If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Amsterdam University Medical Centres, Department of Gastroenterology and Hepatology, Amsterdam, the NetherlandsAmsterdam Gastroenterology Endocrinology Metabolism (AGEM) Research Institute, Amsterdam, the Netherlands
Amsterdam University Medical Centres, Department of Gastroenterology and Hepatology, Amsterdam, the NetherlandsVrije Universiteit, School of Medicine, Amsterdam, the Netherlands
Amsterdam University Medical Centres, Department of Gastroenterology and Hepatology, Amsterdam, the NetherlandsAmsterdam Gastroenterology Endocrinology Metabolism (AGEM) Research Institute, Amsterdam, the Netherlands
Amsterdam Gastroenterology Endocrinology Metabolism (AGEM) Research Institute, Amsterdam, the NetherlandsAmsterdam UMC location University of Amsterdam, Laboratory for Experimental Oncology and Radiobiology, Centre for Experimental and Molecular Medicine, Amsterdam, the NetherlandsCancer Centre Amsterdam, Amsterdam, the NetherlandsOncode Institute, Amsterdam, the Netherlands
Amsterdam Gastroenterology Endocrinology Metabolism (AGEM) Research Institute, Amsterdam, the NetherlandsAmsterdam UMC location Vrije Universiteit Amsterdam, Department of Paediatric Gastroenterology, Amsterdam, the Netherlands
Amsterdam University Medical Centres, Department of Gastroenterology and Hepatology, Amsterdam, the NetherlandsAmsterdam Gastroenterology Endocrinology Metabolism (AGEM) Research Institute, Amsterdam, the NetherlandsVrije Universiteit, School of Medicine, Amsterdam, the Netherlands
The faecal immunochemical test (FIT) suffers from suboptimal performance and participation in colorectal cancer (CRC) screening. Urinary volatile organic compounds (VOCs) may be a useful alternative. We aimed to determine the diagnostic potential of urinary VOCs for CRC/adenomas. By relating VOCs to known pathways, we aimed to gain insight into the pathophysiology of colorectal neoplasia.
Methods
A systematic search was performed in PubMed, EMBASE and Web of Science. Original studies on urinary VOCs for CRC/adenoma detection with a control group were included. QUADAS-2 tool was used for quality assessment. Meta-analysis was performed by adopting a bivariate model for sensitivity/specificity. Fagan’s nomogram estimated the performance of combined FIT-VOC. Neoplasm-associated VOCs were linked to pathways using the KEGG database.
Results
Sixteen studies—involving 837 CRC patients and 1618 controls—were included; 11 performed chemical identification and 7 chemical fingerprinting. In all studies, urinary VOCs discriminated CRC from controls. Pooled sensitivity and specificity for CRC based on chemical fingerprinting were 84% (95% CI 73–91%) and 70% (95% CI 63–77%), respectively. The most distinctive individual VOC was butanal (AUC 0.98). The estimated probability of having CRC following negative FIT was 0.38%, whereas 0.09% following negative FIT-VOC. Combined FIT-VOC would detect 33% more CRCs.
Total 100 CRC-associated urinary VOCs were identified; particularly hydrocarbons, carboxylic acids, aldehydes/ketones and amino acids, and predominantly involved in TCA-cycle or alanine/aspartate/glutamine/glutamate/phenylalanine/tyrosine/tryptophan metabolism, which is supported by previous research on (colorectal)cancer biology.
The potential of urinary VOCs to detect precancerous adenomas or gain insight into their pathophysiology appeared understudied.
Conclusion
Urinary VOCs hold the potential for non-invasive CRC screening. Multicentre validation studies are needed, especially focusing on adenoma detection. Urinary VOCs elucidate underlying pathophysiologic processes.
]. As a result, many countries have implemented population-wide CRC screening programmes.
Currently, most countries use the faecal immunochemical test (FIT) for CRC screening. However, FIT suffers from the suboptimal performance: it does not detect 9–29% of CRC and 60–75% of advanced adenomas, and its suboptimal specificity results in unnecessary, burdensome and costly colonoscopies [
Performance characteristics of fecal immunochemical tests for colorectal cancer and advanced adenomatous polyps: a systematic review and meta-analysis.
]. Hence, an alternative, non-faecal test with improved accuracy would be of great value.
Within the field of non-invasive colorectal neoplasia biomarker exploration, analysing volatile organic compounds (VOCs) is a novel and promising approach [
]. VOCs are gas-phase metabolites that are produced during (patho)physiologic processes amongst other sources that are present in all bodily excrements [
]. VOC analysis can be performed using different instruments as well as data processing techniques, either revealing individual compounds or a 'chemical fingerprint'. To detect colorectal neoplasia, VOCs have predominantly been evaluated in faeces, exhaled breath and urine [
Systematic review with meta-analysis: volatile organic compound analysis to improve faecal immunochemical testing in the detection of colorectal cancer.
Urinary VOC-analysis has clear advantages over faecal and exhaled VOC-analysis, favouring their use in clinical practice. First, urinary VOC analysis seems to display a lower degree of background noise from external factors like diet, medication and analysis setting [
]. Other advantages of analysing urine over faeces are that urine can generally be collected on demand and is associated with greater patient acceptability [
], which may improve compliance. In contrast to breath, urine can easily be stored for the long term at a low cost. Considering this attractive profile, it is important to determine the potential of urinary VOCs.
The aim of the current study was to systematically review and meta-analyse previous studies on urinary VOCs for CRC and adenoma detection. As VOCs represent tumour/human metabolism, this study also sought to explore metabolic pathways associated with colorectal tumourigenesis. This review will serve as a stepping stone for further research on the potential of urinary VOCs, which would be valuable as an optimal non-invasive test for colorectal neoplasia is not yet available nor has its underlying pathophysiology been unravelled.
2. Methods
2.1 Data sources and search strategy
This systematic review was registered in PROSPERO (CRD42021233274) and conducted in accordance with PRISMA. The search was performed in July 2022 using PubMed, Embase and Web of Science. We combined keywords and free text terms representing 'colorectal neoplasm', 'volatolome' and 'urine' (Appendix Tables 1–3), without restrictions for language, publication type or publication date. Both chemical identification and chemical fingerprinting of VOCs were assessed. To identify additional eligible studies, reference lists of included records and relevant reviews found with the search were screened.
], two independent researchers screened titles and abstracts and thereafter performed full-text review. Any disagreement was discussed with the senior author. Studies were eligible if reporting on urinary VOC analysis for CRC or adenoma detection in humans. Studies were excluded if they had no control group or described outcomes from a mixed cancer cohort without CRC-specific outcomes. Non-original articles, case reports and abstracts were also excluded.
2.3 Data extraction and quality assessment
Data were extracted by one reviewer and double-checked by another. The following data were collected: study design, population and neoplasm characteristics, sampling and analysis methodology, neoplasm-associated urinary VOCs and diagnostic performance of urinary VOCs (sensitivity, specificity, area under the curve [AUC]). The quality of studies was evaluated by two independent researchers using the QUADAS-2 tool [
]. True positives/negatives and false positives/negatives were extracted from sensitivity and specificity. Heterogeneity was measured using the I2-statistic.
2.5 Combined FIT-VOC performance
Using our pooled outcomes as well as those from a recent meta-analysis on FIT [
Performance characteristics of fecal immunochemical tests for colorectal cancer and advanced adenomatous polyps: a systematic review and meta-analysis.
], Fagan’s nomogram in R-software (version 4.2.1) estimated the benefit of combined urinary VOC-FIT, including the number of CRCs detected and missed and the number needed to colonoscope [
]. Post-test probability of CRC in FIT-negatives was used as pre-test probability for the VOC-nomogram.
2.6 CRC pathophysiology
Urinary VOCs found to be associated with CRC were standardised against the Human Metabolome Database and PubChem. Subsequently, they were linked to metabolic pathways using MetaboAnalyst 5.0, which integrates enrichment analysis (hypergeometric test) and pathway topology analysis (relative-betweenness centrality test) based on the KEGG-pathway database.
3. Results
Total 16 studies—involving 2721 participants, 837 CRCs, 266 adenomas and 1618 controls—met the inclusion criteria (Appendix Fig. 1) [
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
], three targeted chemical identification using GC–MS and seven chemical fingerprinting using an electronic nose, field asymmetric ion mobility spectrometry (FAIMS) or gas chromatography–ion mobility spectrometry (GC–IMS) [
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
]. Study characteristics, identified neoplasm-associated VOCs and performance outcomes are summarised in Table 1. In most studies, urine samples were first frozen and then analysed; only two studies analysed fresh samples, prior to targeted chemical identification [
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
≥ 5 Adenomas or ≥ 3 adenomas with at least one of which is ≥ 10 mm.
(n = 23/24): 0.53 (0.36–0.70)
All adenomas (n = 55/24): 0.61 (0.47–0.75)
AUC = area under the curve; CI = confidence interval; CRC = colorectal cancer; FAIMS = field asymmetric ion mobility spectrometry; GC–IMS = gas chromatography ion mobility spectrometry; GC–MS = gas chromatography mass spectrometry; GC–qMS = gas chromatography–quadrupole mass spectrometry; GC–TOFMS = gas chromatography time-of-flight mass spectrometry; LC–FAIMS-MS = liquid chromatography–field asymmetric ion mobility spectrometry; n.a. = not applicable; SD = standard deviation; VOC = volatile organic compound; 8-oxoGua = 8-oxo-7,8-dihydroguanine; 8-oxodG(uo) = 8-oxo-7,8-dihydro-2’-deoxyguanosine; 5-hmUra = 5-hydroxymethyluracil.
* Whether levels were increased or decreased in CRC was not described.
† All participants.
‡ Median age.
§ nmol/mmol creatinine.
** Named PEN3; a device consisting of 10 different thick film metal oxide sensors.
†† Named WOLF; a device consisting of an array of electrochemical and optical sensors.
‡‡ Data was provided upon request.
§§ Adenoma with high-grade dysplasia, villous histology, ≥ 10 mm or ≥ 3 adenomas.
*** ≥ 5 Adenomas or ≥ 3 adenomas with at least one of which is ≥ 10 mm.
Seven of the 11 studies using GC–MS were deemed at moderate risk of bias (Appendix Table 4 and Fig. 2); they used a case-control design and controls did not undergo colonoscopy/cross-sectional imaging, so the absence of CRC could not be confirmed. The studies performing chemical fingerprinting were deemed at lower risk of bias, yet some failed to report whether colonic assessment was of high quality. Fifteen studies did not match the review question as these included patients once the diagnosis had been established [
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
]. Fifteen unique VOCs contributed to this separation, with a 95% sensitivity and 73% specificity upon cross-validation (AUC was not reported). In the second largest study, GC–MS distinguished 103 controls from 101 age- and gender-matched patients with CRC [
]. In total, 18 VOCs were discriminative; the most promising were putrescine, citrate, 2-aminobutyrate, p-cresol and hippurate (AUC per metabolite: 0.70–0.92). In another study, GC–MS discriminated 63 controls from 60 age-matched patients with CRC [
]. Sixteen VOCs were significantly different between groups, with succinic acid being the most distinctive.
Two studies compared patients with CRC with controls having CRC-suspected colorectal symptoms. In the first study, performance for early-stage (n = 24) versus late-stage (n = 34) CRC was modest (AUC 0.56) while high for CRC (n = 58) versus controls (n = 38, AUC 0.93) [
]. Twenty-three CRC-specific VOCs were identified. In the second study, positive outcomes were observed upon comparing CRC (n = 18) with controls and adenomatous/hyperplastic polyps (n = 65, AUC 0.91), CRC/polyps with controls (AUC 0.90) and CRC with polyps (AUC 0.90) [
]. Thirteen VOCs differed between CRC and controls/polyps.
Four studies performed GC–MS after isolating VOCs using different micro-extraction techniques. Using three-phase microextraction, VOCs with a concentration of> 20% were assessed in patients with CRC (n = 8) compared to controls (n = 35) and other cancer types (n = 38, Table 1) [
]. The authors created group-specific reference chromatograms, which they subsequently compared with the individual chromatograms from the same cohort. Not surprisingly, this resulted in a sensitivity of 100% to detect CRC. In another study, using solid phase microextraction, twelve CRC-specific VOCs were detected [
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
]. Anisole was most promising, with an 835% mean increase in urine concentration in CRC versus controls. Using needle trap microextraction (divinylbenzene/carboxen 1000/carbopack X sorbent), 99% of patients (n = 30) and controls (n = 30) were correctly classified upon cross-validation [
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
]. Forty-one unique VOCs were discriminative, of which butanal, p-tert-butyl-2-bromophenol and p-tert-butylphenol had an AUC of> 0.9. Lastly, using in-fibre extraction of volatile amines before targeted GC–MS, six amines could accurately discriminate CRC (n = 18) from controls (n = 24), lung cancer (n = 18) and breast cancer (n = 24) upon cross-validation [
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
]. After purification with HPLC and matching by age/sex/diet/bodyweight/smoking, both studies observed higher concentrations of 8-oxoGua and 8-oxodGuo in CRC versus controls. Additionally, Rozalski et al. found lower 5-hmUra levels in CRC [
] whereas 8-oxoGua and 5-hmUra levels were equal in both studies. CRC detection using either 8-oxoGua, 8-oxodGuo or 5-hmUra was suboptimal (AUC 0.64–0.69), but improved upon combining these VOCs (AUC 0.78 [95% CI: 0.70–0.86], sensitivity 79% and specificity 75%) [
]. Upon comparison with controls (n = 50) having comparable characteristics, CRC was correctly classified by FAIMS in 74% of cases, corresponding with a 88% sensitivity and 60% specificity (AUC not reported). In another study, an electronic nose device could discriminate CRC (n = 58) from controls having bowel symptoms suggestive of CRC (n = 38) with an AUC of 0.81 [
Upon comparing patients with CRC (n = 56) with their spouses and first-degree relatives (n = 82)—who share environmental and genetic factors, respectively—using FAIMS, the CRC volatolome appeared different with an AUC of 0.71 [
]. Analysing patients by either CRC-stage or -site did not demonstrate any significant difference. In another study, patients with CRC (n = 39) and irritable bowel disease (n = 35), two diseases with partly overlapping symptoms, were compared by an electronic nose device [
]. Disease-state class distinction was observed, but also some inter-class overlap and intra-class spreading, particularly in the irritable bowel disease group. CRC samples were successfully classified with a 78% sensitivity and 79% specificity (AUC not reported).
Three studies included patients with a higher pre-test probability of CRC compared to the screening population. The first study evaluated 562 patients with gastrointestinal symptoms suspicious of CRC [
]. Using FAIMS, CRC was detected with a modest sensitivity of 63%, specificity of 63% and AUC of 0.67, whereas outcomes for high-risk adenomas were 93%, 16% and 0.56, respectively. The second study also included patients with CRC-suspected colorectal symptoms [
]. Using FAIMS, CRC (n = 18) could be discriminated from controls and adenomatous/hyperplastic polyps (n = 356) with an AUC of 0.86. Performance for CRC and polyps versus controls was suboptimal (AUC 0.66, with particularly a low sensitivity); however, CRC could be distinguished from polyps with high accuracy (AUC 0.86, sensitivity 72%, specificity 89%). The last cohort consisted of 163 patients with a positive FOBT [
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
]. Upon comparing CRC with matched controls, a high degree of separation was observed: AUC 0.82 for GC–IMS and AUC 0.98 for FAIMS. However, a modest AUC of 0.61 was reported for all types of adenomas versus controls (GC–IMS). When patients with CRC and high-risk adenomas were grouped together and compared with controls, accuracy was modest as well (AUC 0.53 [GC–IMS] and 0.62 [FAIMS]). Interestingly, FAIMS could discriminate CRC (n = 12) from high-risk (n = 7), intermediate-risk (n = 12) and low-risk adenomas (n = 12) with high accuracy (AUC 0.92–0.83). Comparing patients having other gastrointestinal pathology (e.g. inflammatory bowel disease, diverticulosis, haemorrhoids) with those having CRC or adenomas resulted in high sensitivity but low specificity using either FAIMS (91% and 25%, respectively) or GC–IMS (71% and 55%, respectively), corresponding to a AUC of 0.56 (FAIMS) and 0.61 (GC–IMS).
3.4 Meta-analysis
Upon meta-analysing performance of chemical fingerprinting techniques for urinary VOC analysis, we found a pooled sensitivity and specificity to detect CRC of 84% (95% CI 73–91%) and 70% (95% CI 63–77%), respectively (Fig. 1). Due to lack of publicly available data, studies performing chemical identification techniques could not be included and pooled accuracy to detect adenomas could not be generated.
Fig. 1Forest plots showing the sensitivity and specificity (with corresponding 95%-CI) of urinary volatile organic compounds for the detection of colorectal cancer.
The estimated probability of having CRC following negative FIT at a threshold > 20 μg (used in most screening programmes) was 0.38%, whereas 0.13% when using a low threshold of 10 μg (Appendix Fig. 3). Upon further testing FIT-negatives for urinary VOCs, probability of CRC following negative FIT-VOC decreased to 0.09% and 0.03% (threshold >20 μg and 10 μg, respectively). Both FIT-VOC > 20 μg and 10 μg resulted in all 12 CRCs being detected per 1000 patients, instead of missing n = 3 and n = 1 CRCs with single FIT > 20 μg and 10 μg, respectively (Table 2). By applying the FIT-VOC test strategy, the number needed to colonoscope increased three to fourfold.
Table 2Number of colorectal cancers (CRC) detected and missed when using solely FIT (white rows) and when FIT-negatives are further tested with urinary VOCs (light blue rows), including the number of patients needed to colonoscope to detect one CRC per 1000 patients tested.
Within the eight studies using untargeted GC–MS, 100 unique urinary VOCs were associated with CRC (Appendix Table 5). These 100 metabolites were particularly hydrocarbons, followed by carboxylic acids, aldehydes/ketones and amino acids (Fig. 2). Carboxylic acids were either increased or decreased in CRC. In contrast, amino acids were predominantly elevated (Fig. 2). Whether hydrocarbons and aldehydes/ketones were either increased or decreased in CRC could not be investigated due to lack of precise metabolite levels. When we linked the 100 metabolites to metabolic pathways, we found that predominantly markers involved in the tricarboxylic acid (TCA) cycle, alanine/aspartate/glutamate metabolism, glutamine/glutamate metabolism, and phenylalanine/tyrosine/tryptophan biosynthesis and metabolism were associated with CRC (Fig. 3).
Fig. 2Chemical classes associated with colorectal carcinogenesis based on untargeted gas chromatography–mass spectrometry analysis.
Fig. 3Metabolic pathways found to be significantly associated with colorectal carcinogenesis, arranged by p-values (from pathway enrichment analysis) on the Y-axis and pathway impact values (from pathway topology analysis) on the X-axis. Larger node radii represent higher pathway impact values, whereas darker colours increased p-values.
In this systematic review, we determined the diagnostic potential of urinary VOCs for colorectal neoplasia in a screening fashion. In all 16 studies, involving 2721 participants, urinary VOCs discriminated patients with CRC from controls. Pooled sensitivity and specificity for CRC using chemical fingerprinting techniques were 84% (95% CI 73–91%) and 70% (95% CI 63–77%), respectively. The most distinctive solitary VOC was butanal, an oxidative stress marker [
Evaluation of a multi-parameter biomarker set for oxidative damage in man: increased urinary excretion of lipid, protein and DNA oxidation products after one hour of exercise.
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
]. Although this finding was not confirmed in all other studies, it is supported by previous research in which oxidative damage has been shown to play an important role in (colorectal) carcinogenesis and to persist in CRC [
Reactive oxygen species in colorectal cancer: the therapeutic impact and its potential roles in tumor progression via perturbation of cellular and physiological dysregulated pathways.
], as well as by the fact that in our review three other oxidative stress markers (8-oxoGua, 8-oxodGuo, 5-hmUra) also demonstrated promising results (AUC 0.78).
None of the identified VOCs were associated with CRC repeatedly across all reviewed studies using chemical identification. This might be explained by the different methods used for sample analysis; a change in GC–MS set-up will lead to different results as a different part of the chemical window is analysed. On the other hand, this might imply that chemical identification is unfeasible for CRC detection. This is supported by previous research which showed overlaps in individual VOCs between various malignancies, while the 'volatile chemical fingerprint' seems to be more unique to specific types of cancer [
]; possibly because a wider spectrum of the chemical information might be analysed. Other advantages of chemical fingerprinting techniques are that they are rapid, easy, portable and low-cost in practice. Hence, we believe chemical fingerprinting would be more appropriate for population-wide screening than detailed chemical identification.
Next to the identification of CRC, screening aims to accurately identify precancerous adenomas [
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
]. It should be noted that both studies were characterised by a low prevalence of adenomas. Conflicting test accuracy was also seen using targeted GC–MS; urinary level of 8-oxodGuo differed significantly in adenomas versus controls in one study [
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
] using FAIMS or GC–IMS resulted in high sensitivity, which suggests that adenomas also generate a distinct urinary VOC profile. Moreover, accurate detection of adenomas has been observed with faecal and exhaled VOCs [
]. Hence, further research exploring whether urinary VOCs could serve as a non-invasive biomarker for adenomas is warranted.
Diagnostic performance of urinary VOCs for CRC and adenomas may improve upon combining these markers with others. We estimated in a screening population that the 0.38% probability of CRC following negative FIT (cut-off >20 μg) decreased to 0.09% following combined FIT-VOC. Moreover, 33% more CRCs would be detected with the FIT-VOC test strategy. In view of the cost- and resource-burden, the clinical utility of FIT-VOC for screening remains questionable, though would require formal evaluation. This evaluation should include the positive effect of FIT-VOC on adenoma detection, which might reduce the number needed for colonoscope to detect one (pre)cancerous lesion.
The second aim of the current review was to gain insight into the underlying pathophysiology of colorectal neoplasia. We found an association between the presence of CRC and altered urinary levels of carboxylic acids, amino acids, hydrocarbons and aldehydes/ketones. Altered levels of carboxylic acids and amino acids have also been observed in faeces, serum and CRC tissue [
In-depth mapping carboxylic acid metabolome reveals the potential biomarkers in colorectal cancer through characteristic fragment ions and metabolic flux.
Metabolic profiling of human colorectal cancer using high-resolution magic angle spinning nuclear magnetic resonance (HR-MAS NMR) spectroscopy and gas chromatography mass spectrometry (GC/MS).
]. The increase in amino acids may be explained by three phenomena frequently observed in CRC: enrichment of the mucin-degrading bacteria Akkermansia muciniphila and Bacteroides (mucins are composed of amino acids) [
]. Carboxylic acids on the other hand contribute to the production of NADPH, which is essential to maintain cellular redox balance, meaning that sufficient reactive oxygen species are generated to support cell proliferation while abundant levels are avoided as these would lead to cell death [
Reactive oxygen species in colorectal cancer: the therapeutic impact and its potential roles in tumor progression via perturbation of cellular and physiological dysregulated pathways.
Next to these four chemical classes, we found that CRC was related to the following metabolic pathways: TCA cycle, alanine/aspartate/glutamate metabolism, glutamine/glutamate metabolism, and phenylalanine/tyrosine/tryptophan biosynthesis and metabolism. This is in line with previous research evaluating the serum metabolome [
In-depth mapping carboxylic acid metabolome reveals the potential biomarkers in colorectal cancer through characteristic fragment ions and metabolic flux.
] and dysfunction of this cycle—as a result of dysregulated enzymes like succinate—has been shown to play an important role in the development of (colorectal) cancer [
Targeted metabolomics in formalin-fixed paraffin-embedded tissue specimens: liquid chromatography-tandem mass spectrometry determination of acidic metabolites in cancer research.
]. The association between CRC and glutamine/phenylalanine/tyrosine/tryptophan pathways may be explained by the fact that glutamate, phenylalanine and tyrosine can be converted to citrate [
]. In addition, glutamine is crucial for the production of other amino acids (like alanine and aspartate) and of glutathione, which contributes to the redox balance [
]. As patients with adenomas were not included in the untargeted GC–MS studies, we were not able to gain insight into the underlying pathophysiology of these premalignant lesions.
Our review is limited by the modest quality of included studies: most studies did not resemble a true screenings setting, various sampling methodologies and statistical models were used (or these were not reported in sufficient detail) and several were deemed at moderate or unclear risk of bias. The studies included in the meta-analysis (which all used chemical fingerprinting) were at lower risk of bias, nonetheless, the meta-analysis showed heterogeneity among included studies. Regarding the studies using chemical identification, on the other hand, the variation in methodology limited the matching CRC-associated VOCs between studies.
To move urinary VOCs forward towards clinical practice, standardised practices for sample preparation, sample analysis and data analysis are crucial, to ensure accuracy, head-to-head comparison and reproducibility of studies. Whilst the optimal analytical strategy for urinary VOC analysis remains to be established, protocols and important considerations have been published [
]. Regarding sample collection, samples should be collected in sealed containers, immediately frozen after collection and analysed within 9–12 months [
]. Next to studies on the optimal analytical strategy, further studies should evaluate cost-effectiveness as well as external confounding variables that limit the detection of colorectal neoplasia (so should be corrected for) [
]. Lastly, multicentre validation studies on diagnostic performance for both CRC and its precursor lesions (being adenomas) are warranted. To establish the patient number needed to produce solid evidence, our pooled data can be used for a formal power calculation. These studies should resemble a real-life screening setting, meaning that consecutive pre-diagnosed patients with a normal pre-test probability of CRC should be included. Ideally, all participants undergo colonoscopy as reference standard.
In conclusion, urinary VOC analysis—particularly by means of chemical fingerprinting—seems promising for non-invasive CRC screening. Multi-centre validation studies are needed, as well as studies evaluating performance for adenomas, the optimal analytical strategy, confounding factors and cost-effectiveness. In addition to their diagnostic potential, we showed that VOC identification on molecular level (as measured by for example GC–MS) offers insight into underlying pathophysiologic processes.
Conflict of interest statement
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: EvL, LvD, MH, GB and SB have nothing to declare. LV received consultancy fees from Bayer, MSD, Genentech, Servier, and Pierre Fabre, but these had no relation to the content of this publication. LV is a New York Stem Cell Foundation–Robertson Investigator. NdB has served as a speaker for AbbVie and MSD and has served as a consultant and principal investigator for TEVA Pharma BV and Takeda. He has received a research grant (unrestricted) from Dr. Falk, TEVA Pharma BV, Dutch Digestive Foundation (MLDS) and Takeda; all outside the submitted work. TdM has served as a speaker for Nutricia, Mead Johnson and Winclove. He has served as an advisory board member for Nutricia. DR has received a research grant (unrestricted) from AbbVie, outside the submitted work. He has served as a member of the Data Safety Monitoring Board of Vivoryon Therapeutics.
Funding statement
This study was funded by Dutch Digestive Foundation (MLDS). The funding organisation did not have any involvement in the study’s design, conduct or reporting.
Contributions
Guarantor of the article: NKH de Boer. EvL, DR and NdB conceived the study design. EvL, LD and GB developed and performed the systematic literature search. EvL and LD collected data and drafted the manuscript. MW performed the meta-analysis and the analysis for combined FIT-VOC. All authors critically revised the manuscript for important intellectual content. All authors had full access to all the data and approved the final version of the manuscript, including the authorship list.
Data sharing statement
The main data supporting the findings of this study are included within the article and its supplementary materials. Additional data are available from the corresponding author upon reasonable request.
Performance characteristics of fecal immunochemical tests for colorectal cancer and advanced adenomatous polyps: a systematic review and meta-analysis.
Systematic review with meta-analysis: volatile organic compound analysis to improve faecal immunochemical testing in the detection of colorectal cancer.
Colorectal cancer and adenoma screening using urinary volatile organic compound (VOC) detection: early results from a single-centre bowel screening population (UK BCSP).
Exploring the potential of needle trap microextraction combined with chromatographic and statistical data to discriminate different types of cancer based on urinary volatomic biosignature.
Investigation of urinary volatile organic metabolites as potential cancer biomarkers by solid-phase microextraction in combination with gas chromatography-mass spectrometry.
Nanoconfinement effect based in-fiber extraction and derivatization method for ultrafast analysis of twenty amines in human urine by GC-MS: application to cancer diagnosis biomarkers' screening.
Evaluation of a multi-parameter biomarker set for oxidative damage in man: increased urinary excretion of lipid, protein and DNA oxidation products after one hour of exercise.
Reactive oxygen species in colorectal cancer: the therapeutic impact and its potential roles in tumor progression via perturbation of cellular and physiological dysregulated pathways.
In-depth mapping carboxylic acid metabolome reveals the potential biomarkers in colorectal cancer through characteristic fragment ions and metabolic flux.
Metabolic profiling of human colorectal cancer using high-resolution magic angle spinning nuclear magnetic resonance (HR-MAS NMR) spectroscopy and gas chromatography mass spectrometry (GC/MS).
Targeted metabolomics in formalin-fixed paraffin-embedded tissue specimens: liquid chromatography-tandem mass spectrometry determination of acidic metabolites in cancer research.
00124-7&id=gr1.jpg){kind=link}
00124-7&id=gr2.jpg){kind=link}
00124-7&id=gr3.jpg){kind=link}