, and reaction time, a linear relation was obtained between the NP absorbance intensity and LTG concentration in the range of 0.02–0.4 μg/mL, with a 5 ng/mL LOD. Moreover, the LTG measurements in the EBC samples of three LTG-treated epileptic individuals showed a significant detection of the drug in comparison to traditional HPLC measurements. In addition, no significant interference of co-administered drugs with this method was recorded, except for phenytoin sodium. Similar physicochemical changes were obtained for indoxyl sulfate-capped silver nanoparticles (InS-AgNPs) upon its employment for the detection of the phenytoin (PHT) drug in EBC. Following these parameters and InS-AgNP-PHT reaction optimization, the PHT sensitivity of this nanoprobe was proved unaffected by co-existing drug administration. Also, promising PHT detection by InS-AgNPs in two PHT-spiked EBC samples was reported. However, InS-AgNPs did not present additional PHT quantification properties. Hence, to further validate the efficiency of this nanoprobe, it is recommended to increase the pool of studied EBC samples and test it on the EBC of treated patients [71]. A simpler and sensitive method relied on the redox reaction between [Ag (NH₃)²⁺] and DOX to generate AgNPs, portraying increased surface plasmon resonance (SPR) intensity proportionally to DOX concentrations. The reaction parameters were optimized, and AgNPs effectively determined the concentration of DOX in DOX-spiked EBC samples with acceptable results and a 4.16 ng/mL LOD. However, the investigation of co-existing compound interference with AgNPs in EBC and testing on EBC samples collected from real DOX-treated patients were missing [70]. Further NP forms introduced SDS-AgNPs for tobramycin antibiotic detection in EBC while relying on ASA-AgNP aggregation with an SPR shift in the presence of tobramycin and sodium metaborate. After testing on EBCs withdrawn for 16 min from six volunteers exposed to a 150 mg dose of tobramycin, SDS-AgNP sensitivity was accurate across all samples and showed a significant recovery rate after external spike control. The LOD and LR of tobramycin were at nano levels, which are comparable to results from other analytical methods derived from urine, plasma, or serum biopsies. The sensitivity was achieved without a pre-concentration, usually performed in solid- or liquid-phase extraction methods and in a fast response manner. Also, good reproducibility from batch to batch during the same day and on different days was reported. However, certain limitations should be tackled, including the shelf stability of the nanoprobes and variability in EBC sampling [73]. Similarly, SDS-AgNPs were employed as catalyzers in a 3,3, 5,5–tetramethyl benzidine–hydrogen peroxide (TMB-H₂O₂) system for the determination of the aspirin drug in EBC. Further elaborated, SDS-AgNPs degraded H₂O₂ to OH radicals responsible for converting the non-colored TMB to blue-colored, oxidized TMB. Since aspirin is prone to oxidation prior to TMB, the higher the aspirin concentration, the lower the obtained UV-Vis spectra signal of the oxidized TMB. Interestingly, drug co-administration in the therapeutic safe doses did interfere with SDS-AgNP sensitivity toward aspirin in aspirin-spiked EBC samples. Moreover, this system presented a linear relationship with aspirin concentration in a 10–250 μg/mL range with a 4.1 μg/mL LOD. The successful aspirin detection in the real EBC samples of six aspirin-treated patients further confirmed the feasibility of this approach [72].

For biomarker detection, AgNPs successfully detected malondialdehyde [74,75] once deposited onto a poly-dopamine–chitosan (poly (DA)-CS) composite. AgNPs enhanced the electrochemical reaction between the analyte, MDA, and its interface. Moreover, AgNPs increased the surface area and aided in the chemical stability and adhesion properties of the analyte. The stability of the sensor was also mediated via the strong electrostatic interaction between the positively charged chitosan and negatively charged dopamine that aided in the dispersion of AgNPs over a wide range of pH from 4.5 to 11. Based on the electrochemical oxidation status from human EBC samples with MDA addition, the platform seems stable and can be reused for up to 100 replicas. Conductivity was more significant in platforms with AgNPs compared to nano-devoid electrodes. Moreover, the accuracy of read-outs, sensitivity, and reproducibility proved effective using their design model. Compared to other methods (HPLC, MS, surface-enhanced Raman spectroscopy) collected from various biopsies (urine, plasma, serum), the LOD performance of this EC nanosensor is 0.817 μM and outperforms the EC from serum samples not yet reaching low pico- or nano-molar levels from urine and plasma using high-performance analytic techniques. Following sensor evaluation on pre-prepared MDA solutions, RF-PT-AgNPs showed a 0.089–1.68 mM linear dynamic range and a low limit of quantification of 0.59 μM. Moreover, this nanosensor achieved efficient MDA detection in real EBC samples. Nevertheless, the sensor was not tested against the interference of other products with RF-PT-AgNPs, which reduces the validity of the reported data.

Table 3. Studies on silver nanoparticles with EBC methodology and nanosensor outcomes.

(a) Extraction Method and Nanosensor Detection Outcomes
EBC Methodology					Nanosensor
Refs.	Condenser	Extraction Temp/Time	Vol	Storage Temp/Time	Biomarker/Drug	Outcomes
Mohamadian et al., 2017 [68]	Lab-made setup	−20 °C/10 min	n/a *	−20 °C/ 3 weeks	Deferiprone (DEF) drug	Enhanced fluorescence emission signal of the Tb3+–DEF complex by AgNP complex (more than 2-fold) under the same conditions Optimal system performance in Tris-HCL buffer at a pH of 9.5, 11.43.10−4 M of Tb3+, 25 min reaction time at 0 °C, and 90 µL of (2.33 × 10−4 M) AgNPs Insignificant interference of co-existing drugs on the fluorescence intensity and selectivity of Tb3+–DEF–AgNPs system Successful DEF quantification in patient’s EBC of very low concentrations (0.06–0.17 μg/mL)
Jouyban et al., 2017 [69]	Lab-made setup	n/a	1 mL	Freshly used	Lamotrigine (LTG) drug	Proven optimal ASA-AgNPs performance in a 1:1 molar ratio of (nAgNO3: nASA), 17.5 × 10−10 M of ASA-AgNPs, a pH of 8.5, and 30 min incubation period Insignificant interference of co-existing drugs with the ASA-AgNPs system, except for phenytoin sodium Significant detection of LGT by ASA-AgNPs in comparison with HPLC measurement method
Jouyban et al., 2019 [70]	Lab-made setup	n/a	n/a	n/a	Doxorubicin (DOX) drug	Proven optimal AgNP performance in 1.0 × 10−3 mol/L of NaOH, 1.06 × 10−3 mol/L of ammonia, no added stabilizing agent, 1.0 × 10−4 mol/L of silver nitrate, 70 °C, and 10 min incubation time Insignificant interference of anions/cations and organic/inorganic compounds with AgNPs performance Highly sensitive determination of DOX in DOX-spiked EBC No testing available on EBC from DOX-treated patients
Hasanzadeh et al., 2018 [74]	Lab-made pipe with a cooling mouthpiece	Room temp	1 mL	−25 °C	Malondialdehyde	AgNPs increase the surface area and aid in the chemical stability and adhesion properties of the analyte AgNPs increase conductivity Proven sensitivity and reproducibility compared to plasma levels
Khoubnasab et al., 2019 [71]	Lab-made system	n/a	n/a	n/a	Phenytoin drug	Proven optimal InS-AgNP performance in 100 μL of Britton–Robinson buffer, a pH of 7.0, 200 μL of AgNPs, and a 35 min incubation time Insignificant interference of co-existing drugs on PHT detection by InS-AgNPs Successful PHT detection by InS-AgNPs without PHT quantification
Jafari et al., 2019 [75]	Lab-made setup	−20 °C/n/a	1 mL	n/a	Malondialdehyde	Suitable performance of electrode for repeatability of the analysis, at least for 50 cycles Increased contact area with MDA Successful MDA detection in EBC samples Missing interference analysis reports
Abachi et al., 2022 [72]	Lab-made system	n/a	n/a	Dark/ 4 °C	Aspirin drug	Proven optimal system performance at a pH of 5, 0.12 mmol/L of TMB, 0.30% v/v H2O2, 3 min reaction time, and 12.5 µL of SDS-AgNPs Adequate selectivity for aspirin analysis with other co-administered drug interference
Rezaei et al., 2020 [73]	Lab-made system	Room temp	0.25 mL	n/a	Tobramycin, gentamycin, and amikacin antibiotics	Highly sensitive with detection levels 4 to 6 times greater for gentamycin and tobramycin
(b) Type and Performance Properties of Silver NPs
Refs.	Age	Groups (N)/ Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Mohamadian et al., 2017 [68]	15–24 years	>Healthy (1) (used for calibration after spiking with deferiprone) >DEF-treated (16)			JASCO FP-750 spectrofluorometer	RSD inter-day analysis: 2.65–5.10% RSD intra-day: 3.68–6.21%	0.06–1.50 μg/mL	0.06 μg/mL
Jouyban et al., 2017 [69]	n/a	>Healthy (n/a) (used for calibration after spiking with lamotrigine) >Epileptic LTG-treated (3)			UV-2550 spectrophotometer	RSD of 2.37% for five replicates for 0.1 μg·mL−1 of LTG ASA-AgNPs	0.02–0.4 µg/mL	0.005 µg/mL
Jouyban et al., 2019 [70]	n/a	>Healthy (n/a) (used for calibration after spiking with doxorubicin) >Healthy (4) (DOX-spiked)			Shimadzu UV–visible double-beam UV-1800	RSD of 2.7% (n = 3)	0.02–2 µg/mL	0.004 µg/mL
Hasanzadeh et al., 2018 [74]	n/a	Healthy (4)			Electrochemical technique: square-wave voltammetry (SWV)	Highly stable for up to 100 replicates with an error of 2.1%	0.817–3.15 µmol/L	0.817 μmol/L
Khoubnasab et al., 2019 [71]	n/a	>Healthy (n/a) (used for calibration after spiking with phenytoin) >Healthy (2)			UV-2550 spectrophotometer	RSD of 3.95% for five replicated determinations of a 100 μg·L−1 PHT InS-AgNP	0.025–0.45 μg/mL	0.01 μg/mL
Jafari et al., 2019 [75]	n/a	Healthy (n/a)			Electrochemical sensor	n/a	0.089–1.68 mM	0.59 μmol/L
Abachi et al., 2022 [72]	n/a	>Healthy (n/a) (used for calibration after spiking with aspirin) >Aspirin-treated patients (6)			Double-beam spectrophotometer UV-1800	Intra-day and inter-day RSD for five replications: 1.0 % and 3.5 % batch-to-batch reproducibility with RSD = 5.2% (n = 3)	10–250 μg/mL	4.1 μg/mL
Rezaei et al., 2020 [73]	n/a	>Healthy (n/a) (calibration spiking with tobramycin, gentamycin, or amikacin) >Healthy (6)			UV-vis spectrophotometer	Good reproducibility from batch to batch during the same day and on different days with an RSD < 4.2%	1.0–50.0 ng/mL	0.5 ng/mL

* n/a indicates that the item has not been assigned.

Table 3. Studies on silver nanoparticles with EBC methodology and nanosensor outcomes.

(a) Extraction Method and Nanosensor Detection Outcomes
EBC Methodology					Nanosensor
Refs.	Condenser	Extraction Temp/Time	Vol	Storage Temp/Time	Biomarker/Drug	Outcomes
Mohamadian et al., 2017 [68]	Lab-made setup	−20 °C/10 min	n/a *	−20 °C/ 3 weeks	Deferiprone (DEF) drug	Enhanced fluorescence emission signal of the Tb3+–DEF complex by AgNP complex (more than 2-fold) under the same conditions Optimal system performance in Tris-HCL buffer at a pH of 9.5, 11.43.10−4 M of Tb3+, 25 min reaction time at 0 °C, and 90 µL of (2.33 × 10−4 M) AgNPs Insignificant interference of co-existing drugs on the fluorescence intensity and selectivity of Tb3+–DEF–AgNPs system Successful DEF quantification in patient’s EBC of very low concentrations (0.06–0.17 μg/mL)
Jouyban et al., 2017 [69]	Lab-made setup	n/a	1 mL	Freshly used	Lamotrigine (LTG) drug	Proven optimal ASA-AgNPs performance in a 1:1 molar ratio of (nAgNO3: nASA), 17.5 × 10−10 M of ASA-AgNPs, a pH of 8.5, and 30 min incubation period Insignificant interference of co-existing drugs with the ASA-AgNPs system, except for phenytoin sodium Significant detection of LGT by ASA-AgNPs in comparison with HPLC measurement method
Jouyban et al., 2019 [70]	Lab-made setup	n/a	n/a	n/a	Doxorubicin (DOX) drug	Proven optimal AgNP performance in 1.0 × 10−3 mol/L of NaOH, 1.06 × 10−3 mol/L of ammonia, no added stabilizing agent, 1.0 × 10−4 mol/L of silver nitrate, 70 °C, and 10 min incubation time Insignificant interference of anions/cations and organic/inorganic compounds with AgNPs performance Highly sensitive determination of DOX in DOX-spiked EBC No testing available on EBC from DOX-treated patients
Hasanzadeh et al., 2018 [74]	Lab-made pipe with a cooling mouthpiece	Room temp	1 mL	−25 °C	Malondialdehyde	AgNPs increase the surface area and aid in the chemical stability and adhesion properties of the analyte AgNPs increase conductivity Proven sensitivity and reproducibility compared to plasma levels
Khoubnasab et al., 2019 [71]	Lab-made system	n/a	n/a	n/a	Phenytoin drug	Proven optimal InS-AgNP performance in 100 μL of Britton–Robinson buffer, a pH of 7.0, 200 μL of AgNPs, and a 35 min incubation time Insignificant interference of co-existing drugs on PHT detection by InS-AgNPs Successful PHT detection by InS-AgNPs without PHT quantification
Jafari et al., 2019 [75]	Lab-made setup	−20 °C/n/a	1 mL	n/a	Malondialdehyde	Suitable performance of electrode for repeatability of the analysis, at least for 50 cycles Increased contact area with MDA Successful MDA detection in EBC samples Missing interference analysis reports
Abachi et al., 2022 [72]	Lab-made system	n/a	n/a	Dark/ 4 °C	Aspirin drug	Proven optimal system performance at a pH of 5, 0.12 mmol/L of TMB, 0.30% v/v H2O2, 3 min reaction time, and 12.5 µL of SDS-AgNPs Adequate selectivity for aspirin analysis with other co-administered drug interference
Rezaei et al., 2020 [73]	Lab-made system	Room temp	0.25 mL	n/a	Tobramycin, gentamycin, and amikacin antibiotics	Highly sensitive with detection levels 4 to 6 times greater for gentamycin and tobramycin
(b) Type and Performance Properties of Silver NPs
Refs.	Age	Groups (N)/ Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Mohamadian et al., 2017 [68]	15–24 years	>Healthy (1) (used for calibration after spiking with deferiprone) >DEF-treated (16)			JASCO FP-750 spectrofluorometer	RSD inter-day analysis: 2.65–5.10% RSD intra-day: 3.68–6.21%	0.06–1.50 μg/mL	0.06 μg/mL
Jouyban et al., 2017 [69]	n/a	>Healthy (n/a) (used for calibration after spiking with lamotrigine) >Epileptic LTG-treated (3)			UV-2550 spectrophotometer	RSD of 2.37% for five replicates for 0.1 μg·mL−1 of LTG ASA-AgNPs	0.02–0.4 µg/mL	0.005 µg/mL
Jouyban et al., 2019 [70]	n/a	>Healthy (n/a) (used for calibration after spiking with doxorubicin) >Healthy (4) (DOX-spiked)			Shimadzu UV–visible double-beam UV-1800	RSD of 2.7% (n = 3)	0.02–2 µg/mL	0.004 µg/mL
Hasanzadeh et al., 2018 [74]	n/a	Healthy (4)			Electrochemical technique: square-wave voltammetry (SWV)	Highly stable for up to 100 replicates with an error of 2.1%	0.817–3.15 µmol/L	0.817 μmol/L
Khoubnasab et al., 2019 [71]	n/a	>Healthy (n/a) (used for calibration after spiking with phenytoin) >Healthy (2)			UV-2550 spectrophotometer	RSD of 3.95% for five replicated determinations of a 100 μg·L−1 PHT InS-AgNP	0.025–0.45 μg/mL	0.01 μg/mL
Jafari et al., 2019 [75]	n/a	Healthy (n/a)			Electrochemical sensor	n/a	0.089–1.68 mM	0.59 μmol/L
Abachi et al., 2022 [72]	n/a	>Healthy (n/a) (used for calibration after spiking with aspirin) >Aspirin-treated patients (6)			Double-beam spectrophotometer UV-1800	Intra-day and inter-day RSD for five replications: 1.0 % and 3.5 % batch-to-batch reproducibility with RSD = 5.2% (n = 3)	10–250 μg/mL	4.1 μg/mL
Rezaei et al., 2020 [73]	n/a	>Healthy (n/a) (calibration spiking with tobramycin, gentamycin, or amikacin) >Healthy (6)			UV-vis spectrophotometer	Good reproducibility from batch to batch during the same day and on different days with an RSD < 4.2%	1.0–50.0 ng/mL	0.5 ng/mL

* n/a indicates that the item has not been assigned.

3.4. Copper Nanoparticles

Copper nanoparticles (CuNPs) possess antifungal, antimicrobial, and anti-cancer properties. Compared to other NPs, CuNPs possess attractive properties such as high thermal and electrical conductivity for the build-up of nanosensors. Furthermore, the size and surface characteristics of CuNPs can regulate the release of copper ions and overall stability [76]. In EBC detection, CuNPs were used in different forms for H₂O₂ and drug detection (Table 4a,b). Fluorometric copper oxide nanoparticles (CuONPs) for H₂O₂ detection in EBC relied on catalyzed 5-aminosalicylic acid/hydrogen peroxide (5 ASA/H₂O₂) systems. CuONPs exhibited horseradish peroxidase (HRP)-like properties by catalyzing and breaking the O-O bond of H₂O₂ to release OH-, which in turn reacted with 5-ASA to form a non-fluorescent compound (5-ASAox). Subsequently, a 50–500 nmol/L LR and 33.6 nmol/L LOD were presented by CuONP-catalyzed 5-ASA/H₂O₂ assay after optimization using H₂O₂-spiked EBC samples. The CuONP-based sensor successfully detected H₂O₂ in unmanipulated EBC samples of four healthy subjects and was not subjected to significant interference by co-administered drugs from the EBCs of sick patients [77]. Interestingly, the electro-deposition of CuNPs onto a PANI film-modified nickel foam electrode (PANI/NFE) drastically enhanced the obtained current with an LR for H₂O₂ of 0.01–500 nM and an LOD of 0.0026 nM. Also, interfering substances such as CO₂, N₂, NO, NH₃, and acetone did not affect the electrode performance even at a 10-fold concentration of H₂O₂ (0.01 μM). Moreover, comparable concentrations of H₂O₂ were obtained by this electrode sensor and HPLC, suggesting that the Cu/PANI/NFE sensor is reliable for field testing. Finally, the Cu/PANI/NFE sensor can be used to detect H₂O₂ levels in the EBCs of smokers compared with non-smokers [63].

In another study by Hatefi et al., fluorescently modified copper nanoclusters coated with cetyl trimethylammonium bromide (Cu-NC-CTAB) successfully detected the carbamazepine (CBZ) drug in EBC. The CTAB coat captured CBZ and blocked the non-radiative e⁻/h⁺ recombination defect sites on the NCs, hence enhancing the NC fluorescence intensity. After routine optimization of the parameters of this nanoprobe using CBZ-spiked EBC, an LR between 0.2 and 20 μg/mL CBZ concentrations was obtained with a 0.08 μg/mL detection limit. Moreover, the sensitivity of Cu-NC-CTAB was not subjected to interference by anions/cations or over-the-counter drugs administered in safe doses. Also, Cu-NC-CTAB successfully determined CBZ in four EBC samples of CBZ-treated patients, once compared with HPLC. Following the same detection mechanism, a copper nanocluster (Cu-NC)-based sensor was designed to detect the vancomycin antibiotic in EBC with a 0.1–8 μg/mL LR [78]. However, like most sensors, the Cu-NC-based SFS model should be tested on a larger pool of subjects and possibly on EBC collected by tidal breathing to further confirm its vancomycin detection efficiency [79]. The latest model of Cu-NPs is designed with copper-doped graphene quantum dots (Cu-doped GQDs) to detect carbamazepine through fluorescence quenching, enabling the monitoring of drug-dose uptake in epilepsy patients. Detection occurred within a concentration range of 0.02–2.0 µg/mL, primarily driven by dynamic quenching, as indicated by the Stern–Volmer plot. The method demonstrated high precision, with an RSD of 1.8% for intra-day and 4.8% for inter-day analysis [80].

Table 4. Copper Nanoparticles.

(a) EBC Methodology and Nanosensor Outcomes
EBC methodology					Nanosensor
Refs.	Condenser	Extraction Temp/Time	Vol	Storage Temp/Time	Biomarker/Drug	Outcomes
Jouyban et al., 2017 [77]	Lab-made setup	n/a * 20 min	2 mL	−20 °C and of RT/12 h, 24 h, and 36 h	H2O2	Proven peroxidase-like activity of CuONPs Insignificant interference of inorganic ions, organic compounds, and co-existing drugs on the fluorescence intensity of the CuONP-catalyzed 5-ASA/H2O2 assay Comparable performance with HRP-catalyzed 5-ASA/H2O2 assay Most stable method for EBC sample storage at −20 °C for more than 36 h Successful detection of H2O2 in EBC samples by CuONP-catalyzed 5-ASA/H2O2 assay
Liu et al., 2018 [63]	Teflon bags (3 L and absorption water of 5 mL)	n/a	n/a	n/a	H2O2	Optimal reaction conditions at a pH of 5.5, applied potential of −0.6 V, and at room temperature (25 °C) Short runtime (minutes) Insignificant interference of other substances with electrode function Comparable performance with HPLC Higher H2O2 levels detection in EBCs of smokers compared to non-smokers
Hatefi et al., 2019 [79]	Lab-made system	n/a	n/a	Freshly used	CBZ	Optimal Cu-NC-CTAB performance at a pH of 6.0, 16.5 mmol/L of PBS, 0.05 μg/mL of Cu-NCs, and for 10 min Insignificant interference of anions/cations, co-existing and over-the-counter drugs on the fluorescence intensity of the Cu-NC-CTAB nanoprobe Accurate and great potential for the determination of CBZ in EBC samples by Cu-NC-CTAB compared with HPLC
Rahimpour et al., 2021 [78]	MVEBC samples collected from the waste of the ventilator	RT n/a	n/a	Freshly used	Vancomycin antibiotic	Acceptable selectivity for vancomycin determination in the presence of co-administered and/or non-prescribed over-the-counter drugs. Optimal parameters with maximum signal intensity at a pH of 6, 0.05 mL of Cu-NCs, 0.01 mol/L of phosphate buffer, and a 5 min incubation time Successful determination of vancomycin in EBCs of newborns receiving vancomycin treatment (average level of 0.56 μg/mL).
(b) Classification of CuNPs and Associated Detection Characteristics
Refs.	Age	Groups (N)/ Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Jouyban et al., 2017 [77]	n/a	>Healthy (n/a) (used for calibration after spiking with hydrogen peroxide) >Healthy (4)			FP-750 spectrofluorometer	Inter-day RSD of 4.2–6.9% and intra-day RSD of 2.4–4.9% for replicates at three levels of H2O2	50–500 nmol/L	33.6 nmol/L
Liu et al., 2018 [63]	n/a	>Healthy before/after smoking (12)			Electrochemical sensor	RSD of 4.2% (C = 0.1 μM, n = 7)	0.01–500 nmol/L	0.0026 nmol/L
Hatefi et al., 2019 [79]	n/a	>Healthy (n/a) (used for calibration after spiking with carbamazepine) >CBZ-treated patients (4)			FP-750 spectrofluorometer	Inter-day RSD of 4.8 and intra-day RSD of 3.9% for six replicas; 1 μg/mL of CBZ Batch-to-batch RSD of 5.2% Cu-NC-CTAB stored in a dark place at 4 °C	0.2–20 μg/mL	8 μg/mL
Rahimpour et al., 2021 [78]	Newborns or premature babies	>Under mechanical ventilator/MV (n/a) (used for calibration after spiking with vancomycin) >Under mechanical ventilator/MV and treated with vancomycin (5)			FP-750 spectrofluorometer	Inter-day RSD of 4.8 and intra-day RSD of 6.0% for six replicas; 1 μg/mL of vancomycin Cu-NCs stored in a dark place at 4 °C	0.1–8 µg/mL	0.06 µg/mL

* n/a indicates that the item has not been assigned.

Table 4. Copper Nanoparticles.

(a) EBC Methodology and Nanosensor Outcomes
EBC methodology					Nanosensor
Refs.	Condenser	Extraction Temp/Time	Vol	Storage Temp/Time	Biomarker/Drug	Outcomes
Jouyban et al., 2017 [77]	Lab-made setup	n/a * 20 min	2 mL	−20 °C and of RT/12 h, 24 h, and 36 h	H2O2	Proven peroxidase-like activity of CuONPs Insignificant interference of inorganic ions, organic compounds, and co-existing drugs on the fluorescence intensity of the CuONP-catalyzed 5-ASA/H2O2 assay Comparable performance with HRP-catalyzed 5-ASA/H2O2 assay Most stable method for EBC sample storage at −20 °C for more than 36 h Successful detection of H2O2 in EBC samples by CuONP-catalyzed 5-ASA/H2O2 assay
Liu et al., 2018 [63]	Teflon bags (3 L and absorption water of 5 mL)	n/a	n/a	n/a	H2O2	Optimal reaction conditions at a pH of 5.5, applied potential of −0.6 V, and at room temperature (25 °C) Short runtime (minutes) Insignificant interference of other substances with electrode function Comparable performance with HPLC Higher H2O2 levels detection in EBCs of smokers compared to non-smokers
Hatefi et al., 2019 [79]	Lab-made system	n/a	n/a	Freshly used	CBZ	Optimal Cu-NC-CTAB performance at a pH of 6.0, 16.5 mmol/L of PBS, 0.05 μg/mL of Cu-NCs, and for 10 min Insignificant interference of anions/cations, co-existing and over-the-counter drugs on the fluorescence intensity of the Cu-NC-CTAB nanoprobe Accurate and great potential for the determination of CBZ in EBC samples by Cu-NC-CTAB compared with HPLC
Rahimpour et al., 2021 [78]	MVEBC samples collected from the waste of the ventilator	RT n/a	n/a	Freshly used	Vancomycin antibiotic	Acceptable selectivity for vancomycin determination in the presence of co-administered and/or non-prescribed over-the-counter drugs. Optimal parameters with maximum signal intensity at a pH of 6, 0.05 mL of Cu-NCs, 0.01 mol/L of phosphate buffer, and a 5 min incubation time Successful determination of vancomycin in EBCs of newborns receiving vancomycin treatment (average level of 0.56 μg/mL).
(b) Classification of CuNPs and Associated Detection Characteristics
Refs.	Age	Groups (N)/ Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Jouyban et al., 2017 [77]	n/a	>Healthy (n/a) (used for calibration after spiking with hydrogen peroxide) >Healthy (4)			FP-750 spectrofluorometer	Inter-day RSD of 4.2–6.9% and intra-day RSD of 2.4–4.9% for replicates at three levels of H2O2	50–500 nmol/L	33.6 nmol/L
Liu et al., 2018 [63]	n/a	>Healthy before/after smoking (12)			Electrochemical sensor	RSD of 4.2% (C = 0.1 μM, n = 7)	0.01–500 nmol/L	0.0026 nmol/L
Hatefi et al., 2019 [79]	n/a	>Healthy (n/a) (used for calibration after spiking with carbamazepine) >CBZ-treated patients (4)			FP-750 spectrofluorometer	Inter-day RSD of 4.8 and intra-day RSD of 3.9% for six replicas; 1 μg/mL of CBZ Batch-to-batch RSD of 5.2% Cu-NC-CTAB stored in a dark place at 4 °C	0.2–20 μg/mL	8 μg/mL
Rahimpour et al., 2021 [78]	Newborns or premature babies	>Under mechanical ventilator/MV (n/a) (used for calibration after spiking with vancomycin) >Under mechanical ventilator/MV and treated with vancomycin (5)			FP-750 spectrofluorometer	Inter-day RSD of 4.8 and intra-day RSD of 6.0% for six replicas; 1 μg/mL of vancomycin Cu-NCs stored in a dark place at 4 °C	0.1–8 µg/mL	0.06 µg/mL

* n/a indicates that the item has not been assigned.

3.5. Quantum Dots, Carbon Nanotubes, Cobalt Oxide, and Polymer Nanoparticles

Nanocomposites of quantum dots (QDs) with carbonaceous compounds, polymers, and cobalt oxides have improved the capacitor outputs of nanosensors with sensitivity to detect drugs, VOCs, and fatty acids in EBC (Table 5a,b).

Mokhtari et al. applied polymers to enhance the spectrofluorometric detection of the phenobarbital drug in EBC using a coordinated luminol–terbium polymer nanoparticle (luminol-Tb CPNPs). Tb acts as a metal ion bridging luminol, the fluorophore, to the substrate and causing aggregation. Subsequently, the fluorescence emission of luminol is proportionally amplified to a phenobarbital concentration. The optimized luminol-Tb CPNP system was not affected by over-the-counter drugs in phenobarbital-spiked samples and showed a linear relationship with the phenobarbital concentration in the range of 0.1–10.0 μg/mL with an LOD of 0.024 μg/mL. Additionally, luminol-Tb CPNPs presented simple and reliable EBC phenobarbital quantification in mechanical ventilator EBCs (MVEBCs) of newborns as compared to HPLC. This probe provided satisfying results yet lacks storage stability [81]. Another nanosensor was based on N-doped carbon dots (N-CDs) modified with molecularly imprinted polymers (MIPs). MIP@ N-CD excitation at 310 nm resulted in a spectrofluorometric emission peak at 417 nm due to the N-CD core. This is due to the conformational change in MIP on MIP@ N-CDs upon phenobarbital binding. This, in turn, quenched the MIP-based fluorescence emission of MIP@ N-CDs upon excitation, which confirmed the phenobarbital presence in the samples. Interestingly, after examining phenobarbital-spiked EBC samples, a linear relationship of a 0.01–8.0 μg/mL range was found between this approach and thephenobarbital concentration, with a 0.006–0.18 μg/mL LOD. Moreover, MIP@ N-CDs portrayed adequate phenobarbital selectivity after analyzing the interference of some co-prescribed and over-the-counter drugs in the determination of 0.10 μg/mL of phenobarbital. Even though promising outcomes were obtained for phenobarbital detection and quantification in the EBC samples of one pre-term infant, increasing the number of tested subjects would further confirm MIP@ N-CD accuracy and reliability for phenobarbital determination in real samples [82].

Cobalt oxide nanoparticles (CO₃O₄ NPs), hematite nanorods (α-Fe₂O₃ nanorods), and nickel oxide nanoparticles (NiO NPs) were used for H₂O₂ detection following a chemiluminescence NP-based EBC assay. The metal ions’ sensitivity and catalytic effects on the luminol–H₂O₂ CL system rendered this NP-based assay with high sensitivity. Interestingly, microarray-based analysis revealed a concentration LR of 1.0 nM–1000 nM and a 0.3 nM LOD for H₂O₂ by CO₃O₄ NPs. The chemiluminescence assay relied on a small sample volume without the need for extended incubation and analytical runtime (<1 min). Moreover, the chemiluminescence readings revealed that Co₃O₄ NPs significantly detected higher H₂O₂ levels in the human EBCs of feverish subjects compared with healthy and rheumatic subjects [62].

Another quantum dot sensor was coated for the first time with amino acids to detect MDA in EBC using polyarginine–graphene (PARG-GQDs) [83]. The graphene QDs increased the sensitivity by a 10-fold difference in the analyte signal with a low detection limit of 0.329 nM. The team explained this fact by the increased electroconductivity and surface area printed by GQDs. Compared to the conventional UV-Vis spectrophotometer, the detection of MDA in EBC was more sensitive with lower values.

Lastly, a chemiresistive single-walled carbon nanotube sensor was functionalized with ionic liquids (ILs) to detect VOCs [84]. The main sensing mechanism is the matrix swelling of the sensor, which leads to a decrease in conductance during the sensing exposure. Human breath was condensed by tidal breathing for 20 min in a custom-made EBC collection system, mainly by breathing into PVC tubing containing a saliva trap and cooled with dry ice to condense the exhaled breath. The collected EBC was either used as untreated (healthy) or as toluene-spiked (diseased). Principal component analysis differentiated between healthy breath, diseased breath, and toluene vapor samples. However, the range of the sensors was higher than the physiological concentration of VOCs in the EBCs. The NSPCl-doped CD nanoprobe detected vancomycin through fluorescence quenching, forming a non-fluorescent complex within a 0.01–2.0 µg/mL range with RSDs of 1.4% (intra-day) and 3.2% (inter-day) [85].

Table 5. Quantum dots/carbon nanotubes/cobalt oxide and polymer NPs.

(a) EBC Extraction Methodology and Detection Outcomes
EBC Methodology					Nanosensor
Refs.	Condenser	Extract Temp/Time	Vol.	Storage Temp/Time	Biomarker/Drug	Outcomes
Li et al., 2013 [62]	Mouthpiece connected to a 50 mL centrifuge tube	n/a *	Diluted to 2 mL with purified water	−20 to −25 °C 1 week	H2O2	High specific catalytic effect of Co3O4 nanoparticles on the luminol H2O2 CL reaction in alkaline medium with a lower concentration of luminol (1 × 10−8 M) No CL intensity interference by the presence of typical metal ions High stability of luminol–H2O2–Co3O4 nanoparticles system Optimal Co3O4 nanoparticles performance in a hydroxide solution at a pH of 13 Small sample and no incubation time with analytical runtime of <1 min Significant detection of higher H2O2 levels in EBCs of feverish subjects by Co3O4 nanoparticles compared with healthy and rheumatic subjects
Hasanzadeh et al., 2017 [83]	Lab-made pipe with a cooling trap mouthpiece	Room temp	2 mL	−25 °C	Malondialdehyde	Biocompatible sensors with low toxicity on cell lines, even at 200 ppm levels of GQDs Cyclic voltammograms recorded values between −1.0 and 1.0 mV in 0.1 M of PBS with higher peaks in the presence of GQD Good long-term stability due to the low solubility of the sensor in water Good sensitivity without any pre-treatment at physiological pH
Park et al., 2018 [84]	Custom-made collection system	0 °C 20 min	n/a	n/a	VOCs	Optimal performance using pastes containing 9 wt % of SWCNT Responses correlate linearly with the concentrations of the VOCs Each VOC generates a different fingerprint pattern from the IL sensor array Long-term stability of SWCNT-ILs and robustness to humidity and air Successful SWCNT-IL differentiation between diseased and healthy breath
Nasehi et al., 2022 [82]	Lab-made setup MVEBC samples collected from the waste of the ventilator	n/a	n/a	n/a	Phenobarbital drug	Optimal parameters for maximum fluorescence are pHs of 8.0, 0.02 mol/L, and 0.02 mL of MIP@ N-CDs and an 8 min incubation time Adequate nanosystem selectivity for the quantification of phenobarbital Successful quantification of phenobarbital in MVEBC of pre-term newborn babies
Mokhtari et al., 2022 [81]	Lab-made setup MVEBC samples collected from the waste of the ventilator	n/a	n/a	Freshly used	Phenobarbital drug	Proven maximum fluorescence at 3 replications, 1.0 μg/mL of 1phenobarbital, a pH of 9, 0.10 μL of phosphate buffer (0.1 mol/L), 40 μL of luminol-Tb CP NPs, and a 10 min incubation time No interference by co-administered drugs with the quantification of phenobarbital Successful detection of phenobarbital in MVEBC of newborns as compared with HPLC
(b) Performance of Quantum Dots/Carbon Nanotubes/Cobalt Oxide and Polymer Nanoparticles Properties
References	Age	Groups (N)/Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Li et al., 2013 [62]	n/a	>Feverish (9) >Rheum (13) >Healthy (13)			Microarray scanner of SynergyTM 2 Multi-Mode Microplate Reader IFFM-A CL analyzer	RSD of 2.1% for 100 nM of H2O2 (n = 11)	0.3 nmol/L	1.0 nmol/L–1000 nmol/L
Hasanzadeh et al., 2017 [83]	n/a	Healthy (n/a)			Electrochemical technique: square-wave voltammetry (SWV)	Simultaneous testing of the electrode with 3 replicates, with RSD of 6.16%.	0.06–0.2 μM	0.329 ng/mL
Park et al., 2018 [84]	n/a	>Healthy (n/a) >Diseased (n/a)			Chemiresistive sensor	n/a	n/a	n/a
Nasehi et al., 2022 [82]	Newborns and premature babies	>Healthy (n/a) >Under mechanical ventilator receiving phenobarbital (1)			Jasco FP-750 spectrofluorometer	Inter-day RSD of 6.1 and 6.9, and intra-day RSD of 1.8 and 1.8% for five replicates	0.01–8.0 μg/mL	0.006–0.18 μg/mL
Mokhtari et al., 2022 [81]	Newborns and premature babies	>Healthy (n/a) >Under mechanical ventilator receiving phenobarbital (3)			FP-750 spectrofluorometer	Inter-day RSD of 3.6 and intra-day RSD of 5.4% Batch-to-batch reproducibility with RSD = 6.2% (n = 5); luminol-Tb CP NPs stored at 4 °C	0.1–10.0 μg/mL	0.024 μg/mL

* n/a indicates that the item has not been assigned.

Table 5. Quantum dots/carbon nanotubes/cobalt oxide and polymer NPs.

(a) EBC Extraction Methodology and Detection Outcomes
EBC Methodology					Nanosensor
Refs.	Condenser	Extract Temp/Time	Vol.	Storage Temp/Time	Biomarker/Drug	Outcomes
Li et al., 2013 [62]	Mouthpiece connected to a 50 mL centrifuge tube	n/a *	Diluted to 2 mL with purified water	−20 to −25 °C 1 week	H2O2	High specific catalytic effect of Co3O4 nanoparticles on the luminol H2O2 CL reaction in alkaline medium with a lower concentration of luminol (1 × 10−8 M) No CL intensity interference by the presence of typical metal ions High stability of luminol–H2O2–Co3O4 nanoparticles system Optimal Co3O4 nanoparticles performance in a hydroxide solution at a pH of 13 Small sample and no incubation time with analytical runtime of <1 min Significant detection of higher H2O2 levels in EBCs of feverish subjects by Co3O4 nanoparticles compared with healthy and rheumatic subjects
Hasanzadeh et al., 2017 [83]	Lab-made pipe with a cooling trap mouthpiece	Room temp	2 mL	−25 °C	Malondialdehyde	Biocompatible sensors with low toxicity on cell lines, even at 200 ppm levels of GQDs Cyclic voltammograms recorded values between −1.0 and 1.0 mV in 0.1 M of PBS with higher peaks in the presence of GQD Good long-term stability due to the low solubility of the sensor in water Good sensitivity without any pre-treatment at physiological pH
Park et al., 2018 [84]	Custom-made collection system	0 °C 20 min	n/a	n/a	VOCs	Optimal performance using pastes containing 9 wt % of SWCNT Responses correlate linearly with the concentrations of the VOCs Each VOC generates a different fingerprint pattern from the IL sensor array Long-term stability of SWCNT-ILs and robustness to humidity and air Successful SWCNT-IL differentiation between diseased and healthy breath
Nasehi et al., 2022 [82]	Lab-made setup MVEBC samples collected from the waste of the ventilator	n/a	n/a	n/a	Phenobarbital drug	Optimal parameters for maximum fluorescence are pHs of 8.0, 0.02 mol/L, and 0.02 mL of MIP@ N-CDs and an 8 min incubation time Adequate nanosystem selectivity for the quantification of phenobarbital Successful quantification of phenobarbital in MVEBC of pre-term newborn babies
Mokhtari et al., 2022 [81]	Lab-made setup MVEBC samples collected from the waste of the ventilator	n/a	n/a	Freshly used	Phenobarbital drug	Proven maximum fluorescence at 3 replications, 1.0 μg/mL of 1phenobarbital, a pH of 9, 0.10 μL of phosphate buffer (0.1 mol/L), 40 μL of luminol-Tb CP NPs, and a 10 min incubation time No interference by co-administered drugs with the quantification of phenobarbital Successful detection of phenobarbital in MVEBC of newborns as compared with HPLC
(b) Performance of Quantum Dots/Carbon Nanotubes/Cobalt Oxide and Polymer Nanoparticles Properties
References	Age	Groups (N)/Health Status			Read-Out Method	Reproducibility/Storage	Linear Range	LOD
Li et al., 2013 [62]	n/a	>Feverish (9) >Rheum (13) >Healthy (13)			Microarray scanner of SynergyTM 2 Multi-Mode Microplate Reader IFFM-A CL analyzer	RSD of 2.1% for 100 nM of H2O2 (n = 11)	0.3 nmol/L	1.0 nmol/L–1000 nmol/L
Hasanzadeh et al., 2017 [83]	n/a	Healthy (n/a)			Electrochemical technique: square-wave voltammetry (SWV)	Simultaneous testing of the electrode with 3 replicates, with RSD of 6.16%.	0.06–0.2 μM	0.329 ng/mL
Park et al., 2018 [84]	n/a	>Healthy (n/a) >Diseased (n/a)			Chemiresistive sensor	n/a	n/a	n/a
Nasehi et al., 2022 [82]	Newborns and premature babies	>Healthy (n/a) >Under mechanical ventilator receiving phenobarbital (1)			Jasco FP-750 spectrofluorometer	Inter-day RSD of 6.1 and 6.9, and intra-day RSD of 1.8 and 1.8% for five replicates	0.01–8.0 μg/mL	0.006–0.18 μg/mL
Mokhtari et al., 2022 [81]	Newborns and premature babies	>Healthy (n/a) >Under mechanical ventilator receiving phenobarbital (3)			FP-750 spectrofluorometer	Inter-day RSD of 3.6 and intra-day RSD of 5.4% Batch-to-batch reproducibility with RSD = 6.2% (n = 5); luminol-Tb CP NPs stored at 4 °C	0.1–10.0 μg/mL	0.024 μg/mL

* n/a indicates that the item has not been assigned.

4. Discussion

It remains crucial to evaluate the variety of reported nanosensors in terms of sensitivity and applicability, given their diverse characteristics and read-out methods. The comparison would be significant if the same functionalization or detection techniques were used across various nanosensors. For instance, the H₂O₂ biomarker was detected in EBC using different nanosensors of copper or cobalt type. The more complex copper sensor, CuNPs, relied on an electrochemical signal [63,75]. On the other hand, simpler methods, including cobalt Co₃O₄ NPs, relied on chemiluminescent emission and spectrophotometric measurement [62,77]. Interestingly, the more complex CuNP-based electrochemical method showed a significantly lower LOD (0.0026 nmol/L) and a wider LR (0.01–500 nmol/L) as compared to simpler Co₃O₄ NP-based methods.

Phenobarbital drug detection in EBC was only performed using quantum dots. In fact, MIP@ N-CDs were more reliable detectors as presented by broader phenobarbital LODs (0.006–0.18 μg/mL) and LRs (0.01–8.0 μg/mL) compared to luminol-Tb CP NPs with LODs of 0.024 μg/mL and LRs of 0.1–10.0 μg/mL [81]. As for reproducibility assessments, both nanosensors presented a comparable RSD of a 5–6% range. AgNPs and QDs were employed for the oxidative stress biomarker detection in EBC using electrochemical approaches. PARG-GQDs presented higher MDA sensitivity with LODs (0.329 nM) as compared to LODs (0.817 µmol/L and 0.59 µmol/L) by poly (DA-CS)-AgNPs and RF-PT-AgNPs, respectively. Except for MDA biomarkers, silver AgNP-based nanosensors were solely presented as drug sensors. Calibration, co-administered drugs interference, and sensitivity assessments of these nanosensors were commonly applied by almost all these studies. The drug sensitivity of these differently coated AgNPs was read using simple spectrophotometric or spectrofluorimetric measurements. In a single approach, DOX in EBC samples was measured proportionally to reduced AgNP levels in the tested sample. This rapid one-step approach presented a low LOD (0.004 µg/mL) [67]. Still, the absence of DOX-AgNP testing on the EBCs of real DOX-treated patients was a major drawback. Other AgNPs’ sensitivity to drugs depended on AgNP surface coats that either caused drug binding or initiated NP aggregation. Either way, such changes were translated as increased or decreased emission signals, proportionally to the drug concentration upon NP excitation. Interestingly, low LODs confirmed high sensitivity toward drug detection in EBC, as shown by detection values of 0.005, 0.01, and 0.5 µg/mL for LTG drug-specific ASA-AgNPs, PHT drug-specific InS-AgNPs, and antibiotic-specific SDS-AgNPs (gentamycin and tobramycin), respectively. It is worth noting that InS-AgNPs were only applicable as PHT sensors without the quantification of the drug in the samples and were not tested on the EBCs of real PHT-treated patients [71]. Also, the previously mentioned SDS-AgNPs presented a drastically higher LOD of 4.1 mg/mL when used for aspirin detection instead of antibiotics [72]. Similarly, a high LOD of 0.06 mg/mL was obtained for DEF detection using AgNPs. These obstacles were not faced in CuNP-based EBC drug detection while still maintaining low LODs and RSD rates. By following the same experimental setup and assessments as in AgNP studies, Cu-NC-CTAB [79] and Cu-NCs [78] presented 0.08 and 0.06 µg/mL LODs, respectively, with <6% RSD.

Remarkably, femtomolar LOD was only reported by dopamine–Fe/Fe₃O₄NPs nanosensors for the detection of cancer and immune biomarkers in EBC samples. However, it lacked any specification of LR values. Independent from all the previously mentioned NP types, AuNPs presented a promising sensing ability of several EBC biomarkers, with constant nanosensor improvement over the years, as shown by successive publications in the literature. AuNPs were mostly applied for CEA, NSE, and SCC cancer biomarkers detection in EBC. For this, two consecutive studies designed an AuNP-based Love Wave immunosensor, starting with a primary design to successfully detect CEA, NSE, and SCC and later refined for enhanced specificity toward CEA detection. Interestingly, both methods revolved around coating AuNPs with specific anti-cancer biomarker antibodies. This step ensured a wide LR (1–16 ng/mL) and low LOD (1 ng/mL) for CEA detection. However, this assay was only replicated three times and was outperformed by other gold nanosensor types with picomolar LOD and LR values. Hence, picomolar levels of CEA in EBC were successfully measured using an SEECL-based AuNP approach with an RSD < 2% [52]. This nanosensor was further improved and reported a similar % of RSD and sensitivity of picomolar LOD and LR, with a wider selectivity for CEA, NSE, and SCC [51]. Beyond cancer biomarkers, citrate-functionalized AuNPs have also been successfully applied for EBC analysis. A simple citrate–AuNP setup efficiently captured polyamines, yielding a low LOD (0.070–0.17 ng/mL), wide LR (0.20–20 ng/mL), and an RSD <4% [60]. Incorporating citrate–AuNPs into a more complex electrochemical sensor setup produced comparable results, consistently achieving RSD values ≤4% [55].

5. Conclusions

Based on our analysis, AuNPs outperform CuNPs and AgNPs given their optical colorimetric characteristics, high excitation coefficient, and improved LSPR properties [86]. Moreover, AuNPs possess chemical stability that, unlike CuNPs and AgNPs, can withstand oxidation and show a long shelf-life duration mainly due to their synthesis and size control via the citrate reduction method [87].

Nanosensors can be successfully functionalized by organic or inorganic compounds to detect various EBC analytes with high ranges of sensitivity and selectivity. The physicochemical characteristics of nanoparticles allow their deposition on electrode chips in millimeter sizes with single- or multi-analyte detection mode. Moreover, the relatively large surface area of NPs facilitates analyte collection and signal amplification based on electron donation and surface plasmon resonance shifts. NPs hold promising prospects to push EBC testing toward the point-of-care, aiding practitioners, physicians, and patients in immediate clinical management. The ability to differentiate metabolic patterns influenced by disease progression, environmental factors, and drug metabolism highlights the promise of nanosensors in EBC metabolomics research. Proteomic biomarker detection in EBC has been primarily focused on lung cancer, with nanosensors identifying CEA, NSE, and SCC antigens. These cancer-related proteins have been effectively detected using AuNP-based immunosensors, which exhibit low LODs and high reproducibility.

To the best of our knowledge, genomic detection in EBC remains unexplored, with no current EBC nanosensors specifically designed to detect nucleic acids, microRNAs, or gene expression markers. Given the success of AuNPs functionalized with DNA/RNA probes in other biofluid analyses, future studies should explore their application in EBC-based genomic biomarker detection. Integrating AuNP-based nanosensors with genomics and transcriptomics approaches could provide a deeper understanding of lung cancer pathogenesis, airway inflammation, and respiratory disease progression, paving the way for more comprehensive and personalized diagnostic strategies.

6. Future Directions

A standardized collection method with a specified EBC volume should be set for future EBC-related studies. In addition, an internal control using self-reference data from urine, plasma, or saliva should be clearly identified prior to clinical applications. Based on our review, the literature lacks EBC–nano applications for genomics and transcriptomics analyses that can be used to decipher the brain–gut microbiota imbalances. Such research is crucial to interpret the effects of diet, stress, and environmental pollutants on the development of airway and systemic diseases.