Applications for direct breath analysis by mass spectrometry (MS) are rapidly expanding. One of the more recent mass spectrometry-based approaches is secondary electrospray ionization coupled to high-resolution mass spectrometry (SESI-HRMS). Despite increasing usage, the SESI methodology still lacks standardization procedures for quality control and absolute quantification. In this study, we designed and evaluated a custom-built standard delivery system tailored for direct breath analysis. The system enables the simultaneous introduction of multiple gas-phase standard compounds into ambient MS setups in the lower parts-per-million (ppm) to parts-per-billion (ppb) range. To best mimic exhaled breath, the gas flow can be heated (37 °C–40 °C) and humidified (up to 98% relative humidity). Inter-laboratory comparison of the system included various SESI-HRMS setups, i.e. an Orbitrap and a quadrupole time-of-flight mass spectrometer (QTOF), and using both single- as well as multi-component standards. This revealed highly stable and reproducible performances with between-run variation <19% and within-run variation <20%. Independent calibration runs demonstrated high accuracy (96%–111%) and precision (>95%) for the single-compound standard acetone, while compound-specific performances were obtained for the multi-component standard. Similarly, the sensitivity varied for different compounds within the multi-component standard across all SESI-Orbitrap and -QTOF setups, yielding limits of detections from 3.1 ppb (for p-xylene) to 0.05 ppb (for 1,8-cineol). Routinely applying the standard system throughout several weeks, allowed us to monitor instrument stability and to identify technical outliers in exhaled breath measurements. Such routine deployment of standards would significantly improve data quality and comparability, which is especially important in longitudinal and multi-center studies. Furthermore, performance validation of the system demonstrated its suitability for reliable absolute quantification while it illustrated compound-dependent behavior for SESI.

The direct analysis of exhaled breath has met with increasing interest from various research fields, with mass spectrometry (MS) being the primary technology applied today [1]. This is primarily due to the high sensitivity of modern mass spectrometers, which allows to detect volatile organic compounds (VOCs) that are present in breath at low concentration, parts per million (ppm) to parts per trillion (ppt). This enabled breath analysis in real time, bypassing any sample preparation, and together with the wide range of detectable metabolites resulted in expanding applications, including clinical monitoring, phenotyping or diagnostics [2].

Secondary electrospray ionization high-resolution mass spectrometry (SESI-HRMS) is a relatively young technique for direct breath analysis, with applications ranging from basic research to biomarker discovery to clinical trials [3–7]. In contrast to the more established MS-based techniques for direct breath analysis, i.e. selected ion flow tube mass spectrometry (SIFT-MS) [8, 9] and proton transfer reaction mass spectrometry (PTR-MS) [10, 11], the SESI source operates at ambient pressure, which results in a softer analyte ionization. Furthermore, this enables coupling to state-of-the-art mass spectrometers and, as a result, detection with very high mass resolution, as well as compound identification by tandem MS.

One current drawback has been the lack of well-defined standardization procedures for SESI-HRMS. While efforts to standardize exhalation manoeuvres have been made [12], no procedures for identifying technical variations have been agreed upon. Nonetheless, this is fundamental for data comparability from different instruments or sites, and it would strengthen the potential of SESI-HRMS for clinical application. Furthermore, SESI-HRMS-based analysis is currently restricted to relative quantification because its ionization process is not yet fully understood. Particularly in clinical diagnostics, however, absolute quantification is required to define thresholds for breath-derived biomarkers.

MS-based methods that are well established in clinical laboratories all employ routine standardization measures to validate instrument performance, most commonly using internal standards. This is not applicable for direct analysis, as no sample storage is involved to which an internal standard can be added to in a controlled fashion. Thus, external standards are a viable option, particularly when introduced in the gas phase to mimic ionization of breath-derived compounds. This can be realized by commercially available systems (e.g. Owlstone Medical, UK) which are based on permeation tubes [13, 14], but they are limited in the number of standards.

To generate robust and comparable data by SESI-HRMS based breath analysis, methodological standardization procedures are required. Reproducibility and comparability have recently become even more important, since SESI-HRMS is applied in an increasing number of clinical trials and laboratories [15]. This also corresponds to the current need for quantifying potential breath biomarkers, which have been identified in various studies [1, 6, 16], as pointed out recently by Jeerage et al [17]. To address this need, we present a standardization approach for routine use in MS-based breath analysis. We developed a custom-designed gas-phase standard delivery system that mimics exhaled breath in the best possible way, including the simultaneous delivery of multiple standard compounds. In this study, the system was thoroughly evaluated on multiple SESI-HRMS setups across different laboratories and using standards that included compound classes commonly detected in breath analysis.

2.1. Technical design of the delivery system

The gas-phase standard was diluted in a stepwise fashion using calibrated mass flow controllers (Bronkhorst High-Tech B.V., Ruurlo, Netherlands). By employing different gas cylinders, various gas-phase standards could be introduced. All stainless-steel tubings and swagelok fittings in contact with the standard were silica coated (SilcoNert®2000, SilcoTek, Bellefonte, PA, USA) to avoid sample adhesion to and reaction with the material (see figure S1 for more details).

2.2. Gas-phase reference material

In this work, a single-component standard containing 1.67 ppm acetone (PanGas AG, Dagmersellen, Switzerland) and a multi-component standard containing different VOCs at 0.1–1 ppm (Apel Riemer Environmental Inc., Miami, FL, USA) served as gas-phase reference standards (table 1). Medical air or VERISEQ® pharmaceutical air (both PanGas AG, Dagmersellen, Switzerland) were applied for stepwise dilution. The standards were humidified by passing a fixed flow of dilution gas (250 ml min−1) through a glass frit placed in optima™ water (Fisher Scientific GmbH, Schwerte, Germany). Humidified standards were subsequently passed through a heated tube to avoid condensation. For performance tests, dilution ratios were fixed: a 3:25 dilution for the multi-component standard and a 1:2 dilution for acetone with dry gas was followed by a 1:6 dilution with humidified gas, resulting in the injection of 15.8 ppb 1,8-cineol (exemplarily for the standard mixture) and 139.2 ppb acetone.

Table 1. Specifications of the gas-phase standards.

StandardCompoundMolecular formulaconc.* (ppb)m/z exp. [M + H]+DetectedMultiaDimethyl sulfideC2H6S93863.03Breath [10, 18], SESI [19]c PhenolC6H5OH76395.05Breath [10], SESIc StyreneC8H8788105.07Breath [20], SESIc  p-XyleneC8H10789107.09Breath [21, 22], SESId Isobutyl nitrateC4H9NO3783120.07SESIc  α-PineneC10H16792137.13Breath [23], SESI [12] 1,8-CineolC10H18O789155.14Breath [23, 24]  β-CaryophylleneC15H2477205.20SESI [25]SinglebAcetoneC3H6O167059.05Breath [10, 26–28], SESI [1, 29]

Reference material: *concentration with ±5% uncertainty, supplied by aApel-Riemer and bPanGas in nitrogen, m/z: expected (exp.). cidentity uncertain—assigned based only on exact mass in breath SESI-HRMS spectra (mass error <0.005 Da), ddetected by SESI-HRMS using a PTR-MS reference gas cylinder [30].

2.3. SESI-HRMS analysis

Exhaled breath, humidified air and humidified gas-phase standards were directly ionized using commercial SESI ion sources (FIT Fossilion Technology, Madrid, Spain) and analyzed in real time on two different high-resolution mass spectrometers coupled to the ion source: an Orbitrap (Q Exactive™ Plus, Thermo Fisher Scientific Inc., Waltham, MA, USA) and a quadrupole time-of-flight (QTOF) MS (TripleTOF 5600+, AB Sciex, Concord, ON, Canada).

To determine the system performance, the standard delivery system was coupled to the respective instrument setups, which were installed at different sites: a SESI-Orbitrap instrument at the University Hospital Zurich (USZ), and two SESI-QTOF instruments, one each at USZ and the University Children's Hospital Zurich. On all sites, the ion source was operated at 130 °C and with a net sample flow of 0.3 l min−1. Further details on operating at the ion source have been described elsewhere [12]. The pure nano electrospray was generated using silica emitters (50 cm length, 20 µm diameter, New Objective Inc., Woburn, MA, USA) and a 0.1% (v/v) aqueous formic acid solution as the primary electrospray solvent. Spectral data were recorded in the mass range from 50 to 500 Da. The spray voltage in positive ionization mode was set to +3.5 kV on the Orbitrap. On the QTOF-MS setups, a spray voltage of +4.5 kV was applied and spectra were recorded with an accumulation time of 0.5 s. MS/MS spectra of the standard mixture components were recorded on the SESI-Orbitrap in positive ion mode with an isolation window of 0.4 Da and compared against library MS/MS spectra available in the human metabolome data base [31].

For exhaled breath analysis, test subjects refrained from eating, drinking, brushing their teeth, or using facial cosmetics for at least one hour prior to the breath measurements. Sitting in an upright position, the subjects exhaled with an overpressure of 10 mbar directly into the SESI-HRMS through a single-use mouthpiece (ACE Handels- und Entwicklungs GmbH, Freilassing, Germany), that was connected to the ion source by a Teflon adapter. Exhaled breath was analyzed in positive ionization mode, recording six consecutive exhalations. The part of this study with human volunteers had been approved by the local ethical committee (KEK-ZH 2019-00030) and was conducted according to the Declaration of Helsinki. Written informed consent was obtained from all subjects.

2.4. Data preprocessing

Raw mass spectral data files were converted into .mzXML files using MSConvert (ProteoWizard v3.0.2) [32] and preprocessed in Matlab® R2020a (MathWorks Inc., Natick, MA, USA). The data preprocessing included interpolation, baseline adjustment, peak signal integration and peak picking. For relative intensities, signal intensities were normalized to the total ion current (TIC). Furthermore, the first 50 scans were excluded to ensure well controlled gas flows. Signal averaging of all scans resulted in the final intensity matrix.

3.1. A tailored standard system for breath analysis

The gas standard delivery system developed in this work was primarily used to monitor instrumental fluctuations on different SESI-HRMS breath analysis setups. The system was designed to deliver low mole fractions of humidified gas-phase standards which are introduced from a gas cylinder and diluted in a stepwise fashion to the desired concentration (figure 1). Furthermore, the standards were humidified (98% rel. humidity) and heated (37 °C) before entering the ion source to simulate exhaled breath and the ionization process of breath-borne metabolites in the best possible way. While the first dilution step served to vary gas-phase standard concentrations, the second ensures constant relative humidity.

Figure 1. Design of the gas-phase standard delivery system. (a) Schematic illustration, dynamic dilution of the standard controlled by multiple mass flow controllers (MFCs) allows to vary the standard concentration (MFC 1 and 2) but keeping the sample humidity constant (MFC 3). Typical flow rates: f1= 6 ml min−1, f2= 44 ml min−1 and f3= 250 ml min−1 for a net flow rate of 300 ml min−1 introduced to the SESI source. (b) Photograph (top view, opened) of the setup showing the positioning of the MFCs and the humidifier.

Download figure:

Standard image High-resolution image

Gas-phase standards were selected for positive ionization mode and to cover compounds of different mass and a range of compound classes (table 1). The compounds are expected to be detectable by SESI-HRMS with some of them having been previously detected in exhaled breath (see table 1 for references) [1, 10, 12, 18–30] . The design of the system allows to deliver standards with concentrations in the parts-per-billion (ppb) range. Taking into account the optimal gas flow rates specified for the SESI source (0.2–0.5 l min−1), this resulted in a working range of about 1.5–300 ppb, well in line with the sensitivity reported for SESI-HRMS [33, 34]. Concentration ranges can be further adjusted by employing more highly concentrated standards or installing an additional dilution step in front of the humidification step.

3.2. System performance evaluation

The selected gas-phase standards were mainly detected as protonated molecular ions ([M + H]+), with two exceptions: 1,8-cineol showed two additional fragments at m/z 81 and m/z 137 with strong signal response (figure S2), as identified by MS/MS analysis (figure S3). This included the base peak at m/z 137, which unfortunately coincides with the protonated molecular ion of α-pinene. Isobutyl nitrate was not detected in the present concentration range (figures S4 and S5).

The performance of the standard system was first assessed by introducing the multi-component standard at fixed concentrations. Over a test period of 1 h, the standard was introduced every 10 min for a minimum of two minutes. The system showed high stability between runs on the SESI-Orbitrap setup with a coefficient of variation (CV) <2% for the prominent 1,8-cineol ions (figure 2) and CV <6% for all mixture compounds (table 2). Repetition on the SESI-QTOF setups for inter-laboratory comparison revealed an increased variation for 1,8-cineol of up to 12% (figure 2(b)). The remaining standard compounds showed comparable variations between the QTOF2 and Orbitrap setup, but increased variability on the QTOF1 setup (table 2). Nevertheless, the obtained between-run variations, with exceptions of that for dimethyl sulfide, phenol and p-xylene measured on the QTOF1, were well in line with official guidelines for bioanalytical method validation recommending <15% CV [35, 36]. Within-run variations of <20% CV were obtained for most standard compounds, with the highest stability (<5% CV) for the highest-intensity features m/z 137 (α-pinene/1,8-cineol fragment) and m/z 155 (1,8-cineol) across all laboratories (table S1). Accordingly, the two exceptions showing strong within-run fluctuations (dimethyl sulfide and β-caryophyllene) were also the lowest intensity features.

Figure 2. Standard system performance based on between-run variation upon repeated introduction of the standard mixture during 1 h. Performance exemplarily shown for 16 ppb 1,8-cineol. (a) Signal response of 1,8-cineol (m/z 155) and its fragment ions (m/z 81, m/z 137; see figure S3) on a SESI-Orbitrap, intensities normalized to the TIC. (b) Inter-laboratory comparison of the SESI-HRMS setups: Orbitrap, QTOF1 and QTOF2. Dotted line: recommendation of max. 15% CV. Represented data included n = 6 timepoints (⩾180 scans each, every 10 min), averaged CV defined after outlier removal (Tukey's fence).

Download figure:

Standard image High-resolution image

Table 2. Inter-laboratory comparison of the system performance based on between-run variations of the multi-component standard.

Standard compoundm/zCV (Orbitrap)CV (QTOF1)CV (QTOF2)Dimethyl sulfide63.035.218.62.2Phenol95.051.715.93.6Styrene105.072.214.51.9 p-Xylene107.093.715.91.4 α-Pinene/1,8-Cineol fragment137.131.510.65.11,8-Cineol155.141.87.57.2 β-Caryophyllene205.203.68.73.4

CV: coefficient of variation in %; different SESI-HRMS setups: Orbitrap = QExactive+ (Thermo Fisher), QTOF = TripleTOF (AB Sciex); n = 6 timepoints during 1 h.

3.3. Evaluating multiple delivery units

To achieve comparability between different SESI-HRMS operating sites, standard delivery systems are required in each laboratory. Therefore, three units of the system were manufactured and evaluated on the same SESI-HRMS setup. All three units were tested on a SESI-QTOF setup throughout one day, introducing the single-component standard acetone. Furthermore, the units were applied in randomized order to avoid confounding effects from instrumental drifts, which might occur over time and for which a monitoring system is currently not available.

All three system units performed with similar precisions as previously obtained for the multi-component standard on the QTOF setups. A discrepancy between the units in signal intensity was observed that could not be corrected for by signal normalization to the TIC (figure 3(a)). Nevertheless, the between-run variation for acetone with <10% CV (figure 3(b)) and even lower within-run variabilities (table S2) demonstrated a highly stable and repeatable performance regardless of the units and compounds tested.

Figure 3. Performance of multiple system units. (a) Signal response of 139 ppb acetone (m/z 59) applying the different system units in randomized order, intensities normalized to the TIC. (b) CV obtained for water cluster (m/z 55) and acetone (m/z 59) on each system unit. Dotted line: recommendation of max. 15% CV, data acquired on a SESI-QTOF with n = 9 data points (⩾180 scans each).

Download figure:

Standard image High-resolution image 3.4. Method validation for quantitative analysis

In addition to the qualitative performance, we evaluated the use case of the standard system for quantification. The concentration-dependent response of the standards was validated by repeated independent runs within one day, between several days and throughout the different laboratories as specified in table 3. The monitored concentration range for each standard was defined by their initial gas-phase concentration.

Table 3. Calibration curve characteristics of the standards detected by different SESI-HRMS setups.

SetupValidationStandard ([M + H]+)Polyn. fit orderR2 run1R2 run2R2 run3Acc.* (%)Prec.* (%)Sensitivity (a.u./ppb)LOD (ppb)LOQ (ppb)SESI-QTOFBetween-daysAcetone10.94350.99320.981496–11152.75 × 1021.525.08SESI-QTOFWithin-dayDimethyl sulfide40.99990.99990.999891–100255.53 × 101n.a.n.a. Phenol10.99610.99370.9914101–10952.53 × 1031.143.80  Styrene40.99940.99910.999997–107114.19 × 1022.598.64   p-Xylene30.98330.99750.994093–123115.09 × 1022.869.54  1,8-Cineol40.99290.99910.999687–10122.22 × 1040.361.19   β-Caryophyllene40.99760.99630.991791–104162.55 × 1030.321.07SESI-OrbitrapBetween-daysDimethyl sulfide40.98760.96660.984288–112514.03 × 1022.337.76Phenol10.99540.93010.989197–115133.99 × 1042.428.05 Styrene20.99610.98570.996790–109171.48 × 1042.909.66   p-Xylene10.99790.96740.996396–109161.76 × 1043.0710.25  1,8-Cineol40.99870.99570.995687–10181.16 × 1070.050.17   β-Caryophyllene40.99310.98020.996379–105131.00 × 1060.060.20

Independent runs were performed in triplicates on different SESI-HRMS setups as indicated. *Accuracy (Acc.) and precision (Prec.) were estimated based on the signal response of three calibration standard samples at low, medium, and high concentration, that were excluded from calibration curve fitting, and averaged over three independent runs. LOD: limit of detection, LOQ: limit of quantification. Within-day: runs on the same day, between-days: runs during 3 d (acetone) or 4 d (multi-component standard).

Isobutyl nitrate was not detected since no change in response was obtained for the expected [M + H]+ ion upon increasing concentrations. This was consistent across all analyzers and laboratories (figures S4 and S5), and could be due to an interference present at the exact same m/z value. Among the detectable standards, different behaviors were identified by least-square regression analysis of the calibration curves (table 3). For example, the acetone response was linear (figures 4(a) and S6(a)), while regression for 1,8-cineol (m/z 155) required a 4th order polynomial fit (figures 4(b) and S6(b)). The linear range for 1,8-cineol was between 13 and 53 ppb. Altogether, lower orders of polynomials were observed for the standards on the SESI-Orbitrap setup, indicating that for some of the standards the present concentration range approached saturation on the SESI-QTOF setup.

Figure 4. Calibration curves obtained for (a) acetone at m/z 59 (red diamonds) and (b) 1,8-cineol at m/z 155 (blue circles) on the SESI-QTOF setup. Each data point represents the averaged intensity from at least 120 scans per sample. (c) and (d): exemplary standard curves for accuracy and precision calculation. Grey markers: calibration samples used for curve fitting; colored markers: three calibration samples at low, medium, and high concentrations that were excluded from curve fitting and used to estimate accuracy and precision for each standard compound.

Download figure:

Standard image High-resolution image

Between-run accuracy and precision were assessed based on the independent runs of the calibration curves. Here, three standard samples at low, medium, and high concentration across the concentration range were excluded from curve fitting and used to calculate both accuracy and precision (figures 4(c) and (d)), averaged over the three independent runs and the three selected calibration standards. For most of the tested compounds, we obtained high accuracy and precision (table 3). Specifically, precision values for 1,8-cineol confirmed the between-run variation obtained from the previous stability test results (figure 2). Only dimethyl sulfide showed high variations on both SESI-QTOF and SESI-Orbitrap setups, which have not been observed during the stability test.

For all standard compounds, the limit of detection (LOD) and limit of quantification (LOQ) were determined by three times (LOD, for LOQ: 10 times) the standard deviation of the blank sample divided by the sensitivity, which was derived by dividing absolute signal intensity by the sample's concentration (ppb) and averaging calculated sensitivities across all concentration samples. The blank sample consisted of introducing humidified dilution gas into SESI-HRMS. While the LODs for 1,8-cineol and β-caryophyllene ranged in the lower ppt region (0.05 ppt and 0.06 ppt, respectively) on the SESI-Orbitrap setups, the LODs were significantly higher for the remaining multi-component compounds (2.3–3.1 ppb). The overall LOD range for each standard compound individually was comparable across the SESI-Orbitrap and SESI-QTOF setups.

3.5. Application to monitor technical fluctuations

One of the key motivations for the development of the standard system was to assess technical fluctuations on SESI-HRMS setups. Therefore, the multi-component gas-phase standard was introduced into a SESI-QTOF by the standard system over several weeks. In parallel, exhaled breath measurements of one subject were performed shortly before or after the gas-phase standard measurements. This resulted in 40 measurements each for the standard and exhaled breath.

Within-run variability for the most prominent standard 1,8-cineol remained below 10% CV, indicating a stable standard introduction, while more pronounced between-day variations were observed (table S4). To evaluate the standard system's monitoring potential, we focused on the water cluster signal at m/z 55 in exhaled breath, which is detectable on a SESI-QTOF. Because the humidity in breath is expected to be constant, water cluster signal is also assumed to be constant throughout the measurement period and was thus used hereafter to compare breath signal with the standard signal. Outliers in breath samples were defined based on normalization to the most prominent standard signal of the multi-component standard, i.e. 1,8-cineol (figure 5(a)). After outlier removal, Pearson's correlation analysis between the gas-phase standard and exhaled breath (water cluster) of the same subject showed a positive correlation (R = 0.64, figure 5(b)), revealing fairly synchronized fluctuations over the test period of several weeks. An improved correlation was observed between the water clusters in breath and the water clusters in standard samples (R = 0.72, figure S7). Additionally, the correlation analysis revealed that fragmentation of 1,8-cineol (m/z 137) was constant on this instrument.

Figure 5. Monitoring of technical fluctuations on the SESI-QTOF setup. (a) Boxplots of normalized water clusters in breath (m/z 55) normalized to the most prominent signals in the multi-component standard (m/z 137 and m/z 155), n = 40 independent measurements, outliers marked by orange circles. (b) Correlation matrix of the Pearson´s correlation coefficients for breath samples with parallel measurements of standard compounds (n = 40). Along the diagonal: intensity distribution for each m/z feature.

Download figure:

Standard image High-resolution image

The direct analysis of exhaled breath by highly sensitive instruments such as mass spectrometers has gained increasing interest from various research fields [1]. As demonstrated by SESI-HRMS, the scope of ionization technologies is expanding to fully explore and exploit insights provided by the analysis of exhaled breath. However, with new technologies comes the need for standardization procedures to ensure reliable analysis [15]. First suggestions for standardization in SESI-HRMS analysis have already been made [12]. While these focused on standardizing the procedure of breath sampling, quality control approaches for the entire instrumentation have never been tackled. For this reason, we created a standard system that allows for the simultaneous introduction of multiple (more than five) gas-phase standards with the goal of mimicking the complexity and humidity of breath samples [17, 22]. Inter-laboratory evaluation of the standard system revealed highly stable and reproducible delivery of gas-phase standards on various SESI-HRMS setups. Furthermore, estimated accuracy and precision demonstrated reliable performance of the system, as well as its potential for quantitative analysis.

In this work, we were able to thoroughly validate the manufactured gas-phase standard delivery system. Comparing its performance in various laboratories included, for the first time, evaluation on both SESI-Orbitrap and SESI-QTOF setups that are used for breath analysis. We applied a multi-component standard, similar to the one used by Liu et al [37], in addition to a single-component standard to extensively evaluate accuracy and precision in multiple independent runs, and were able to demonstrate the system's monitoring suitability over a period of several weeks.

While the standard delivery system was evaluated with an emphasis on exhaled breath measurements, the system is equally applicable for VOC analysis in general, using any direct ambient ionization techniques, particularly SIFT- or PTR-MS. The humidification step in the delivery system is optional and can be replaced by dynamic dilution with dry gas, depending on the humidity-dependence or independence of the respective detection method.

We observed different behaviors for the tested standard compounds based on their calibration curves (table 3). This not only agrees well with previously reported differences in sensitivity for a multi-component standard by SESI-HRMS [37], but also demonstrates the range of possible responses. Interestingly, a discrepancy between the different SESI-setups was obtained: the standard responses on the SESI-QTOF setup required a higher order polynomial fit than on the SESI-Orbitrap, e.g. for p-xylene and styrene, which showed a saturation trend. While the absolute signal of both standards is far from the detector saturation limit, this may rather imply a decreasing ionization efficiency for both compounds in the QTOF-connected ion source version as the standard mixture concentration increases. Nonetheless, the different responses observed in this study emphasize the importance of including multiple compound classes in a reference material for breath analysis and, therefore, underlines recently reported recommendations about a multi-component standard for breath analysis [17].

Applying the standard system in parallel to breath samples over several weeks demonstrated its suitability to monitor instrumental fluctuations and define technical outliers in exhaled breath. Therefore, we propose the routine inclusion of such a quality control in standard operating procedures on SESI-HRMS setups. Here, we observed a slightly