Special article
13271
Volatile organic compounds as new biomarkers for colorectal
cancer: a review
Abstract
Analysis of the volatile part of the metabolome (volatile
organic compounds, VOC) present in the gas phase of
excreted materials is a promising new screening tool for
several cancers, including colorectal cancer (CRC). The
VOC signature can reflect health status, like a ‘finger-
print’, and can be modified in several diseases. Technical
difficulties still limit the widespread use of VOC analysis
in the clinical setting, but this approach has already
been applied successfully in the diagnosis of CRC. The
present study reviews the available data on VOC present
in the headspace (the gaseous constituents of a closed
space above a liquid or solid) of blood, urine, faeces
and breath as a potential screening tool for CRC. A sys-
tematic electronic literature search was conducted in
PubMed, Scirus and Google using the following key-
Introduction
Colorectal cancer (CRC) is one of the commonest
tumours and is an important cause of cancer-related
mortality [1]. The adenoma–carcinoma sequence has
allowed for a mass screening with the aim of decreasing
mortality [2]. Colonoscopy is the gold standard for the
diagnosis of CRC, although its cost prevents its use for
mass screening. Faecal immunochemical blood testing
(FIT) is the most widely used noninvasive screening
tool, showing fairly good specificity but a high variation
in sensitivity (61–91%) and adherence to screening pro-
grammes rarely reaches 70% of the target population
[3]. Volatile organic compounds (VOC) are present in
various excreted biological materials and their analysis
offers a possibility for cancer screening.
Analysis of the volatile part of the metabolome,
named the ‘volatome’, was initiated in 1971 by the
Nobel laureate Pauling et al. [4], who described the
presence of about 250 VOC in human breath and urine
using gas chromatography–mass spectrometry (GC-
MS). In 1989 Williams and Pembroke [5] demon-
strated that sniffer dogs could be trained to detect the
presence of melanoma, suggesting the idea of a charac-
teristic ‘smellprint’ for any disease. This ‘volatile
hypothesis’ was confirmed by other studies that used
the canine sense of smell for the diagnosis of bladder
cancer [6], lung and breast cancer [7] and CRC [8].
VOC are the final products of cellular metabolism prob-
654 Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
doi:10.1111/codi.
words: Metabolomic, Volatile Organic Compounds
(VOC), Electronic-nose and Colorectal Cancer. Only
articles published in English between 2000 and 2015
were selected and these were independently checked by
two of the authors. Ten papers describing the reliability
of VOC analysis in breath and faeces, blood and urine
were selected; all indicated good reliability in detecting
CRC. The use of different substrates and different ana-
lytical platforms has led to the identification of different
patterns of VOC. The reliability of a metabolomic
approach as a noninvasive biomarker for use in CRC
screening is supported by this review despite several lim-
itations due to the number of patients included in each
study, the different analytical platforms and biological
materials used and different VOC identified.
ably produced by the oxidative stress or peroxidation of
cell membranes as a consequence of gene or protein
alterations in cancer cells. VOC can reflect any meta-
bolic changes in response to inflammation, necrosis,
cancer degeneration or alteration of microbiota or can
be related to external factors such as environmental pol-
lution, medication and diet [8]. These metabolites are
released into the bloodstream, and wherever they are
produced in the body will reach the alveoli or renal
tubules, where they are excreted. VOC released into the
gut may be detected in the faeces.
In the last 20 years this approach has been applied in
different clinical settings, including cancer. Several stud-
ies have demonstrated the high reliability of a metabo-
lomic approach for detecting and discriminating
patients with lung [9], prostate [10], breast [11], skin
[12], gastric and oesophageal [13,14] and thyroid [15]
cancer from controls. The present review aims to evalu-
ate the potential role of VOC analysis as a mass screen-
ing tool for CRC.
Method
A systematic electronic literature search was conducted
by accessing the United States National Library of Med-
icine database (Medline–PubMed), Scirus (for scientific
information) and the Google search engine. The search
was supplemented by further scanning of the reference
lists obtained from retrieved articles in order to identify
M. Di Lena et al. VOC as biomarkers for colorectal cancer

any relevant additional study material missed on the ini-
tial surveys. The selected literature was screened by two
authors (MDL and FP) independently and then cross-
matched. Any duplicate nonoriginal articles such as edi-
torials and reviews were excluded.
The search terms ‘Metabolomic’, ‘Volatile Organic
Compounds (VOC)’, ‘Electronic-nose’ and ‘Colorectal
Cancer’ were cross-checked and the risk of bias in indi-
vidual studies was eliminated by selecting only papers
dealing with VOC excluding other metabolites and only
screening for primary CRC. All the eligible studies on
VOC in CRC were included irrespective of the out-
come. Only original articles published in the English
language and appearing in peer-reviewed journals
between January 2000 and June 2015 were considered.
Results
Ninety-six titles or abstracts were identified from the
electronic search and reference lists in the preliminary
review. According to the PRISMA flow diagram
(Fig. 1) 84 did not meet the criteria for inclusion. After
removal of duplicates and screening for relevant titles,
10 articles were submitted for a full-text analysis and
included in this systematic review. Three were related to
VOC identified in urine [16–18], four in exhaled breath
[8,19–21], one in blood [22] and two in faeces [23,24]
using different analytical platforms.
Analytical platforms
Several types of technologies using different analytical
platforms have been applied to the analysis of VOC for
CRC.
Gas chromatography–mass spectrometry
This is the gold standard hybrid analytical platform
employed successfully for the chemical characterization
of VOC biomarkers. This method was used in seven
[8,16,18–22] of the 10 studies. GC-MS allows the

Records identified through Additional records identified

database searching through other sources
(n = 96) (n = 4)

Records after duplicates removed

(n = 23)

Records screened Records excluded

(n = 77) (n = 65)

Full-text articles assessed Full-text articles excluded,

for eligibility with reasons
(n = 12) (n = 2)

Studies included in
qualitative synthesis
(n = 10)

Figure 1 PRISMA flow diagram.

Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663 655
VOC as biomarkers for colorectal cancer
physical separation of the volatile molecules in samples
and the subsequent identification of their nature and
quantification, permitting identification of the exact
composition of the sample analysed. Giving qualitative
and quantitative information about sample composition,
this analytical technique allows the identification of the
metabolic profile characteristic of CRC patients com-
pared with healthy controls.
Despite the advantage of its high efficiency, GC-MS
has a high cost and low manageability, which limit its
use in routine application for mass screening. The tech-
nique requires three different operative steps including
breath collection, sample preconcentration and analysis.
Different sample preconcentration methods have been
employed. In four studies [8,16,21,22] solid phase
microextraction (SPME) was used. In this technique
equilibrium is established between the sample and the
headspace above and analytes from the sample are
absorbed by a polymer-coated fused fibre. They are
then desorbed from the fibre to a chromatography col-
umn. In two studies a multi-bed sorbent cartridge tech-
nique was adopted which focused on breath monitoring
[19] and in one study [18] a Combi-PAL ITEX auto-
mated preconcentrator system was coupled to GC-MS
for urine analysis.
In a paper on urinary VOC [18] another innovative
technique, field asymmetric ion mobility spectrometry
(FAIMS), was used. This technique achieves separation
of chemical compounds according to their different ion-
ized chemical mobilities in an electric field, working at
atmospheric pressure and room temperature, unlike
GC-MS. After ionization, the neutral molecules of the
sample become ions of various sizes and types. By
applying an asymmetric high-voltage waveform, differ-
ent movements of these molecules within the high elec-
tric field can be measured, and a consequent separation
of the complex mixture will be observed. This technol-
ogy has the advantage of supplying a physical separation
of the compounds and has better manageability than
GC-MS, but the absence of a reference database limits
the ability to obtain reliable qualitative and quantitative
information about the VOC contained in the samples.
These compounds, therefore, can be recognized and
quantified only if their nature is already known.
Selected ion flow tube mass spectrometry (SIFT-MS)
This technique was employed in a study which focused
on the monitoring of VOC from faecal samples [23].
SIFT-MS is a quantitative MS technique that allows
the real-time measurement of concentrations of trace
gaseous molecules in humid air samples. It is based on
chemical ionization induced by precursor ions H3O+,
NO+ and O2+ generated in a microwave discharge and
656 Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
M. Di Lena et al.
selected by a quadrupole mass filter. The selected ion
is injected into a fast flowing helium carrier gas down
a flow tube. The sample is then introduced into the
flow tube and the precursor ion reacts with the trace
gases and VOC in the sample. The precursor and pro-
duct ions in the carrier gas are separated in a second
quadrupole mass spectrometer and counted in a detec-
tor. Data may be obtained through scanning a spec-
trum at a user-defined range of mass-to-charge ratio
(m/z) values and absolute concentrations of trace
compounds can be calculated in real time from the
ratio of the precursor and product ion signal ratios
without the need for sample preparation or calibration
with standard mixtures.
Electronic-nose (e-nose) technology
This was used in four of the studies [8,17,20,24] and is
based on an array of sensors able to generate an electri-
cal response in the presence of a chemical class of VOC,
giving a qualitative nonspecific response. An e-nose
does not enable a single VOC to be recognized, but as
in human olfaction, it detects the presence of a combi-
nation of different chemical classes, such as alkanes,
alcohols and aromatic compounds. E-nose technology is
faster and easier to perform than GC-MS and can be
trained using pattern recognition to identify a particular
state. At present there are different kinds of commercial
and custom e-noses, containing an array of 6–32 sensors.
Study characteristics
All the studies considered were prospective, compara-
tive, observational case–control studies with relatively
small samples.
Volatile organic compounds in the urine headspace
The high concentration of volatile metabolites in the
urine makes it an attractive target for VOC analysis.
Silva et al. [16] evaluated differences in the urinary
VOC profile of cancer patients and controls using
dynamic headspace microextraction (DHS-SPME) in
combination with GC-MS. The cancer group included
14 cases of leukaemia, 12 CRC patients and seven lym-
phoma patients; this group was compared with 21
healthy controls. Fifteen VOC were identified that dif-
fered between CRC and controls, although the authors
concluded that CRC patients were better characterized
by the presence of 1,4,5-trimethylnaphthalene, 2,7-
dimethylquinoline and 2-methyl-3-phenyl-2-propenal
(Table 1). This study was the first application of VOC
analysis of urinary samples from CRC patients, although
it is nonspecific for CRC and evaluated three other
types of cancers that are very different from each other.
 Table 1 Analytical platforms, biological matrices and chemical compounds identified in the 10 publications selected in the present review.

Analytical platform Patients Biomarkers Chemical classes Sensitivity Specificity P

M. Di Lena et al.

Breath
Peng e-nose 26 CRC 1,10 -(1-Butenylidene)- Aromatic compound < 30% – –
(2010) [8] bis-benzene
GC-MS 22 HC 1,3-dimethylbenzene Aromatic compound
2-Amino-5-isopropyl- Amino derivative
8-methyl-1-azulenecarbonitrile
1-Iodononane Haloalkane
[(1,1-Dimethylethyl)thio] Acid
acetic acid
4-(4-Propylcyclohexyl)- Aromatic acid
40 -cyano[1,10 -biphenyl]-
4-yl ester benzoic acid
Altomare GC-MS 37 CRC Nonanal Aldehyde 86% 83% –
(2013) [19] 41 HC Decanal Aldehyde
4-Methyl-2-pentanone Ketone
2-Methylbutane Alkane
4-Methyloctane Alkane
4-Methylundecane Alkane
2-Methylpentane Alkane
3-Methylpentane Cycloalkane
Methylcyclopentane Cycloalkane
Cyclohexane Alkene
Methylcyclohexane Aromatic compound
Trimethyldecane-1,2-

Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
pentadiene
1,3-Dimethylbenzene
1,4-Dimethylbenzene
Wang GC-MS 20 CRC Cyclohexanone Ketone – – < 0.05
(2014) [21] 20 HC Dodecane Alkane
2,2-Dimethyldecane Alkane
4-Ethyl-1-octyn-3-ol Alcohol
Trans-2-dodecen-1-ol Alcohol
Cyclooctylmethanol Alcohol
Ethylaniline Aromatic amine
derivative
VOC as biomarkers for colorectal cancer

657
 Table 1 (Continued).

658
Analytical platform Patients Biomarkers Chemical classes Sensitivity Specificity P

3-Hydroxy-2,4,4- Ester
trimethylpentyl 2
methylpropanoate
6-t-Butyl-2,2,9,9- Alkyne
tetramethyl-
3,5-decadien-7-yne
Amal GC-MS 65 CRC Acetone Ketone For e-nose: For e-nose: 94% –
(2015) [20] e-nose 12 advanced aden. Ethylacetate Ester 85% (CRC vs HC) (CRC vs HC)
VOC as biomarkers for colorectal cancer

10 nonadvanced Ethanol Alcohol 94% (CRC 88% (CRC

aden. vs aden.) vs aden.)
122 HC 4-Methyloctane Alkane 94% (control 94% (control
vs aden.) vs aden)
Blood
Wang GC-MS 16 CRC Phenylmethylcarbamate Carbamic – – < 0.05
(2014) [22] acid ester
20 HC Ethylhexanol Alcohol
6-t-Butyl-2,2,9,9- Alkyne
tetramethyl-3,5-
decadien-7-yne
1,1,4,4-Tetramethyl-2,5- Cycloalkane
dimetylene-cyclohexane
Faeces
de Meij Cyranose 320 40 CRC – 85% for CRC 87% for CRC < 0.001 for CRC
(2014) [24] 60 advanced aden. 62% for adenomas 86% for < 0.001 for
57 HC adenomas adenomas
Batty SIFT-MS 62 FOBT-positive Hydrogen sulphide Sulphur compound 72% 78% –
(2015) [23] patients: 31 from Dimethylsulphide Sulphur compound
normal controls and Dimethyldisulphide Sulphur compound
31 from high-
risk group
Urine
Silva DHS-SPME/ 12 CRC Hexanal Aldehyde < 0.05
(2011) [16] GC-MS 21HC Heptanal Aldehyde
2-Methyl-3-phenyl- Aldehyde
2-propenal;
1,2,4-Trimethylbenzene Aromatic compound

Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
M. Di Lena et al.
 Table 1 (Continued).

Analytical platform Patients Biomarkers Chemical classes Sensitivity Specificity P

p-Cymene Aromatic compound

M. Di Lena et al.

1,2-Dihydro-1,1,6- Aromatic compound

trimethyl-naphthalene
1,4,5-Trimethylnaphthalene Aromatic compound
Anisole Aromatic ether
1-Octanol Alcohol
4-Methylphenol Aromatic alcohol
c-Terpinene Terpene
Bomylene Terpene
Dimethyl disulphide Sulphur compound
2-Methoxythiophene Thiophene derivative
2,7-Dimethylquinoline Quinoline derivative
Westenbrink e-nose 39 CRC – 78% 79% –
(2015) [17] 35 IBS
18 HC
Arasaradnam FAIMS/GC-MS 83 CRC Acetaldehyde (or ethylene Aldehyde 88% 60% < 0.001
(2014) [18] oxide or oxalic acid)
50 HC Acetone Ketone
2-Pentanone (or 3-methyl- Ketone
2-butanone or 2,3-
butanedione)
4-Heptanone (or Ketone
3-heptanone or 2,4-
dimethyl-3-pentanone)
Dimethyldiazene (or Nitrogen compound

Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
cyclobutylamine or
oxepane)
Acetyloxime-pyridine Pyridine derivative
carboxaldehyde (or
hydrocinnamoylbenzene-
ethanamine or styrene or
dimethylthiourea)
Allylisothiocyanate (or Isothiocyanato
isothiocyanato- derivative
cyclopropane or 2-
cyanoacetamide)
VOC as biomarkers for colorectal cancer

659
VOC as biomarkers for colorectal cancer M. Di Lena et al.

Two further studies [17,18] were conducted at the

CRC, colorectal cancer; HC, healthy control; aden, adenoma; GC-MS, gas chromatography-mass spectrometry; e-nose, electronic nose; SIFT-MS, selected ion flow tube mass spec-
trometry; FOBT, faecal occult blood test; FAIMS, field asymmetric ion mobility spectrometry; IBS, irritable bowel syndrome; DHS-SPME, dynamic headspace solid phase microex-
University Hospital Coventry and Warwickshire (UK).
These studies considered that since for patients admit-
ted to a CRC screening programme collecting urine can
be more acceptable than collecting faeces compliance
P

could be improved. The first of the studies evaluated

urine samples obtained from 83 CRC patients and 50
healthy controls using a FAIMS instrument and a
Specificity

Combi-PAL ITEX automated preconcentrator system

(CTC Analytics AG, Switzerland) coupled to a GC-MS.
This showed a sensitivity of 88% and a specificity of
60% in discriminating between the two groups [18].
The VOC identified by GC-MS are shown in Table 1.
In the second study [17] the same authors developed a
dedicated e-nose, the WOLF system, made up of 13
Sensitivity

sensors (10 amperometric electrochemical sensors, two

nondispersive infrared optical devices and one photoion-
ization detector), testing its discriminatory power on
the headspace of 92 urine samples including 39 CRC
patients, 35 patients with irritable bowel syndrome
Oxime derivative
Chemical classes

(IBS) and 18 healthy controls. This new technology

showed a sensitivity of 78% and specificity of 79% in
detecting CRC patients compared with the other two
Alcohol

groups, confirming the presence of a metabolomic

derangement identifiable in the urine samples of CRC
patients and the efficiency of urine as a potential good
biological material for screening.
Methoxy-phenyl-oxime (or
ethylbenzoic acid (pentyl

Hexen-1-ol (or 4-methyl-

ester) or carbamic acid

Volatile organic compounds in the faecal headspace

1-hexene or hexanal)

The first report of the analysis of faecal samples was by

(methyl ester))

de Meij et al. [24]. They used the Cyranose 320

(Smiths Detections, Pasadena, California, USA) to
Biomarkers

monitor 40 CRC patients, 60 patients with advanced

colonic adenomas and 57 healthy controls. A FIT was
performed on the faecal samples of 35 CRC, 46
advanced adenomas and 52 controls from the same
cohort. The Cyranose 320, a commercial e-nose based
on an array of 32 nanocomposite organic polymeric
sensors, showed a sensitivity and a specificity of 85%
and 87%, in the comparison of CRC patients and con-
Patients

trols. A sensitivity of 62% and a specificity of 86% was

found when comparing advanced adenoma patients with
controls and a sensitivity of 75% and a specificity of 73%
Analytical platform

was obtained for the comparison of patients with CRC

and advanced adenoma. FIT used in this study was an
automated OC-sensor test, showing a specificity of
100% in the three groups and a sensitivity of 63% for
Table 1 (Continued).

CRC and 7% for advanced adenoma. Batty et al. [23]

studied 31 faecal samples from normal controls and 31
from a group of high-risk patients with a high grade
adenoma or adenocarcinoma, stratified after colono-
traction.

scopy performed on the basis of positivity on a faecal

occult blood test. SIFT-MS analysis was performed

660 Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
M. Di Lena et al.
using each of the three precursor ions available (H3O+,
NO+, O2+). Four ions were identified that were statisti-
cally different in the two groups monitored (m/z 35,
m/z 90, m/z 62, m/z 94) most likely derived from
hydrogen sulphide, dimethylsulphide and dimethyldisul-
phide (the fourth being unknown). The technique
showed a specificity of 78% and a sensitivity of 72%.
Volatile organic compounds in the blood headspace
The metabolomic approach has been applied in blood/
serum samples from CRC patients, showing different
altered metabolic pathways compared with healthy sub-
jects [25,26], but only one study analysed the volatile
part of the metabolome in the complex matrix of
blood/serum [22]. Based on the hypothesis that VOC
are first released into the bloodstream and then reach
the alveoli and are exhaled, Wang et al. [22] investi-
gated the VOC present in blood samples from 16 CRC
patients (eight colon cancer and eight rectal cancer) and
20 healthy controls using SPME-GC-MS. They found
three VOC to be present in significantly lower amounts
(phenylmethylcarbamate, ethylhexanol and 6-t-butyl-
2,2,9,9-tetramethyl-3,5-decadien-7-yne) and one in a
significantly higher amount (1,1,4,4-tetramethyl-2,5-
dimetylene-cyclohexane) (P < 0.05) in CRC patients
compared with healthy subjects.
Volatile organic compounds in the breath
Peng et al. [8] published the first report of metabolo-
mic breath analysis in CRC patients using a tailor-made
14-nanosensor array based on organically functionalized
gold nanoparticles (GNPs) and a SPME-GC-MS analy-
sis to identify suitable representative VOC to distinguish
CRC from other types of cancer. Breath samples were
collected from 26 CRC patients, 30 lung cancer
patients, 22 breast cancer patients, 18 prostate cancer
patients and 22 healthy controls. The GNP array
showed high discriminant ability in identifying CRC
patients from healthy controls, while using the six VOC
identified (see Table 1) by GC-MS in CRC patients the
sensitivity dropped to about 30%.
In 2013 our group analysed breath samples from 37
CRC patients and 41 healthy controls with thermal des-
orption (TD)-GC-MS, and we were able to identify a
pattern of 15 VOC (see Table 1) showing a sensitivity
of 86%, a specificity of 83% and an accuracy of 85% in
discriminating between the two groups, with an area
under the receiver operating characteristic curve (AUC)
of 0.85. The discriminant power of the exhaled VOC
pattern was then validated in a blind phase on a further
series of 25 subjects showing an accuracy of 76% [19].
A further study on breath analysis was published by
Wang et al. [21], who analysed breath samples from 20
Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
VOC as biomarkers for colorectal cancer
CRC patients and 20 healthy controls using SPME-GC-
MS. Nine VOC were significantly increased in exhaled
breath from CRC patients (P < 0.05) and another one
was significantly reduced compared with healthy volun-
teers (P < 0.05) (see Table 1). Recently Amal et al. [20]
published the results of an analysis of exhaled breath from
65 patients with CRC, 22 with adenoma and 122 healthy
controls. GC-MS and an e-nose made from cross-reactive
nanoarrays in combination were used for the analysis.
GC-MS revealed four compounds that were significantly
different in the CRC and control groups. In particular,
acetone and ethyl-acetate were higher in the CRC group,
while ethanol and 4-methyl-octane were higher in the
control group. The performance of the sensors in dis-
criminating CRC and control groups resulted in a sensi-
tivity of 85%, a specificity of 94% and an accuracy of 91%.
The use of nanoarray technology resulted in good dis-
crimination between the CRC and adenoma groups, the
adenoma and control groups and between advanced and
nonadvanced adenoma, even if the small number of ade-
noma patients made the results unreliable The authors
evaluated the role of different confounding factors such
as age, gender, current smoking and fasting of at least
12 h, revealing any influence on sampling and the infor-
mation contained in exhaled air.
Discussion
Analysis of the 10 studies reviewed shows that a
derangement of metabolomics related to the presence
of CRC really exists and that it could be detected by
VOC analysis as a cancer fingerprint, showing fair relia-
bility, with a sensitivity ranging between 30% and 94%
and a specificity of 60–94%. This metabolomic ‘signal’
can be differently evaluated according to the analytical
platform used and the matrix analysed, but the lack of a
standardized experimental procedure makes an overview
of the results difficult.
All the experimental projects discussed in this review
were pilot studies containing small numbers of patients
(ranging from 12 to 83) in which different tools, ana-
lytical platforms or experimental procedures were
employed, resulting in noncomparable patterns of
metabolites or electrical signals, although the most fre-
quent chemical classes of metabolites identified were
alkanes, aromatic compounds and alcohols (Table 1).
Two [17,18] of the studies reviewed were conducted
by the same authors using different analytical platforms
(FAIMS, SPME-GC-MS and the WOLF system) to
analyse urine, leading to results which were not compa-
rable. The same was true of the two studies conducted
by Wang et al. [21,22] in which different biological
samples, including exhaled breath and blood from CRC
661
VOC as biomarkers for colorectal cancer
patients, were examined using the same technique
(SPME-GC-MS). The two VOC patterns were not
completely matched, probably due to the different char-
acteristics of the samples. Interestingly, these authors
found that 6-t-butyl-2,2,9,9-tetramethyl-3,5-decadien-
7-yne was present at lower levels in breath and blood
samples [21].
The four studies aiming to detect VOC in breath
samples suggested different patterns of potential
biomarkers which did not fit each other, with the
exception of 1,3-dimethylbenzene, identified by both
Peng et al. [8] and Altomare et al. [19], and of 4-
methyloctane, considered a potential biomarker by Alto-
mare et al. [19] and by Amal et al. [20], probably
because of the different procedures employed in collec-
tion and analysis of the samples [8,19–21]. So far only
one study has focused on the monitoring of VOC from
blood samples [22] and two from faeces [23,24], and
both studies had small sample sizes. These studies
should therefore be considered as only preliminary. Fur-
thermore, the two studies of faeces were carried out
using different techniques (Cyranose 320 and SIFT-
MS) so that the results are not comparable.
Initials publications on the use of e-noses for diagno-
sis reported exciting results with the possibility of devel-
oping a faster, more reliable instrument for cancer
screening. The reliability of the four sensor arrays tested
on CRC patients seems to be good, showing a sensitiv-
ity ranging from 78 to 94% and a specificity of 79–94%
[8,17,20,24]. Although the application of e-noses
seems to receive support from the scientific community,
it must be pointed out that e-noses cannot identify any
specific CRC biomarker. It would be very interesting to
interface GC-MS analysis and e-nose technology to
establish a dedicated e-nose equipped with sensors able
to detect CRC-specific compounds or classes of VOC
identified by GC-MS.
Another debate concerning the metabolomic approach
to CRC screening is the type of biological sample chosen
for VOC analysis. The choice of which biological matrix to
analyse is crucial for increasing patients’ compliance and
adherence to the screening programme, one of the keys to
success for a new screening tool. In this respect, faecal
samples represent the worse choice, as demonstrated by
the insufficient compliance obtained using the FIT. Blood
collection is invasive compared with urine and breath col-
lection. Furthermore, urine, faeces and blood need to be
processed, stored within a few hours of collection, then
defrosted and warmed-up to obtain the headspace for
analysis. Breath testing therefore is the most promising
choice owing to its gaseous state, with no need for pro-
cessing before analysis, easy non-invasive collection and
low cost.
662 Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
M. Di Lena et al.
The idea of a breath test as a screening tool for CRC
is old, and was first proposed in 1977 by Haines et al.
[27], who evaluated the presence of higher methane
concentrations in the breath of CRC patients. This
hypothesis was then rejected by consecutive studies
[28,29] and remained neglected for 30 years. Analysis
of the VOC content breath has reawakened this idea;
however, it is not focused on the monitoring of a single
marker but allows the simultaneous identification of
several compounds. Our group [30] has demonstrated
that the pattern of exhaled VOC in CRC patients is
changed after removal of the cancer, suggesting that the
profile before surgery is strictly linked to cancer meta-
bolism, confirming a close relationship between tumour
metabolism and exhaled VOC. In this study, 11 VOC
were selected which were consistently involved in the
discriminatory models, even in our previous study [19],
and which could represent the core ‘breath print’ of
CRC patients, suggesting, if confirmed in a larger
cohort, the use of the VOC pattern not only for screen-
ing but also for the secondary prevention of CRC.
For any mass screening programme the cost of the
procedure is of great importance, and VOC analysis
using GCMS equipment is neither cheap nor widely
available. The collection of exhaled breath is probably
cheaper than for urine or faeces but the true cost of
these procedures has never been exactly determined. To
fill the gap between the laboratory and clinical practice
there is a need to develop different online analysers such
as dedicated e-noses or quartz-enhanced photoacoustic
spectroscopy sensors [31].
VOC analysis is a very promising new screening tool
for CRC, although it will require much effort to stan-
dardize all the phases of analysis. The small number of
patients in each study, the different analytical platforms
and the biological material used probably account for
the different VOC patterns, meaning that the results of
different studies are not comparable. Biochemical stud-
ies aimed at defining the hitherto unknown molecular
pathways implicated in VOC production in CRC should
be encouraged. This review underlines the presence of a
‘metabolomic signal’ in the excreta of patients with CRC
that could be used for screening and diagnosis of CRC.
M. Di Lena*, F. Porcelli† and D. F. Altomare*
*Department of Emergency and Organ Transplantation, University ‘Aldo
Moro’ of Bari, Bari, Italy and †Department of Biology, University ‘Aldo
Moro’ of Bari, Bari, Italy
E-mail: donatofrancesco.altomare@uniba.it
Received 18 October 2015; accepted 24 November 2015 Accepted
Article online 11 January 2016
PROSPERO registration number: CRD42015023978
M. Di Lena et al.
References
1 Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013.
CA Cancer J Clin 2013; 63: 11–30.
2 Whitlock EP, Lin J, Liles E et al. Screening for Colorectal
Cancer: An Updated Systematic Review [Internet]. Rock-
ville, MD: Agency for Healthcare Research and Quality
(US), 2008. Report No.: 08-05-05124-EF-1.
3 Stracci F, Zorzi M, Grazzini G. Colorectal cancer screen-
ing: tests, strategies, and perspectives. Front Public Health
2014; 2: 210.
4 Pauling L, Robinson AB, Teranishi R, Cary P. Quantitative
analysis of urine vapour and breath by gas–liquid partition
chromatography. Proc Natl Acad Sci USA 1971; 68:
2374–6.
5 Williams H, Pembroke A. Sniffer dogs in the melanoma
clinic? Lancet 1989; 1: 734.
6 Willis CM, Church SM, Guest CM et al. Olfactory detec-
tion of human bladder cancer by dogs: proof of principle
study. BMJ 2004; 329: 712.
7 McCulloch M, Jezierski T, Broffman M et al. Diagnostic
accuracy of canine scent detection in early- and late-stage
lung and breast cancers. Integr Cancer Ther 2006; 5: 30–9.
8 Peng G, Hakim M, Broza YY et al. Detection of lung,
breast, colorectal, and prostate cancers from exhaled breath
using a single array of nanosensors. Br J Cancer 2010;
103: 542–51.
9 Phillips M, Gleeson K, Hughes JMB et al. Volatile organic
compounds in breath as markers of lung cancer: a cross-
sectional study. Lancet 1999; 353: 1930–3.
10 Roberts MJ, Schirra HJ, Lavin MF, Gardiner RA. Meta-
bolomics: a novel approach to early and noninvasive
prostate cancer detection. Korean J Urol 2011; 52: 79–
89.
11 Mangler M, Freitag C, Lanowska M, Staeck O, Schneider
A, Speiser D. Volatile organic compounds (VOCs) in
exhaled breath of patients with breast cancer in a clinical
setting. Ginekol Pol 2012; 83: 730–6.
12 Kwak J, Gallagher M, Ozdener MH et al. Volatile
biomarkers from human melanoma cells. J Chromatogr B
Analyt Technol Biomed Life Sci 2013; 931: 90–6.
13 Chen J-lian, Tang H-qing, Hu J-duo, Fan J, Hong J, Gu
J-zhong. Metabolomics of gastric cancer metastasis
detected by gas chromatography and mass spectrometry.
Metabolomics 2010; 16: 5874–80.
14 Liu R, Peng Y, Li X et al. Identification of plasma metabo-
lomic profiling for diagnosis of esophageal squamous-cell
carcinoma using an UPLC/TOF/MS platform. Int J Mol
Sci 2013; 14: 8899–911.
15 Guo L, Wang C, Chi C et al. Exhaled breath volatile bio-
marker analysis for thyroid cancer. Transl Res 2015; 166:
188–95.
16 Silva CL, Passos M, C^amara JS. Investigation of urinary
volatile organic metabolites as potential cancer biomarkers
Colorectal Disease ª 2016 The Association of Coloproctology of Great Britain and Ireland. 18, 654–663
VOC as biomarkers for colorectal cancer
by solid-phase microextraction in combination with gas
chromatography-mass spectrometry. Br J Cancer 2011;
105: 1894–904.
17 Westenbrink E, Arasaradnam RP, O’Connell N et al.
Development and application of a new electronic nose
instrument for the detection of colorectal cancer. Biosens
Bioelectron 2015; 67: 733–8.
18 Arasaradnam RP, McFarlane MJ, Ryan-Fisher C et al.
Detection of colorectal cancer (CRC) by urinary volatile
organic compound analysis. PLoS One 2014; 9: e108750.
19 Altomare DF, Di Lena M, Porcelli F et al. Exhaled volatile
organic compounds identify patients with colorectal cancer.
Br J Surg 2013; 100: 144–50.
20 Amal H, Leja M, Funka K et al. Breath testing as potential
colorectal cancer screening tool. Int J Cancer 2015; 138:
229–36.
21 Wang C, Ke C, Wang X et al. Noninvasive detection of
colorectal cancer by analysis of exhaled breath. Anal Bioa-
nal Chem 2014; 406: 4757–63.
22 Wang C, Li P, Lian A et al. Blood volatile compounds as
biomarkers for colorectal cancer. Cancer Biol Ther 2014;
15: 200–6.
23 Batty CA, Cauchi M, Lourencßo C, Hunter JO, Turner C.
Use of the analysis of the volatile faecal metabolome in
screening for colorectal cancer. PLoS One 2015; 10:
e0130301.
24 de Meij TG, Larbi IB, van der Schee MP et al. Electronic
nose can discriminate colorectal carcinoma and advanced
adenomas by fecal volatile biomarker analysis: proof of
principle study. Int J Cancer 2014; 134: 1132–8.
25 Tan B, Qiu Y, Zou X et al. Metabonomics identifies serum
metabolite markers of colorectal cancer. J Proteome Res
2013; 12: 3000–9.
26 Qiu Y, Cai G, Su M et al. Serum metabolite profiling of
human colorectal cancer using GC-TOFMS and UPLC-
QTOFMS. J Proteome Res 2009; 8: 4844–50.
27 Haines A, Metz G, Dilawari J, Blendis L, Wiggins H.
Breath-methane in patients with cancer of the large bowel.
Lancet 1977; 2: 481–3.
28 Sivertsen SM, Bjørneklett A, Gullestad HP, Nygaard K.
Breath methane and colorectal cancer. Scand J Gastroen-
terol 1992; 27: 25–8.
29 Kashtan H, Rabau M, Peled Y, Milstein A, Wiznitzer T.
Methane production in patients with colorectal carcinoma.
Isr J Med Sci 1989; 25: 614–6.
30 Altomare DF, Di Lena M, Porcelli F et al. Effects of cura-
tive colorectal cancer surgery on exhaled volatile organic
compounds (VOCs) and potential implications in clinical
follow-up. Ann Surg 2015; 262: 862–7.
31 Patimisco P, Scamarcio G, Tittel Fk, Spagnolo V. Quartz-
Enhanced Photoacoustic Spectroscopy: a Review. Sensors
2014; 14: 6165–206.
663