Thermal desorption tube sampler
11387092 · 2022-07-12
Assignee
Inventors
Cpc classification
H01J49/049
ELECTRICITY
International classification
H01J49/04
ELECTRICITY
Abstract
The disclosure provides a thermal desorption (TD) tube sampler. The sampler comprises a first connector configured to reversibly connect to a TD tube containing a sample, and a second connector configured to couple to a direct injection mass spectrometer. The TD tube sampler is configured to desorb a sample in a TD tube connected thereto, and feed the desorbed sample from the TD tube to a direct injection mass spectrometer such that the desorbed sample does not pass through a cold trap.
Claims
1. A thermal desorption (TD) tube sampler, wherein the sampler comprises a first connector configured to reversibly connect to a TD tube containing a sample, and a second connector configured to couple to a direct injection mass spectrometer, and the TD tube sampler is configured to desorb a sample in a TD tube connected thereto, and feed the desorbed sample from the TD tube to a direct injection mass spectrometer such that the desorbed sample does not pass through a cold trap, wherein the TD tube sampler comprises a cold trap and is configured to couple to a gas chromatography mass spectrometer (GC-MS), and the TD tube sampler has a first operating mode, for use when the TD tube sampler is coupled to a direct injection mass spectrometer, wherein the TD sampler is configured to feed the sample from the TD tube to the direct injection mass spectrometer such that the sample does not pass through the cold trap; and a second operating mode, for use when the TD tube sampler is coupled to a GC-MS, wherein the TD sampler is configured to feed the sample from the TD tube to the direct injection mass spectrometer such that the sample does pass through the cold trap.
2. A TD tube sampler according to claim 1, wherein the TD tube sampler does not comprise a cold trap.
3. A TD tube sampler according to claim 2, wherein the TD sampler comprises a conduit which extends between the first connector and the second connector, and the conduit comprises an aperture therein fluidly connecting an internal channel of the conduit to an external environment.
4. A TD tube sampler according to claim 1, wherein the sample comprises one or more volatile organic compounds (VOCs).
5. A TD tube sampler according to claim 1, wherein the TD sampler may be configured to desorb the sample for between 1 minute and 20 minutes.
6. A TD tube sampler according to claim 1, wherein the TD tube sampler is configured to desorb the sample by being configured to heat the TD tube to a temperature of between 100° C. and 450° C.
7. A TD tube sampler according to claim 1, wherein the TD tube sampler is configured to desorb the sample by being configured to cause an inert gas to flow through the TD tube at a rate of between 50 sccm and 250 sccm.
8. A TD tube sampler according to claim 1, wherein the TD tube sampler is configured to receive a plurality of TD tubes, wherein each tube comprises a sample, and the TD tube sampler is configured to reversibly couple to each of the TD tubes in turn and to feed the sample from the coupled TD tube to the direct injection mass spectrometer such that the sample does not pass through a cold trap.
9. An apparatus comprising the TD tube sampler according to claim 1 and a direct injection mass spectrometer coupled thereto.
10. An apparatus according to claim 9, wherein the direct injection mass spectrometer comprises a selected ion flow tube mass spectrometer (SIFT-MS), a proton transfer reaction mass spectrometer (PTR-MS), or an ion mobility spectrometer (IMS).
11. An apparatus according to claim 10, wherein the proton transfer reaction mass spectrometer is a proton transfer reaction time-of-flight mass spectrometer (PTR-ToF-MS).
12. A method of analysing a sample, the method comprising: desorbing a sample from a thermal desorption (TD) tube; and feeding the desorbed sample into a direct injection mass spectrometer to thereby analyse the sample; wherein the method does not comprise feeding the sample into a cold trap, and wherein the method comprises switching a valve to prevent the sample from being fed into a cold trap.
13. A method according to claim 12, wherein the method is used to identify one or more signature VOCs in a patient sample, a food sample, a cosmetic sample, an environmental sample or an automotive emissions sample.
Description
(1) For a better understanding of the invention, and to show how embodiments of the same may be carried into effect, reference will now be made, by way of example, to the accompanying drawings, in which:—
(2)
(3)
(4)
(5)
(6)
(7)
EXAMPLES
(8) Materials and Methods
(9) Measurements were conducted using a PTR-TOF 1000 instrument (Ionicon Analytik GmbH, Innsbruck, Austria). Optimal conditions for volatile organic compound (VOC) identification and quantification were defined according to a previously described experimental workflow.sup.10, dedicated to method optimization under breath-relevant conditions. Briefly, the workflow consisted of: (i) screening of reduced drift field conditions using different reagent ions, (ii) evaluating the impact of a change in humidity on branching ratios and (iii) gravimetric calibration using permeation or diffusion tubes. Optimal conditions for the drift tube were: temperature 110° C., pressure 2.30 mbar and voltage 350 V, resulting in an E/N of 84 Td (1 Townsend=10-17 V cm2). Reagent and analyte ions used in VOC determination throughout the paper are summarized in table 1. Sample inlet flow rate was set to 200 and 130 sccm for direct injection and thermal desorption analysis, respectively.
(10) TABLE-US-00001 TABLE 1 Mass peaks monitored in breath and standards with respective identifications Analyte ion Reaction (m/z) Reagent ion Channel Identification 57.03 NO.sup.+ Hydride Propanal abstraction 59.05 H.sub.3O.sup.+ Proton transfer Acetone 71.05 NO.sup.+ Hydride Butanal abstraction 89.06 H.sub.3O.sup.+ Proton transfer Butanoic acid 94.05 NO.sup.+ Electron transfer Phenol 103.07 H.sub.3O.sup.+ Proton transfer Pentanoic acid 117.09 H.sub.3O.sup.+ Proton transfer Hexanoic acid 155.14 NO.sup.+ Hydride Decanal abstraction
(11) A TD autosampler (TD100-xr, Markes International Ltd, Llantrisant, UK), which was adapted for PTR-MS analysis, was employed. The original configuration for TD analysis is shown in
(12) The t-piece 1 comprises an overflow aperture with a diameter of 1 mm configured to ensure the flow from the TD autosampler matches the flow of the PTR-MS.
(13) The adapter allowed the collection of VOCs directly from the TD tube, entirely bypassing the cold trap and achieving one-stage desorption. In the final optimized TD method, TD tubes were desorbed for 10 minutes at 280° C. and 130 sccm using zero-grade nitrogen purified by means of a Supelpure HC hydrocarbon trap (Sigma Aldrich, St Louis Mo., USA).
(14) For the calculation of breath VOC concentrations ([VOC].sub.breath) based on TD-PTR data, the inventors employed the following equation:
[VOC].sub.breath=([VOC].sub.TD×t.sub.D×φ)/V.sub.b
(15) Where [VOC].sub.TD is the measured concentration, t.sub.D and φ are desorption time and flow, respectively and V.sub.b is the collected breath volume or the standard volume loaded onto the tube. The same approached was used to compare loaded and desorbed VOC amounts when working with authentic standards.
(16) The inventors employed stainless steel TD tubes containing a mixed sorbent bed consisting of Tenax TA/Carbograph 5TD (Bio-Monitoring C4-C30, Markes Ltd). TD tubes were conditioned using a TC 20 conditioning station (Markes Ltd) and following the manufacturer's recommendations. Authentic standards (Sigma Aldrich) were loaded into TD tubes by means of custom-made permeation tubes. The permeation tubes consisted of PTFE tubing, sealed at both ends. Permeation tubes were incubated at 30° C. at a constant flow rate using a permeation unit (ES 4050P, Eco Scientific, Stroud, Gloucestershire UK). Clean air was supplied using a membrane pump (KNF Neubeger UK, Witney, Oxfordshire), connected to a hydrocarbon trap. Before tube loading, the permeation unit was connected to the PTR-ToF-MS by means of a PEEK union connection and measurement was conducted for 3 minutes; the union was then switched by a TD tube placed between the permeation unit and the mass analyzer. Immediately upon switching a drop in VOC signals down to baseline levels was observed. Standard loading was conducted at a flow of 200 sccm for 2.5 minutes, during which no increase in the signals of the VOCs of interest was observed. This showed that sample VOC loading could be carried out without detectable breakthrough. After loading, the TD tube was removed and the device was connected back to the PTR-ToF-MS and left to equilibrate for 2 minutes.
(17) The choice of compounds used was based upon previous studies on diagnosis of oesophago-gastric cancer and colorectal cancer by means of SIFT-MS.sup.12,13,14. The inventors selected three aldehydes (propanal, butanal, decanal), three fatty acids (butanoic acid, pentanoic acid, hexanoic acid) and phenol. Clinically relevant concentrations for these compounds range from the low-ppbV level to a few hundreds of ppbV.
(18) In order to optimise desorption time on the one stage system, VOCs were loaded onto the TD tubes at a fixed reference concentration. For this experiment, in addition to the previously mentioned seven oxygenated VOCs, acetone was also loaded onto the TD tubes. Loading concentrations were: acetone 283.5 ppbV, butanoic acid 21.0 ppbV, pentanoic acid 70.1 ppbV, hexanoic acid 8.8 ppbV, phenol 88.7 ppbV, propanal 12.5 ppbV, butanal 4.9 ppbV, decanal 2.3 ppbV. Tubes were then analysed by TD-PTR-ToF-MS prolonging the desorption time up to 20 minutes.
(19) The performance parameters for method validation and the techniques to evaluate them were chosen adopting the guidelines established by the European Respiratory Society technical standard for the analysis of exhaled biomarkers in lung disease.sup.15, which recommend evaluating method performance with respect to linearity, limit of quantification (LOQ), limit of detection (LOD) and repeatability. Calibration curves were established based on comparison between loaded and desorbed amounts of the authentic standards. Different loading concentrations were achieved by changing the flow rate applied to the permeation oven. Ten TD tubes were prepared, corresponding to five concentration levels, each in duplicate.
(20) Comparison with TD-GC-MS
(21) The performance of the newly built TD-PTR-ToF-MS platform was directly compared with that of the leading technology for VOC quantitation, i.e. TD-GC-MS. TD-GC-MS analysis was conducted using a previously described analytical method.sup.11. The inventors employed a Markes TD-100 thermal desorption unit, using a two-stage desorption program. GC-MS analysis was performed using an Agilent 7890B GC with 5977A MSD (Agilent Technologies, Cheshire, UK) equipped with a Zebron ZB-642 capillary column (60 m×0.25 mm ID×1.40 μm df; Phenomenex Inc, Torrance, USA). Chromatographic peaks were normalised against an internal standard (d8-toluene, 10 ppm in nitrogen). This was added to each TD tube using a gas canister (Thames Restek). Calibration curves for TD-GC-MS were constructed by loading and analyzing a set of TD tubes on an analytical platform consisting of a commercial automated TD system, coupled to GC-quadrupole MS. The results were extracted as mass area units obtained in the total ion current mode.
(22) Breath Collection and Analysis
(23) The findings obtained on authentic standards were corroborated by means of experiments conducted on breath samples. Multiple breath samples (n=10) were collected within a limited time span from a healthy volunteer. The inventors employed a breath collection device (ReCIVA, Owlstone Ltd, Cambridge UK), using optimized sampling parameters, as determined in a previous study.sup.10. The device allowed for direct breath collection onto TD tubes. The tubes were then immediately analyzed by TD-PTR-ToF-MS. In this experiment, a desorption step of 20 minutes was employed. To assess the usability of the TD-PTR-ToF-MS analytical platform in the high throughput analysis of clinical samples, breath samples were collected from patients attending St Mary's Hospital for upper gastrointestinal endoscopy or surgery. The use of the ReCIVA breath collection device allowed the simultaneous sampling of up to four TD tubes from each patient. Ethical approval was obtained through NHS Health Research Authority (NRES Committee London—Camden and Islington, approval granted on 16 Jul. 2014, REC reference 14/LO/1136).
(24) Quality Control
(25) During the described experiments, a series of quality checks were conducted on the PTR-ToF-MS daily, according to a previously reported protocol.sup.10. Impurities with the three ionization modes were O.sub.2+(<2%) for H.sub.3O.sup.+ as primary ion, NO.sub.2.sup.+ (<3%) for O.sub.2.sup.+ as primary ion and NO.sup.+ and NO.sub.2.sup.+ (<5% altogether) for O.sub.2.sup.+ as primary ion, respectively. Quantitation accuracy with the three ions was within ±10% of a certified standard, represented by a Trace Source™ benzene permeation tube (Kin-Tek Analytical Inc., La Marque Tex.). The consistency of fragmentation with NO.sup.+ and H.sub.3O.sup.+ as primary ions was assessed by measuring the ratio between reference mass peaks with given standard compounds. For NO.sup.+, the inventors used the ratio between peaks m/z 71 and 43 using a butanal permeation tube standard. For H.sub.3O+, the inventors used the ratio between peaks m/z 89 and 71 using a butanoic acid permeation tube standard. The values measured on the different days were within ±2% of the mean. Whenever required, the inventors optimized the voltage of the microchannel plate and the mass resolution (>1,500 m/Δm), using benzene (m/z 78 or 79 with NO.sup.+ or H.sub.3O.sup.+, respectively) as reference compound.
(26) Data Analysis
(27) Data were extracted using PTRMS viewer version 3.2.2.2 (Ionicon Analytik). Additional data analysis was conducted using in-house generated scripts written using R programming language.sup.16.
(28) Results
Example 1—Optimization of TD Cycle Duration
(29) Rapidity of analysis is a highly desirable characteristic for a method intended for large-scale clinical studies: for TD-PTR-ToF-MS the main limiting factor is the duration of the TD analytical cycle. The transition from two-stage to one-stage desorption allows for a first reduction in the TD cycle. Another step affecting TD cycle length is desorption time: the goal is to minimize its duration without compromising analyte (VOC) recovery. Using the permeation unit, eight VOCs (acetone, butanoic acid, pentanoic acid, hexanoic acid, phenol, propanal, butanal and decanal) were loaded onto the TD tubes at fixed concentrations. Concentrations ranged between a few ppbV to a few hundred ppbV, aiming to simulate a breath-relevant concentration for each of the compounds. The possibility to monitor VOC release online throughout desorption allowed an optimal time to be selected by means of a single experiment, without the need to test different desorption times separately. Experiments were conducted using H.sub.3O.sup.+ and NO.sup.+ as reagent ions (11 and 9 replicates, respectively). The mean release profiles in
Example 2—Analytical Performance of TD-PTR-TOF-MS
(30) Calibration curves allowed the assessment of linearity (R.sup.2=0.98-0.99) within a realistic concentration range for each of the VOCs of interest, see table 2. Slopes were reasonably close to unity (0.8-1.3), indicating that quantitative or near-quantitative VOC recovery was achieved. Limits of detection (LOD) and quantification (LOQ) were established by measuring blank TD tubes (n=45 for each of the primary ions). LOD and LOQ were fixed at 3-fold and 5-fold the background standard deviation and were in the order of 0.2-0.9 and 0.3-1.5 ppbV, respectively.
(31) TABLE-US-00002 TABLE 2 Limit of detection (LOD), limit of quantification (LOQ), calibration curve parameters and linearity assessed on seven selected compounds (n = 10) using TD-PTR-ToF-MS. Linear LOD.sup.a LOQ.sup.a range Compound (ppbV) (ppbV) Intercept Slope (ppbV) R.sup.2 Butanoic acid 0.6 0.9 −1.21 1.09 27-533 0.995 Pentanoic acid 0.2 0.3 0.03 1.04 5-83 0.995 Hexanoic acid 0.2 0.4 0.54 0.8 2-17 0.988 Propanal 0.8 1.3 −5.9 1.3 9-109 0.999 Butanal 0.9 1.5 −1.77 1.05 7-95 0.999 Phenol 0.5 0.9 0.44 0.94 3-41 0.997 Decanal 0.3 0.5 −1.24 1.25 2-21 0.999
(32) Repeatability was assessed by applying the same methodology: TD tubes were loaded at three different concentration levels for all compounds, repeating the procedure in duplicate on five different days. Coefficients of variation of repeated measurements were in the range 7-20% and recoveries remained higher than 83% for all tested VOCs. The results are shown in table 3.
(33) TABLE-US-00003 TABLE 3 Repeatability and mean recovery assessed on seven selected compounds (n = 10) Concentration Compound level Mean (±SD) Recovery (%) Butyric acid High 559 (±77) 92% Mid 377 (±62) 97% Low 54 (±4) 89% Pentanoic acid High 97 (±10) 91% Mid 46 (±6) 94% Low 12 (±1) 87% Hexanoic acid High 19 (±4) 100% Mid 11 (±1) 89% Low 5 (±1) 92% Propanal High 116 (±22) 94% Mid 79 (±14) 110% Low 13 (±1) 85% Butanal High 99 (±17) 95% Mid 44 (±6) 110% Low 10 (±1) 83% Phenol High 34 (±5) 104% Mid 19 (±1) 128% Low 4 (±0.5) 94% Decanal High 20 (±0.4) 122% Mid 5 (±0.1) 96% Low 3 (±0.6) 90%
Example 3—Comparison with TD-GC-MS
(34)
Example 4—Breath Analysis by TD-PTR-ToF-MS
(35) The findings obtained on authentic standards were compared to the results obtained by the analysis of breath samples obtained from a healthy volunteer. Tentative identifications were based on accurate mass determination (within ±0.01 mass-to-charge ratio). The perusal of online desorption profiles confirms that, for all considered compounds, the first 10 minutes of desorption encompass at least 90% of the total VOC release, see
(36) The optimized TD cycle has an overall duration of 15 minutes (10 min desorption+5 min between tube loading and unloading and leak test). This results in an expected throughput of nearly 100 samples for 24 hours of continuous operation. This was eventually confirmed by means of a platform usability study. Breath was sampled onto TD tubes using a breath collection device (Experimental Section) from 46 patients attending St Mary's Hospital (Paddington, London) for upper gastrointestinal endoscopy or surgery. Two TD tubes were collected for each patient and each tube was analysed by TD-PTR-ToF-MS, employing either H.sub.3O.sup.+ or NO.sup.+ as the reagent ion. The 92 tubes were processed by the analytical platform over 24 hours of continuous, unattended operation, thus corroborating the usability of the platform for large-scale clinical studies. Table 4 shows concentration ranges for the seven target VOCs. In addition to that, breath acetone was quantified in order to evaluate the platform performance on an abundant, well characterized breath VOC. Aldehydes, fatty acids and phenol remained well above the limit of quantification, with the exception of decanal, with 3 samples out of 46 lower than LOQ. The range of concentration established for acetone was similar to the values reported for adult volunteers using SIFT-MS.sup.19.
(37) TABLE-US-00004 TABLE 4 Platform usability: detection and quantification of 8 VOCs in 46 patients Minimum 1.sup.st quartile Median 3.sup.rd quartile Maximum Compound (ppbV) (ppbV) (ppbV) (ppbV) (ppbV) Acetone 33.6 161 274.8 523.5 1760.3 Butanoic 1.4 2.7 3.8 4.9 74.4 acid Pentanoic 0.7 1.5 2 2.4 3.5 acid Hexanoic 0.4 0.8 1.1 1.7 10.6 acid Phenol 2.2 5.9 7.6 10.2 21.4 Propanal 2.9 3.9 4.9 7.3 9.9 Butanal 1.4 1.9 2.6 3.5 5.4 Decanal 0.5.sup.a 0.7 0.8 0.9 1.9 .sup.aFor decanal, 3 samples were below LOQ
Example 5—Comparison with Two-Stage TD-PTR-ToF-MS
(38) Limits of detection (LOD) and quantification (LOQ) were calculated for four selected compounds using both the modified one-stage TD-PTR-ToF-MS device, according to the present invention, and the prior art unmodified two-stage TD-PTR-ToF-MS device. The methods used were the same as described in example 3. The results are shown in table 5.
(39) TABLE-US-00005 TABLE 5 Performance comparison between two-stage and one-stage TD setup Linearity Primary LOD LOQ Accuracy range Device Compound ion (ppbV) (ppbV) (%) (ppbV) One-stage Butanal NO.sup.+ 0.1 0.5 73-94% 5-102 Two-stage Butanal NO.sup.+ 1.1 3.7 91-117% 5-55 One-stage Phenol NO.sup.+ 0.6 2 125-103% 2-18 Two-stage Phenol NO.sup.+ n.a. n.a. n.a. n.a. One-stage Decanal NO.sup.+ 0.3 1 75-98% 1-16 Two-stage Decanal NO.sup.+ n.a. n.a. n.a. n.a. One-stage Butyric acid H.sub.3O.sup.+ 0.8 2.7 76-115% 20-647 Two-stage Butyric acid H.sub.3O.sup.+ 2.8 9.4 62-106% 44-360
(40) Accuracy was calculated by comparing the results with direct analysis. Linearity was calculated by undertaking a visual inspection of the graphs.
(41) Visual inspection of
(42) A comparison of the time required to analyse the contents of a single TD tube using both the two-stage and one-stage devices is given in table 6.
(43) TABLE-US-00006 TABLE 6 Comparison of analytical cycle duration between the two-stage and one-stage TD setup Step Two-stage One-stage TD tube purge 1′ — TD tube desorb 20′ 10′ CT purge 1′ — CT desorb 6′ — Other 6′ 6′ Cleaning 10′ — TOTAL 44′ 16′
(44) The 6 minutes listed under “other” comprises movement of mechanical parts and self-testing of the TD apparatus which is performed at each cycle.
CONCLUSION
(45) This research work addresses the development of a new analytical platform, based on TD-PTR-ToF-MS, and its applicability to breath analysis. This platform displays high throughput and sensitivity. Limits of detection and quantification were in the order of 0.2-0.9 and 0.3-1.5 parts per billion by Volume (ppbV), respectively. Analytical recoveries from TD tubes were 83% or higher and coefficients of variation were below 20% of mean values. When tested against the VOC panel, the platform showed better linearity and sensitivity than the currently available leading technology (i.e. TD-GC-MS). The usability of the platform was evaluated in the analysis of a set of breath samples of clinical origin, allowing for a throughput of nearly wo TD tubes for 24 hours of continuous operation. All these characteristics enhance the implementation of TD-PTR-ToF-MS for large-scale clinical studies.
(46) Using this approach, breath could be collected from multiple GP practices or clinical hubs, in a similar way to what is already done for blood samples today. Samples would then be analysed in a regional lab by means of TD-PTR-ToF-MS. The platform is based on a new coupling strategy that permits one-stage desorption. This appears to be better suited for coupling to a direct-injection MS instrument, such as PTR-ToF-MS or SIFT-MS. It must be noted that this coupling, however unprecedented, is easily reproducible using commercial instrumentation and readily available materials. Based on the inventors' preliminary comparison with TD-GC-MS, TD-PTR-ToF-MS appears to offer some advantages in terms of analytical performance.
(47) The main limitation of direct injection MS resides in the absence of a chromatographic separation step: hence, unambiguous VOC identification within complex matrices is not always possible. A possible solution could be represented by the simultaneous analysis of a subset of the samples by means of TD-GC-MS for cross-platform comparison and robust identification.
REFERENCES
(48) 1. Amann, A.; Miekisch, W.; Schubert, J.; Buszewski, B.; Ligor, T.; Jezierski, T.; Pleil, J.; Risby, T. Annu Rev Anal Chem (Palo Alto Calif.) 2014, 7, 455-482. 2. Smith, D.; Španěl, P.; Herbig, J.; Beauchamp, J. Journal of Breath Research 2014, 8 (2), 027101. 3. Markar, S. R.; Lagergren, J.; Hanna, G. B. BMJ Open 2016, 6 (1), e009139. 4. Chow, J. C.; Yu, J. Z.; Watson, J. G.; Hang Ho, S. S.; Bohannan, T. L.; Hays, M. D.; Fung, K. K. Journal of Environmental Science and Health, Part A 2007, 42 (ii), 1521-1541. 5. Crespo, E.; Devasena, S.; Sikkens, C.; Centeno, R.; Cristescu, S. M.; Harren, F. J. M. Rapid Communications in Mass Spectrometry 2012, 26 (8), 990-996. 6. Hryniuk, A.; Ross, B. M. International Journal of Mass Spectrometry 2009, 285 (1-2), 26-30. 7. Jordan, A.; Haidacher, S.; Hanel, G.; Hartungen, E.; Mark, L.; Seehauser, H.; Schottkowsky, R.; Sulzer, P.; Mark, T. D. International Journal of Mass Spectrometry 2009, 286 (2-3), 122-128. 8. Erickson, M. H.; Gueneron, M.; Jobson, B. T. Atmospheric Measurement Techniques 2014, 7 (1), 225-239. 9. Holzinger, R.; Williams, J.; Herrmann, F.; Lelieveld, J.; Donahue, N. M.; Röckmann, T. Atmospheric Chemistry and Physics 2010, 10 (5), 2257-2267. 10. Romano, A.; Hanna, G. B. Journal of Mass Spectrometry 2018, 10.1002/jms.4063. 11. Doran, S.; Romano, A.; Hanna, G. B. Journal of Breath Research 2017. 12. Kumar, S.; Huang, J.; Abbassi-Ghadi, N.; Španěl, P.; Smith, D.; Hanna, G. B. Analytical Chemistry 2013, 85 (12), 6121-6128. 13. Kumar, S.; Huang, J.; Abbassi-Ghadi, N.; Mackenzie, H. A.; Veselkov, K. A.; Hoare, J. M.; Lovat, L. B.; Španěl, P.; Smith, D.; Hanna, G. B. Annals of Surgery 2015, 262 (6), 981-990. 14. Markar, S. R.; Chin, S.-T.; Romano, A.; Wiggins, T.; Antonowicz, S.; Paraskeva, P.; Ziprin, P.; Darzi, A.; Hanna, G. B. Breath Volatile Organic Compound Profiling of Colorectal Cancer Using Selected Ion Flow-Tube Mass Spectrometry. Annals of Surgery 2017, 1. 15. Horváth, I.; Barnes, P. J.; Loukides, S.; Sterk, P. J.; Högman, M.; Olin, A.-C.; Amann, A.; Antus, B.; Baraldi, E.; Bikov, A.; Boots, A. W.; Bos, L. D.; Brinkman, P.; Bucca, C.; Carpagnano, G. E.; Corradi, M.; Cristescu, S.; de Jongste, J. C.; Dinh-Xuan, A.-T.; Dompeling, E.; Fens, N.; Fowler, S.; Hohlfeld, J. M.; Holz, O.; Jöbsis, Q.; Van De Kant, K.; Knobel, H. H.; Kostikas, K.; Lehtimäki, L.; Lundberg, J. O.; Montuschi, P.; Van Muylem, A.; Pennazza, G.; Reinhold, P.; Ricciardolo, F. L. M.; Rosias, P.; Santonico, M.; van der Schee, M. P.; van Schooten, F.-J.; Spanevello, A.; Tonia, T.; Vink, T. J. European Respiratory Journal 2017, 49 (4), 1600965. 16. R Core Team. R Foundation for Statistical Computing, Vienna, Austria. 2017. 17. Kloos, D.; Lingeman, H.; Mayboroda, O. A.; Deelder, A. M.; Niessen, W. M. A.; Giera, M. Analysis of Biologically-Active, Endogenous Carboxylic Acids Based on Chromatography-Mass Spectrometry. TrAC Trends in Analytical Chemistry 2014, 61, 17-28. 18. Schieberle, P.; Molyneux, R. J. Quantitation of Sensory-Active and Bioactive Constituents of Food: A Journal of Agricultural and Food Chemistry Perspective. Journal of Agricultural and Food Chemistry 2012, 60 (10), 2404-2408. 19. Španěl, P.; Dryahina, K.; Smith, D. Acetone, Ammonia and Hydrogen Cyanide in Exhaled Breath of Several Volunteers Aged 4-83 Years. Journal of Breath Research 2007, 1 (1), 011001.