Selected Ion Flow Tube-MS Analysis of ... - ACS Publications

Article pubs.acs.org/ac

Selected Ion Flow Tube-MS Analysis of Headspace Vapor from Gastric Content for the Diagnosis of Gastro-Esophageal Cancer Sacheen Kumar,†,‡ Juzheng Huang,†,‡ Julia R. Cushnir,† Patrik Španěl,¶ David Smith,§ and George B. Hanna*,† †

Department of Surgery and Cancer, Imperial College London, 10th Floor QEQM Wing, St Mary’s Hospital, London W2 1NY, United Kingdom ¶ J. Heyrovsky Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejskova 3, 18223 Prague 8, Czech Republic § Institute for Science and Technology in Medicine, Keele University, Guy Hilton Research Centre, Thornburrow Drive, Hartshill, Stoke-on-Trent ST4 7QB, United Kingdom ABSTRACT: Gastric content is a complex biofluid within the human stomach which has an important role in digestive processes. It is believed that gastric content may be a contributory factor in the development of upper gastro-intestinal diseases. In this work, selected ion flow tube mass spectrometry (SIFT-MS) has been applied to the quantification of volatile organic compounds (VOCs) in the headspace vapor of gastric content samples, which were retrieved from three groups of patients, including those with gastro-esophageal cancer, noncancer diseases of the upper gastro-intestinal tract, and a healthy cohort. Twelve VOCs have been investigated in this study; the following 7 VOCs, acetone, formaldehyde, acetaldehyde, hexanoic acid, hydrogen sulphide, hydrogen cyanide, and methyl phenol, were found to be significantly different between cancer and healthy groups by the Mann−Whitney U test. Receiver operating characteristics (ROC) analysis was applied for the combined VOCs of acetaldehyde, formaldehyde, hydrogen sulphide, and methyl phenol to discriminate cancer patients from healthy controls. The area under the curve (AUC) was 0.9. This result raises the prospect that a VOC profile rather than a single biomarker may be preferable in the molecular-orientated diagnosis of gastro-oseophageal cancer, and this warrants further investigation to assess its potential application as a new diagnostic test.

V

concentration of dimethyl disulfide was statistically significantly lower in cancer patients compared to healthy controls, while concentrations of ketones, phenols, and benzene derivatives were found to be higher in the cancer group when compared to the healthy control group.3,4 Xue et al. used SPME/GC/MS to investigate the blood VOCs linked with liver cancer. Nineteen blood VOCs were found to be different between the liver cancer and the control groups (p < 0.05). Of these 19 compounds, 3 (hexanal, 1-octen-3-ol, and octane) were found to be present at much higher concentrations in blood of the liver cancer group compared to the control group.5 These studies all demonstrate promising associations between certain VOCs and cancer states in a variety of biological samples. However, there are several important considerations for the application of GC/MS in the analysis of biological samples. It remains an offline technique requiring appropriate GC-column selection and preanalysis sample preparation. Analysis times are lengthy and VOCs with low molecular weight (some of which

olatile organic compounds (VOCs) emitted from the human body have been of interest to researchers for over four decades. In 1971, Nobel-prize winner Linus Pauling demonstrated that breath is a complex mixture containing about 250 VOCs; he also showed that urine vapor has over 280 VOCs.1 Since that time, research in this field has focused on identifying VOCs associated with certain disease states, the ultimate translational application for this research being the quantification of VOCs as a noninvasive diagnostic and therapeutic monitoring tool in medicine. Several chemical analytical techniques have been employed to study VOCs in disease, the most common being gas chromatography mass spectrometry (GC/MS). One of the most common disease states that has been studied in relation to VOC analysis is cancer. Phillips et al. studied the exhaled breath of lung cancer patients and controls and reported that alkanes, alkane derivatives, and benzene derivatives were able to differentiate those with and without lung cancer.2 GC/MS coupled with sample collection and concentration techniques, in particular solid phase microextraction, SPME, has further enhanced the isolation and identification of VOCs from mixed samples. Using SPME/GC/MS to analyze urine, Silva et al. showed that © 2012 American Chemical Society

Received: August 20, 2012 Accepted: October 4, 2012 Published: October 4, 2012 9550

dx.doi.org/10.1021/ac302409a | Anal. Chem. 2012, 84, 9550−9557

Analytical Chemistry

Article

referral form contains information explaining the indication for the procedure and the patient’s known medical history. These referral forms were screened, and our exclusion criteria included patients with liver disease and any small bowel or colonic conditions. Patients with a biopsy-confirmed invasive gastric or esophageal cancer were included in the cancer cohort. Patients with any noncancer conditions of the upper gastrointestinal tract (e.g., esophagitis, gastritis, and peptic ulcer disease) were included in the positive control cohort. Patients who had a normal OGD test and were H. pylori negative were included in the “healthy” cohort group. Local ethics committee approval was obtained for this study. Patients eligible for the study were given information sheets pertaining to the study, and informed consent was obtained prior to enrollment in the study. Gastric Content Sampling. All patients were fasted for a minimum of 6 h as standard protocol for an OGD. All gastric content samples were taken from patients undergoing OGD. A gastroscope (Olympus, Essex, UK) is introduced through the mouth and then into the esophagus. The gastroscope is passed via the esophagus into the stomach. During this procedure, one is able to take biopsies, and fluid can also be retrieved through a suction channel within the gastroscope. The gastric content retrieved during a procedure was captured in a disposable suction jar. This suction jar was then taken to the SIFT-MS instrument, and all samples were analyzed within 1 h. The pH of gastric content samples was measured immediately after SIFT-MS analysis. SIFT-MS Analysis. The experiments in this study were conducted using a Prof ile-3 SIFT-MS instrument (Instrument Science, Crewe, UK).18 The principle of SIFT-MS has been extensively explained in the literature.9,10,19 Thus, only a brief explanation of the SIFT-MS principle is described in this report. Selected precursor ions are formed in a microwave discharge source and are selected according to their mass-tocharge ratio, m/z, by a mass filter and injected into flowing helium carrier gas where they are convected as a thermalized swarm along a flow tube. H3O+, NO+, and O2+ precursor (reagent) ions are used to ionize the trace gases in an air/vapor sample that is introduced into the helium at a known flow rate (20 mL/min in the present study), as these ions do not react with the major compounds in air; they only selectively ionize the volatile and trace compounds present within a the sample producing characteristic product (analyte) ions. The flow rate of the ion swarm/carrier gas determines the reaction time. By measuring the count rate of both precursor ions and the characteristic product ions at the downstream spectrometer/ detection system, a real-time quantification is achieved, realizing absolute concentration of trace and volatile compounds at the parts-per-billion by-volume (ppbv) or parts-per-million byvolume level (ppmv).20 SIFT-MS operates in two different modes. In the full-scan mode (FS), the ambient air/breath or liquid headspace vapors are introduced at known flow rates into the carrier gas, and the downstream analytical mass spectrometer system is scanned over predetermined m/z ranges for a given time period. Concentrations of VOCs and trace gases are calculated from the count rates of the characteristic product ions using a standardized kinetics library.18,20 In the multi-ion monitoring mode (MIM), the downstream analytical mass spectrometer is rapidly switched between selected m/z values to target selected trace gas species. From the data obtained, together with the rate coefficients of

have been muted as potential biomarkers) can be difficult to accurately quantify with GC/MS. The advent of new ionization techniques such as atmospheric pressure chemical ionization (APCI),6 proton transfer reaction mass spectrometry (PTRMS),7 and selected ion flow tube mass spectrometry (SIFTMS)8 has allowed one to perform real-time, online sample analysis. In comparison to GC techniques, SIFT-MS has the advantage of real-time VOC quantification without the need for any preconcentration steps. The potential of SIFT-MS as a noninvasive clinical tool has been demonstrated in studies involving patients with end-stage renal disease and cystic fibrosis.9−11 As well as exhaled breath analysis, SIFT-MS has been employed to detect and quantify the VOCs within the headspace of biofluids; analysis of urine headspace has been carried out to assess for urinary infection,12 and elevated levels of formaldehyde were detected in the headspace of urine from patients with bladder and prostate cancer.13 In the UK, gastro-esophageal cancer remains a disease with poor patient outcomes. In 2009 in the UK, 15,656 people were diagnosed with gastro-esophageal cancer.14 The red flag symptoms often occur at a more advanced stage of the disease, and therefore, people with such cancers most commonly go to their doctor later in the disease process. This delay in diagnosis results in only 20% of patients being suitable for potentially curative treatment at first presentation. Between 2005 and 2009, the five year survival rate in the UK for esophageal cancer was only 13%,14 this statistic highlighting the poor prognosis associated with this disease. One of the main methods for improving these survival rates lies in the earlier detection of these cancers, and VOC analysis represents a promising area to explore in relation to gastro-esophageal cancer. Biofluids obtained from organ-specific cavities have much potential in assisting in the identification of disease biomarkers. In the literature, biofluids which have been investigated include urine, serum, plasma, saliva, sweat, and bronchoalveolar lavage.15−17 In this study, we selected gastric content as our biofluid of choice for investigation. Gastric content lies within in the stomach and is a combination of gastric acid, saliva, and bile reflux. Specifically in gastro-esophageal cancer, the gastric content comes into direct contact with cancer cells. We hypothesize that the VOCs of the gastric content retrieved from patients with gastro-esophageal cancer differs from patients with benign upper gastro-intestinal conditions and those with healthy upper gastro-intestinal tracts. In this study, we report the first investigation of VOCs through headspace analysis of gastric content using SIFT-MS. The comparative analysis of gastric content vapor metabolites was conducted with a gastroesophageal cancer cohort of 19 patients, a positive control group (i.e., those with noncancer diseases of the upper gastrointestinal tract) of 12 patients, and a “healthy” group of 11 patients (i.e., those with no diseases present within the upper gastro-intestinal tract).

■

EXPERIMENTAL SECTION Patient Selection. Patients for this study were recruited through St. Mary’s Hospital, Paddington (Imperial College London) and via the North West Thames regional gastroesophageal cancer network (UK). Patients referred with upper gastro-intestinal symptoms are commonly sent for an esophago-gastro-duodenoscopy (OGD) by their clinician. An OGD is a routinely performed camera test in which the upper gastro-intestinal tract is visualized (a description of the procedure and sample retrieval is provided below). An OGD 9551



■

Article

RESULTS AND DISCUSSION Analysis of Full Scan Spectra. Full scan (FS) spectra were obtained when the headspace vapor of gastric content were introduced into the helium carrier gas. Representative spectra using H3O+ and NO+ precursor ions are shown in Figure 1. A

the reactions as included in the kinetics library, specifically targeted volatile compounds are quantified in real-time. In the current study, the FS mode of SIFT-MS was initially operated to reveal which characteristic product ions were present in the headspace vapor of the gastric content. However, the MIM mode was mainly employed for this study as it is more sensitive and reproducible than the FS mode21 and allows one to perform targeted analysis. For each measurement, a total of 10 mL of gastric content from the original sample jar was aliquoted into a standard 60 mL specimen jar. The specimen jar was sealed with parafilm and then placed in an oven at 37 °C for five minutes prior to analysis. A 21g Luer-lock sterile hypodermic needle was connected to a 200 cm Vygon extension tube. The extension tubing was directly connected to the SIFT-MS sampling port. The sterile needle end punctured the parafilm layer of the specimen jar to sample the headspace above the gastric content liquid. The headspace vapor automatically flowed into the helium carrier gas at a fixed rate through the sampling line which was set at a temperature of 80 °C, and the analysis was conducted. Throughout the analysis, the specimen jar, needle, and extension tubing were all kept in the oven at 37 °C. For the FS mode, two serials of 9 × 60 s repeated scans were performed using H3O+ and NO+ precursor ions. The O2+ precursor ion was not used for the FS analysis, as it is more energetic than the other precursor ions resulting in spectra that are more difficult to interpret.21 For the MIM mode, selected VOCs from the vapor samples were analyzed for a total of 60 s, and the measured concentrations were averaged over the analysis time for each VOC. Statistical Analysis. Statistical analysis was performed using IBM SPSS statistics 19 (SPSS Inc., Chicago, IL). Mann− Whitney U test was used to compare the measured concentrations of VOCs between cancer patients and healthy controls. A p-value ≤0.05 was taken as the level to indicate statistical significance, in keeping with most clinical studies. With a test being described as significant at 5%, this indicates that there is less than a 1 in 20 chance that a difference as large as that observed in the study could have arisen by chance if there is no true difference. Receiver operating characteristic curves (ROC) is a fundamental test to assess the accuracy of a diagnostic test to discriminate between a diseased group and healthy group.22,23 In a ROC curve, the sensitivity is plotted against (1-specificity) using the variation to differentiate between the two groups. The cutoff points on the ROC curve represent when a discrimination threshold is reached with a paired sensitivity and specificity. The area under curve (AUC) is a measure of how accurate the test is in separating patients with the disease from healthy people. The higher the AUC value, the better the ability of the test to positively identify those persons with the disease and correctly identify those persons without the disease. A diagnostic test is considered highly accurate when an area under the curve (AUC) > 0.9 is obtained.24 In the current study, after Mann−Whitney U test, the statistically significant VOCs which showed a difference between the cancer and healthy cohorts were used as included variables for the ROC curve. To construct the ROC curve, the disease conditions were used as the dependent variable, and the VOCs (which showed statistically significant differences in Mann−Whitney U test) were used as the independent variable. A binary logistic regression method was conducted to test different combinations of selected VOCs as the variables to find the model with the highest R2 value and AUC value for the ROC curve.

Figure 1. Full scan spectra of the headspace of gastric content using (a) H3O+ precursor ion; (b) NO+ precursor ion. The individual characteristic product ions and their hydrates were assigned to the analyzed molecules. In (a), m/z 19, 37, 55, and 73 are H3O+(H2O)0,1,2,3; m/z 59, 77, acetone; m/z 33, 51, 69, methanol; m/z 47, 65, and 83, ethanol; m/z 45, 81, and 83, acetaldehyde; m/z 31, formaldehyde; m/z 35, hydrogen sulphide; m/z 28, hydrogen cyanide; m/z 95, phenol; m/z 18, 36, and 54, ammonia. In (b), m/z 19, 37, 55, and 73 are H 3O+(H2O)0,1,2,3; m/z 30, 48, 66, and 84 are NO+(H2O)0,1,2,3; m/z 90 is acetic acid.

variety of VOCs were detected and identified using the full scan (FS) mode of SIFT-MS from their characteristic product ions. From full scan spectra, those VOCs showing overlapped product ions were not selected for MIM mode analysis. For instance, within the carboxylic acid group, propanoic acid, pentanoic acid, and butanoic acid were not selected for further MIM investigation as their product ions shared overlapping m/ z values and these would be difficult to quantify accurately. GC/MS is often not capable of detecting large volatile compounds such as long chain aldehydes and the smaller, trace volatile organic compounds (H2S and HCN).25 Hydrogen sulphide and hydrogen cyanide were unambiguously identified from their product ions using SIFT-MS. These compounds are considered equally important as those large volatile compounds; these compounds were investigated using MIM mode analysis. Reproducibility and Validity of the Sampling Method. Using the methods described above, the reproducibility of headspace analysis of gastric content was assessed. A total of ten gastric content samples was analyzed, and with each sample, three MIM measurements for selected VOCs were taken. The coefficient of variation (CV) was calculated for the measured concentration of several VOCs from the headspace vapor of the gastric content. These CV values in percent are: acetaldehyde, 6.7%; formaldehyde, 11%; acetone, 6.8%; hexanoic acid, 10.7%; methanol, 15.9%; ethanol, 6.2%; phenol, 22.6%; methyl phenol, 13.4%; hydrogen sulphide, 44.7%; hydrogen cyanide, 22.7%; ammonia, 6.8%. These coefficients of variation demonstrate the reproducibility of the MIM measurement. In contrast to the 9552



Article

Table 1. Summary of Analytical Information of Volatile Organic Compounds Analyzed by the MIM of SIFT-MS Using H3O+, NO+, and O2+ Precursor Ions compounds

molecular formula

precursor ions

m/z

characteristic product ions

references

methanol ethanol acetic acid hexanoic acid acetaldehyde formaldehyde acetone acetone hydrogen sulphide phenol methyl phenol ammonia hydrogen cyanide

CH4O C2H6O C2H4O2 C6H12O2 C2H4O CH2O C3H6O C3H6O H2S C6H6O C7H8O NH3 HCN

H3O+ H3O+ NO+ H3O+ H3O+ H3O+ H3O+ O2+ H3O+ H3O+ O2+ O2+ H3O+

33, 51, 69 47, 65, 83 90, 108 117, 135 45, 81 31 59, 77 43, 58 35, 53 95, 113 108, 126 17 28

CH5O+, CH5O+(H2O), CH5O+(H2O)2 C2H7O+,C2H7O+(H2O), C2H7O+(H2O)2 NO+·CH3COOH, NO+·H2OCH3COOH C6H12O2H+,C6H12O2H+(H2O) CH3CHOH+, CH3CHOH+(H2O)2 CH2OH+ C3H6OH+, C3H6OH+(H2O) C3H6O+, C2H3O+ H3S+, H3S+(H2O) C6H6OH+, C6H6OH+(H2O) C7H8O+, C7H8O+(H2O) NH3+ H2CN+

26 26 27 26 28 28 28 28 21, 29 30 30 12 31

Figure 2. Box-whisker plots of the headspace median concentrations and interquartile ranges (parts-per-billion by-volume, ppbv) of (a) acetaldehyde, (b) formaldehyde, (c) acetone, (d) acetic acid, (e) hexanoic acid, and (f) pH measurements of the gastric content of cancer, positive control, and healthy groups. Acetaldehyde and formaldehyde and hexanoic acid were analyzed using H3O+ precursor ions; acetone was analyzed using O2+ ions, and acetic acid was analyzed using NO+ precursor ions.

present in the headspace of gastric content were analyzed using SIFT-MS. There is a decreasing trend in the median concentrations of acetaldehyde when comparing cancer, positive control, and healthy cohorts (Figure 2). The acetaldehyde concentrations of the cancer and positive control patients have similar median values (67 and 54 ppbv) and data distribution, as shown in Table 2 and Figure 2. In comparison, the median value from healthy group (24 ppbv) is 64.2% less than that of the cancer group. Furthermore, the majority of acetaldehyde values in the healthy cohort are less than 50 ppbv, while in the cancer and positive control groups, the majority of observed values are above 50 ppbv. Acetaldehyde is a metabolite of ethanol oxidation, but in recent years, it has been classified as a Class I carcinogen. Several studies have postulated that deficiencies in enzymes involved in aldehyde metabolism may contribute to upper gastro-intestinal carcinogenesis; Oikawa et al. demonstrated that those deficient in aldehyde dehydrogenase II (ALDH2) in the presence of gastric hypochlorhydria are at increased risk of esophageal squamous

major metabolites (e.g., acetone, acetaldehyde), the higher CV for hydrogen sulphide and hydrogen cyanide is expected given their extremely low concentrations in gastric content and hence the low count rates of their characteristic product ions.20 MIM Analysis of Selected VOCs in the Headspace of Gastric Content. As shown in Table 1, a total of 12 volatile compounds released from the headspace of gastric content were analyzed using the MIM mode of the SIFT-MS. The analytical information, including chemical formula, precursor ions, m/z ratio, and characteristic product ions are given in the table. H3O+ is the most used precursor ion as it ionizes more chemical functional groups than the other two precursor ions. However, NO+ and O2+ have also been employed for specific molecules. In these instances, they have an advantage in accurately detecting and quantifying these compounds over H3O+ (for which there are known overlaps with other product ions). Aldehydes and Ketones. Three short-chain carbonyl compounds, i.e., acetaldehyde, formaldehyde, and acetone, 9553



Article

Table 2. Summary of Median Concentrations and Interquartile Ranges for Those VOCs That Were Significantly Different between Cancer and Healthy Groupsa cancer samples

healthy samples

positive controls

VOCs

precursor ion

median

interquartile range

median

interquartile range

median

interquartile range

acetaldehyde acetone formaldehyde acetic acid hexanoic acid hydrogen sulphide hydrogen cyanide methanol acetone methyl phenol

H3O+ H3O+ H3O+ NO+ H3O+ H3O+ H3O+ H3O+ O2+ O2+

67 69 48 37 35 1.9 14.3 70 296 16

[39−105] [17−253] [40−64] [25−71] [12−269] [0.7−5.5] [5.8−18.1] [47−141] [154−525] [10−20]

24 12 68 21 12 0.6 5.7 48 84 9

[17−56] [9−27] [49−77] [16−45] [9−13] [0.3−1.1] [3.5−9.8] [39−72] [49−130] [6−15]

54 23 52 28 11 1.4 5.8 43 153 11

[24−90] [12−75] [40−79] [18−32] [10−17] [0.7−2.0] [2.3−10.2] [34−58] [79−217] [7−16]

a

The results of the positive control group are included for comparison. The concentrations are measured in parts-per-billion by-volume, ppbv. The cancer group included all patients with a biopsy-confirmed gastro-esophageal cancer. The positive control group included all patients diagnosed with noncancer diseases of the upper gastro-intestinal tract. The healthy group included all persons who had undergone an OGD which demonstrated no diseases of the upper gastro-intestinal tract.

cell carcinoma.32 Linderborg et al. reported that reducing acetaldehyde exposure in those with an achlorhydric stomach may be an important factor in reducing the risk of gastric cancer.33 For formaldehyde, an increasing trend in median value from the cancer, positive control, and healthy groups is apparent. Formaldehyde has been associated with cancer states to varying degrees in previously reported studies. Fuchs et al. studied breath gas aldehydes using GC/MS in patients with lung cancer, smokers, and healthy volunteers.34 Concentrations of exhaled formaldehyde were significantly lower in smokers when compared to healthy volunteers and lung cancer patients. There was no difference observed between lung cancer patients and healthy volunteers in this study. Španěl et al. studied formaldehyde in the headspace of urine from bladder cancer, prostate cancer, and healthy controls.13 This study demonstrated that formaldehyde was clearly elevated in the headspace of urine from cancer patients when compared to healthy controls. The varying findings associated with formaldehyde in different biofluids and breath warrants further investigation. Acetone, the simplest of the ketones, was measured using both H3O+ and O2+ by SIFT-MS. It is the most abundant ketone body (the other two being beta-hydroxybutyric acid and acetoacetic acid) which is known as a product of lipolysis.35 Elevated acetone levels have been observed in people in starving states and specifically in the condition of diabetic ketoacidosis.36−39 There is a decreasing trend in acetone median values from the cancer, positive control, and healthy groups (Figure 2). Patients in the cancer cohort demonstrated the highest concentrations of acetone. As shown in Figure 2, acetone values in the cancer group also have a much wider interquartile range of 154−525 ppbv in contrast to the positive control and healthy groups. This observed increase in acetone may be related to the patient’s cancer status or may reflect a change in their underlying metabolism in response to disease. Patients with gastro-esophageal cancer occasionally have difficulty in maintaining an adequate oral intake, and given their increased energy requirements, this can also result in a reliance on ketogenic metabolic processes. Acids. Two organic acids were measured using SIFT-MS in the headspace of gastric content. Acetic acid is a metabolic intermediate for the generation of the tioester acetyl-CoA. Its presence has been shown in the exhaled breath of lung cancer

patients using SIFT-MS.40 As shown in Figure 2, acetic acid in the headspace of gastric content from the cancer cohort has a higher median value and wider interquartile range, in contrast to the other 2 groups. The median value in the cancer group is found to be higher than that in the healthy group, i.e., 37 and 21 ppbv. Hexanoic acid has been reported as a potential biomarker in a study of urine from patients with leukemia, lymphoma, and colorectal cancer.3 As shown in Figure 2, hexanoic acid is found to be significantly different between cancer and the other two control groups. Both positive and healthy control groups have similar medians and interquartile ranges. The median concentration of hexanoic acid in the cancer group is 35 ppbv, i.e., approximately 3 times greater than the other two groups. As shown in Figure 2, compared to the positive and healthy controls, a much wider distribution of hexanoic acid values in the cancer group are observed (interquartile range 12−269 ppbv). The elevated levels of hexanoic acid observed within the cancer group require further investigation into its source, as it may represent a potential VOC biomarker for gastro-esophageal cancer. The observed differences in the headspace acid concentrations between the groups led us to analyze the pH of the gastric content samples. It was hypothesized that the higher acid concentrations would correlate with a lower pH of the gastric content. Gastric content can have a variable pH range based on the respective contribution from gastric acid and bile reflux. This variation in pH, and specifically the hypoacidity, is believed to contribute to the pathogenesis of upper gastrointestinal diseases. Lu et al. investigated gastric content acidity and found that both gastric cancer and gastric ulcer patients had a higher pH (6.6 and 3.4 vs 2.9) than healthy controls.41 As shown in Figure 2, the gastric content from the healthy group has a median pH value of 1.93. The median pH of gastric content from the positive group is 5.67, and the median value for the cancer group is 6.18. The cancer cohort also has the widest pH range among the three groups. From a chemistry perspective, one would postulate that a low pH would facilitate the release of volatile organic acids into the headspace vapor. However, the results from this study do not support this theory. Figure 2 shows that both the cancer and positive control groups (with higher median pH values) have higher measured concentrations of acetic acid and hexanoic acid than the healthy controls. The pH of a biofluid represents one variable 9554



Article

Figure 3. Box-whisker plots of the headspace median concentrations and interquartile ranges (parts-per-billion by-volume, ppbv) of (a) methanol, (b) methyl phenol, (c) hydrogen sulphide, and (d) hydrogen cyanide from the gastric contents of cancer, positive control, and healthy groups. Methanol, hydrogen sulphide, and hydrogen cyanide were analyzed using H3O+ ions, and methyl phenol was analyzed using O2+ precursor ions in the SIFT-MS.

that may contribute to the measured volatile organic acid concentrations; however, other factors such as bacteria (e.g., H. pylori) or cancer cells may have a greater influence in gastric content. Our group is currently undertaking further investigations to fully understand these observed results. Alcohol and Phenol Derivatives. Methanol, ethanol, phenol, and methyl phenol were investigated in the headspace of the gastric content. Methanol is highly volatile and can be difficult to detect by GC/MS. The measured median values for methanol demonstrate a decreasing trend from the cancer cohort to the healthy group and positive control cohort (Figure 3). Methanol concentration from gastric content of the cancer cohort shows the widest distribution in contrast to the other two groups. Ethanol levels within the headspace of gastric content were also measured during this study; however, their values are difficult to interpret given the potential effects of exogenous ethanol sources within the hospital environment (e.g., disinfectant hand gel). As shown in Figure 3, the distribution of methyl phenol concentrations is similar among the three groups. Methyl phenol concentration from gastric content in the cancer group is found to have the largest median value (16 ppbv), followed by the positive control group (11 ppbv), and the healthy group (9 ppbv). Sulfur and Nitrogen-Containing Compounds. H2S and HCN are highly volatile and have low molecular weights. SIFTMS has an advantage in analyzing these types of compounds in comparison to GC/MS. The measured H2S and HCN levels showed large variation between cancer, positive control, and healthy groups. As shown in Figure 3 and Table 2, the median concentrations and interquartile ranges of H2S showed a decreasing trend from the cancer cohort to the healthy group. The HCN median value is 14.3 ppbv for the cancer group which is significantly higher than the other two groups. The measured median concentrations for positive control and healthy groups were found to be similar, i.e., 5.8 and 5.7 ppbv, respectively.

The other volatile organic compounds shown in Table 1 did not show any significant differences between the groups. Therefore, no further investigations of these compounds were conducted in this study. Statistical Analysis. Mann−Whitney U test (p-value ≤0.05) was performed for all the volatile organic compounds listed in Table 1, to assess any differences in median concentrations between gastro-esophageal cancer patients and people with healthy upper gastro-intestinal tracts. Seven metabolites were found to be statistically significantly different across the two groups, including acetaldehyde (p = 0.027), acetone (p = 0.002), hexanoic acid (p = 0.010), hydrogen sulphide (p = 0.024), hydrogen cyanide (p = 0.035), formaldehyde (p = 0.041), and methyl phenol (p = 0.027). Their median concentrations and interquartile ranges from cancer, healthy, and positive control groups are given in Table 2. Using the statistically significant VOCs identified in the Mann−Whitney U test, binary logical regression analysis was performed to construct a discrimination model. The sum concentrations of acetaldehyde, formaldehyde, hydrogen sulphide, and methyl phenol demonstrated the highest area under curve (AUC) for diagnosing cancer with a value of 0.9. Figure 4 demonstrates a high degree of discrimination between gastro-esophageal cancer and healthy controls with reasonable sensitivity and specificity using the aforementioned compounds. This result raises the prospect that a VOC profile rather than a single biomarker may be preferable in the molecular-orientated diagnosis of gastro-oseophageal cancer. However, further investigations are necessary to fully understand the reported results and the proposed hypothesis.

■

CONCLUSIONS This is the first study of the headspace analysis of gastric content from patients with gastro-esophageal cancer, noncancer diseases of the upper gastro-intestinal tract, and those with 9555



■

ACKNOWLEDGMENTS

■

REFERENCES

Article

The authors express their gratitude to British Research Council (BRC) for funding and to Ms. Linghan Qi for statistical advice.

(1) Pauling, L.; Robinson, A. B.; Teranishi, R.; Cary, P. Proc. Natl. Acad. Sci. U. S. A. 1971, 68, 2374−2376. (2) Phillips, M.; Gleeson, K.; Hughes, J. M. B.; Greenberg, J.; Cataneo, R. N.; Baker, L.; McVay, W. P. Lancet 1999, 353, 1930− 1933. (3) Silva, C. L.; Passos, M.; Câmara, J. S. Br. J. Cancer 2011, 105, 1894−1904. (4) Silva, C. L.; Passos, M.; Câmara, J. S. Talanta 2012, 89, 360−368. (5) Xue, R.; Dong, L.; Zhang, S.; Deng, C.; Liu, T.; Wang, J.; Shen, X. Rapid Commun. Mass Spectrom. 2008, 22, 1181−1186. (6) Taylor, A. J.; Linforth, R. S. T.; Harvey, B. A.; Blake, A. Food Chem. 2000, 71, 327−338. (7) Lindinger, W.; Hansel, A.; Jordan, A. Int. J. Mass Spectrom. Ion Processes 1998, 173, 191−241. (8) Španěl, P.; Smith, D. Eur. J. Mass Spectrom. 2007, 13, 77−82. (9) Smith, D.; Španěl, P. Mass Spectrom. Rev. 2005, 24, 661−700. (10) Smith, D.; Španěl, P. Rapid Commun. Mass Spectrom. 1996, 10, 1183−1198. (11) Smith, D.; Španěl, P. Analyst 2007, 132, 390−396. (12) Smith, D.; Španěl, P.; Holland, T. A.; Al Singari, W.; Elder, J. B. Rapid Commun. Mass Spectrom. 1999, 13, 724−729. (13) Španěl, P.; Smith, D.; Holland, T. A.; Singary, W. A.; Elder, J. B. Rapid Commun. Mass Spectrom. 1999, 13, 1354−1359. (14) Cancer Statistics registrations: Registrations of cancer diagnosed in 2009; Office for National Statistics: London, 2012. (15) Martin, H. J.; Riazanskaia, S.; Thomas, C. L. P. Analyst 2012, 137, 3627−3634. (16) Cheng, W. H. J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng. 2008, 43, 98−104. (17) Julák, J.; Stránská, E.; Rosová, V.; Geppert, H.; Španěl, P.; Smith, D. J. Microbiol. Methods 2006, 65, 76−86. (18) Smith, D.; Pysanenko, A.; Španěl, P. Int. J. Mass Spectrom. 2009, 281, 15−23. (19) Španěl, P.; Smith, D. Med. Biol. Eng. Comput. 1996, 34, 409− 419. (20) Španěl, P.; Dryahina, K.; Smith, D. Int. J. Mass Spectrom. 2006, 230, 249−250. (21) Olivares, A.; Dryahina, K.; Navarro, J. L.; Flores, M.; Smith, D.; Španěl, P. Anal. Chem. 2010, 82, 5819−5829. (22) Metz, C. E. Semin. Nucl. Med. 1978, 8, 283−298. (23) Zweig, M. H.; Campbell, G. Clin. Chem. 1993, 39, 561−577. (24) Song, G.; Qin, T.; Liu, H.; Xu, G. B.; Pan, Y. Y.; Xiong, F. X.; Gu, K. S.; Sun, G. P.; Chen, Z. D. Lung Cancer 2010, 67, 227−231. (25) Davis, B. M.; Senthilmohan, S. T.; Wilson, P. F.; McEwan, M. J. Rapid Commun. Mass Spectrom. 2005, 19, 2272−2278. (26) Španěl, P.; Smith, D. Int. J. Mass Spectrom. Ion Processes 1997, 167−168, 375−388. (27) Smith, D.; Wang, T.; Španĕl, P. Rapid Commun. Mass Spectrom. 2003, 17, 2655−2660. (28) Španěl, P.; Ji, Y.; Smith, D. Int. J. Mass Spectrom. Ion Processes 1997, 165−166, 25−37. (29) Španel, P.; Smith, D. Rapid Commun. Mass Spectrom. 2000, 14, 1136−1140. (30) Wang, T.; Španěl, P.; Smith, D. Int. J. Mass Spectrom. 2004, 239, 139−146. (31) Španěl, P.; Wang, T.; Smith, D. Rapid Commun. Mass Spectrom. 2004, 18, 1869−1873. (32) Oikawa, T.; Iijima, K.; Koike, T.; Uno, K.; Horii, T.; Iwai, W.; Abe, Y.; Asano, N.; Imatani, A.; Shimosegawa, T. Scand. J. Gastroenterol. 2010, 45, 1338−1344. (33) Linderborg, K.; Marvola, T.; Marvola, M.; Salaspuro, M.; Färkkilä, M.; Väkeväinen, S. Alcohol.: Clin. Exp. Res. 2011, 35, 516− 522.

Figure 4. Receiver operating characteristics (ROC) curve of discrimination result for the diagnosis of gastro-esophageal cancer from healthy control using the sum concentrations of acetaldehyde, formaldehyde, hydrogen sulphide, and methyl phenol exhibited. The integrated area under ROC curve (AUC) is 0.9.

normal upper gastro-intestinal tracts. This is also the first study to demonstrate the applicability of SIFT-MS in the chemical analysis of gastric content. In the current study, SIFT-MS also offers a real-time analysis that avoids sample preanalysis treatment, which is particularly desirable in a hospital setting. A total of 12 VOCs have been analyzed in the headspace of gastric content obtained during gastroscopy from patients with gastro-intestinal disease and normal upper gastro-intestinal tracts. Of these compounds, the headspace concentrations of seven VOCs (acetone, formaldehyde, acetaldehyde, hexanoic acid, hydrogen suphide, hydrogen cyanide, and methyl phenol) were statistically different between cancer groups and the healthy cohort. We believe that gastric content represents a good surrogate biofluid to improve our understanding of the pathogenesis of upper gastro-intestinal diseases. By analyzing the VOCs in the headspace of gastric content, we postulate that VOCs associated with gastro-esophageal cancer can be identified which may contribute to the development of new diagnostic tests. Specifically, SIFT-MS represents a technique that can be readily used for targeted analysis of trace volatile compounds in human biofluids. However, further studies are necessary to understand the biochemical basis of the observed differences in the VOCs in the gastric content of the gastroesophageal cancer cohort.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +44(0)20 33122124. Fax: +44 (0) 20 3312 6309. E-mail: [email protected]. Author Contributions

‡ S.K. and J.H. are joint first authors and they contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Notes

The authors declare no competing financial interest. 9556



Article

(34) Fuchs, P.; Loeseken, C.; Schubert, J. K.; Miekisch, W. Int. J. Cancer 2010, 126, 2663−2670. (35) Laffel, L. Diabetes/Metab. Res. Rev. 1999, 15, 412−426. (36) Španěl, P.; Dryahina, K.; Rejšková, A.; Chippendale, T. W. E.; Smith, D. Physiol. Meas. 2011, 32, N23. (37) Smith, D.; Španěl, P.; Davies, S. J. Appl. Physiol. 1999, 87, 1584− 1588. (38) Manolis, A. Clin. Chem. 1983, 29, 5−15. (39) Crofford, O. B.; Mallard, R. E.; Winton, R. E.; Rogers, N. L.; Jackson, J. C.; Keller, U. Trans. Am. Clin. Climatol. Assoc. 1977, 88, 128−139. (40) Smith, D.; Španêl, P.; Sulé-Suso, J. Expert Rev. Mol. Diagn. 2010, 10, 255−257. (41) Lu, P. J.; Hsu, P. I.; Chen, C. H.; Hsiao, M.; Chang, W. C.; Tseng, H. H.; Lin, K. H.; Chuah, S. K.; Chen, H. C. World J. Gastroenterol. 2010, 16, 5496−5501.

9557


Selected Ion Flow Tube-MS Analysis of ... - ACS Publications

Recommend Documents